Steak and Soya: Grilled Meat and Cancer Risk

Abstract

Red meat and processed meat occupy an unusual position in cancer epidemiology: they are among the most studied dietary exposures in existence, they carry formal carcinogen classifications from the World Health Organization's cancer agency, and yet the magnitude of the risk they confer remains actively contested in the peer-reviewed literature.

This review examines the question through two lenses: the Western grilled steak, and soya — the thin-sliced, spice-coated, charcoal-grilled skewered beef known as suya in Nigeria, Ghana, Niger and Sudan and as soya in Cameroon. The two are the same food in chemical terms and very different foods in practice. Soya is cut thinner, cooked longer, basted repeatedly with oil, grilled in direct contact with wood or charcoal smoke, and coated in a spice mixture that is simultaneously one of the most protective and one of the most dangerous things about it. It sets out the epidemiological evidence base and its limitations, then works through the chemistry of the five principal classes of meat-associated carcinogens — heterocyclic aromatic amines (HCAs), polycyclic aromatic hydrocarbons (PAHs), N-nitroso compounds (NOCs), heme-iron-catalysed lipid peroxidation products, and the non-human glycan N-glycolylneuraminic acid (Neu5Gc) — describing for each the precursors, the reaction pathways, the conditions that favour formation, and the metabolic activation steps that convert a food-borne molecule into a DNA adduct.

It then translates that chemistry into quantitative cooking guidance. Temperature, time, heat-transfer geometry, fat dripping, doneness and surface chemistry all influence carcinogen yield by one to two orders of magnitude, and several well-replicated interventions — antioxidant-rich marinades, microwave pre-treatment, frequent turning, indirect heat, lower target internal temperatures — reduce HCA and PAH formation by 50–95% without eliminating steak from the diet.

The soya sections draw on measurements made in Nigeria, Burkina Faso and elsewhere in West Africa rather than on European or North American data. Nigerian studies have repeatedly found benzo[a]pyrene in commercial suya above the European Union maximum permissible level, and traditional wood-fire methods have produced contamination an order of magnitude higher still. A Burkinabè study found apolar heterocyclic amines twelve times more abundant in flame-grilled than in braised chicken.

The review also identifies what may be the single most serious and most neglected cancer risk in the whole suya supply chain, and it is not the smoke. It is aflatoxin B1 in the groundnut-based yaji spice: a Group 1 human carcinogen and a direct cause of liver cancer, found at unsafe levels in the majority of commercial spice samples in published surveys. Any honest article about suya and cancer that discusses only grilling has missed the larger problem.

A concluding synthesis offers practical, evidence-graded protocols for both the home cook and the commercial vendor. Every substantive claim is referenced, and a full annotated review of the cited literature appears at the end.

Disclaimer. This article is a review of published scientific literature written for general education. It is not medical advice, does not establish a clinician–patient relationship, and should not be used to make individual treatment decisions. Anyone with a cancer diagnosis, a family history of cancer, chronic hepatitis B or C, liver disease, haemochromatosis or an inherited colorectal cancer syndrome should discuss dietary changes with their own physician or a registered dietitian.

---

Table of Contents

---

1. Introduction and Scope

Few dietary questions generate as much heat and as little clarity as the relationship between meat and cancer. In October 2015 the International Agency for Research on Cancer (IARC), the specialised cancer agency of the World Health Organization, announced that processed meat would be classified as carcinogenic to humans and red meat as probably carcinogenic to humans.¹² The announcement was widely reported, frequently misunderstood, and in at least one case reduced to the claim that a single burned steak was equivalent to smoking six hundred cigarettes — an assertion so far outside the evidence that it prompted a formal rebuttal in the International Journal of Cancer pointing out that meat consumption is not tobacco smoking.³

The confusion is understandable, because the question involves at least four separable issues that are routinely collapsed into one:

1. Is there a chemical basis for concern? This is a question of food chemistry and toxicology, and the answer is unambiguously yes. Cooked meat demonstrably contains mutagens; the pathways that generate them are well characterised; and several of the resulting compounds are individually classified as probable or possible human carcinogens.
2. Does eating meat measurably increase cancer incidence in populations? This is a question of epidemiology, and the answer is a qualified yes for colorectal cancer, with weaker and more contested signals for several other sites.
3. How large is the effect? This is where the genuine scientific disagreement lies, and where the honest answer is: modest, and considerably smaller than public discourse implies.
4. Can the risk be reduced without eliminating the food? This is a question of practical food science, and the answer is emphatically yes — with reductions in carcinogen formation of 50% to 95% achievable through preparation changes alone.

Soya — the West African grilled skewered meat, called suya in Nigeria and soya in Cameroon — brings all four questions into sharper focus than a Western steak does. It concentrates almost every variable that increases carcinogen formation into a single dish: meat cut very thin, basted repeatedly with oil, grilled in direct contact with charcoal or wood smoke, cooked well-done, and eaten frequently as a street snack. Nigerian analytical studies have measured the result, and the numbers exceed European regulatory limits often enough to matter.

But soya also carries something a plain grilled steak does not: yaji, a heavy coating of groundnut, chilli, ginger, clove and black pepper applied before it ever meets the fire. That spice mixture belongs to precisely the class of antioxidant preparations shown across the food-science literature to suppress heterocyclic amine formation by 50 to 95 percent. It is, in effect, an indigenous mitigation system that predates the chemistry by centuries.

And yaji carries a third thing, which almost no popular article on suya mentions and which may matter more than everything else combined. That is dealt with in Section 9, and readers who want the single most important practical finding in this article should go there first.

This article treats both foods seriously and at length. It is deliberately mechanistic: rather than simply reporting that grilling produces carcinogens, it describes which molecules form, from which precursors, under which conditions, through which reaction steps, and how they are subsequently activated in the human body into the electrophiles that actually damage DNA. That level of detail matters, because the mitigation strategies fall directly out of the mechanisms. Once you know that heterocyclic amines require creatine, that creatine migrates to the meat surface with water, and that phenolic antioxidants intercept the radical intermediates, the practical advice — reduce surface water loss, lower surface temperature, add polyphenols — stops being a list of rules and becomes something you can reason about.

---

2. Definitions

Precision here prevents a great deal of downstream confusion.

Red meat, in the IARC and WCRF sense, means unprocessed mammalian muscle meat: beef, veal, pork, lamb, mutton, horse and goat, including minced or frozen product.² Note that pork is red meat for these purposes, notwithstanding marketing to the contrary. Poultry and fish are not red meat and are not covered by the classification.

Processed meat means meat transformed by salting, curing, fermentation, smoking, or other processes intended to enhance flavour or improve preservation.² Bacon, ham, salami, chorizo, frankfurters, pastrami, biltong, many sausages and most deli meats fall here. Critically, the defining feature is the preservation process, not the presence of a factory. Home-cured bacon is processed meat.

Steak, the specific focus of this article, is unprocessed red meat, usually beef, typically cooked by a high-temperature dry-heat method — grilling, broiling, pan-frying — that maximises surface browning. Steak therefore sits at the intersection of two separate exposures: the intrinsic composition of red meat (heme iron, Neu5Gc, amino acid profile) and the cooking-generated compounds (HCAs, PAHs). These are chemically independent and require independent mitigation.

Soya, in the sense used throughout this article, is the West African street food otherwise spelled suya. It is grilled, skewered, spice-coated meat, and it has nothing whatever to do with the soybean. The terminology is regional: it is called suya in Nigeria, Ghana, Niger and Sudan, and soya in Cameroon, where the equivalent French term is les brochettes.⁵⁴⁵⁵ It originated with the Hausa people of northern Nigeria, Niger and northern Cameroon and has since spread across ethnic and national lines.⁵⁶

It is worth stating the disambiguation plainly for readers who arrive from a search engine: soya (grilled meat) and soya bean (Glycine max) are entirely unrelated foods. This article concerns the former. A short note on the latter appears in Appendix A.

Soya is prepared from boneless meat — beef most commonly, but also ram, goat, pork and chicken, and organ meats including kidney, liver and tripe — sliced thin, threaded onto sticks or skewers, coated with a groundnut-based spice mixture, and roasted around a glowing charcoal fire.⁵⁶ Three main forms are recognised: tsire (the skewered form, and the most commonly preferred), kilishi (sun-dried spiced sheets) and balangu (roasted without the spice coating).⁵⁶

Yaji, also called suya spice or suya sauce, is the compound spice mixture. It is typically built from seven ingredients: groundnut cake powder (kuli-kuli), red pepper (Capsicum annuum), black pepper (Piper nigrum), ginger (Zingiber officinale), clove (Syzygium aromaticum), table salt and monosodium glutamate in the form of a bouillon cube.⁵⁷ The name derives from the Hausa for "hot one".⁵⁷ Regional and vendor variations add garlic, onion powder, nutmeg, Ashanti pepper (Piper guineense), calabash nutmeg or cayenne.⁵⁴ Yaji is applied as a dry rub before grilling, often again during grilling, and served alongside as a dip.

The chemical significance of this composition is developed at length in Section 9. In brief: yaji is a concentrated source of exactly the polyphenolic compounds shown elsewhere in this article to suppress heterocyclic amine formation by 50–95%, and simultaneously the principal vector for aflatoxin contamination in the product.

3. The Epidemiological Foundation

3.1 The IARC evaluation and what a classification actually means

In 2015 an IARC Working Group of twenty-two scientists from ten countries reviewed the accumulated literature on red and processed meat. The scale of the evidence base was substantial: more than seven hundred epidemiological studies on red meat, more than four hundred on processed meat, and more than four hundred on related mechanisms, with the largest single block concerning colorectal cancer.⁴ Greatest weight was given to prospective cohort studies in general populations, with high-quality population-based case-control studies providing supporting evidence.⁴

The conclusions: processed meat was placed in Group 1, carcinogenic to humans, on the basis of sufficient evidence for colorectal cancer. Red meat was placed in Group 2A, probably carcinogenic to humans, on the basis of limited evidence for colorectal cancer together with strong mechanistic evidence.¹² Positive associations were also noted for pancreatic and prostate cancer with red meat, and for stomach cancer with processed meat, though these were not the basis of the classifications.

Three points about this classification are consistently misreported and deserve emphasis.

First, IARC performs hazard identification, not risk assessment. A Group 1 classification states that the agency is confident the agent can cause cancer in humans. It says nothing whatever about how much cancer, at what dose, or with what probability. Tobacco smoking, asbestos, plutonium, solar radiation, alcoholic beverages and processed meat are all Group 1. They are in the same category because the evidence that each can cause cancer is comparably strong, not because the magnitudes are comparable. The magnitudes differ by orders of magnitude.

Second, the Working Group did not evaluate the individual chemicals. The monograph explicitly notes that although the literature on heme iron, heterocyclic aromatic amines, N-nitroso compounds and polycyclic aromatic hydrocarbons is summarised in the mechanistic section, these agents were not themselves evaluated in relation to meat consumption in that volume.⁵ The classification is of the food, not of the constituent molecules.

Third, cooking was excluded from the definition of processing. Handling between abattoir and butcher, and subsequent domestic cooking including the addition of salt and seasoning, were expressly not treated as "processing" for the purposes of the evaluation.⁵ This matters because it means the red meat classification is not principally a statement about grilling.

The associated dose–response figure that entered public consciousness came from the Lancet Oncology summary: an analysis of data across studies estimated that each 50 g daily portion of processed meat was associated with roughly an 18% increase in colorectal cancer risk.¹

3.2 The WCRF/AICR dose–response evidence

The World Cancer Research Fund and American Institute for Cancer Research maintain a Continuous Update Project that performs systematic reviews and dose–response meta-analyses. Their 2018 Third Expert Report concluded that there is convincing evidence that processed meat consumption causes colorectal cancer, and that red meat consumption is a probable cause.⁶⁷

Their quantitative estimates are the ones most often cited in dietary guidelines. A dose–response meta-analysis of fifteen studies found an approximately linear relationship, with about a 12% increase in colorectal cancer risk per 100 g/day increment of red and processed meat combined, and about a 10% increase in colon cancer risk per 50 g/day of unprocessed red meat.⁸

The resulting recommendation is specific: if you eat red meat, limit consumption to no more than about three portions per week, equivalent to roughly 350–500 g of cooked weight (about 12–18 oz), and eat little if any processed meat.⁷ For orientation, 500 g of cooked red meat corresponds to approximately 700–750 g raw, because meat loses roughly a quarter of its weight during cooking.⁷ A moderate portion of 4–6 oz cooked is about the size of one to two decks of cards, and a 4 oz raw portion shrinks to about 3 oz cooked.⁹

The WCRF panel also identified limited suggestive evidence linking red meat to nasopharyngeal, lung and pancreatic cancers, and processed meat to lung, pancreatic and nasopharyngeal cancers — evidence the panel judged too weak to support recommendations.¹⁰

Adherence to the full 2018 WCRF/AICR recommendation set has itself been prospectively evaluated. In cohort analyses, greater adherence was associated with substantially reduced total cancer risk, with each one-point increment in the adherence score corresponding to a 3–4% reduction; notably, more than 90% of participants failed to meet the recommendations on plant food intake and on limiting red and processed meat, which the authors framed as a large unrealised prevention opportunity.¹¹

3.3 The dissenting literature

An honest review must engage with the fact that this consensus is contested by serious methodologists, not merely by industry.

The most prominent challenge came from the NutriRECS consortium, which in 2019 published a series of systematic reviews in Annals of Internal Medicine concluding that the certainty of evidence linking red and processed meat to adverse health outcomes was low to very low, that absolute effect sizes were small, and that most adults could continue current consumption.¹² The associated methodological review by Zeraatkar and colleagues reached similar conclusions on all-cause mortality and cardiometabolic outcomes.¹³ WCRF responded publicly, standing by its recommendations and arguing that the NutriRECS output risked confusing the public and downplaying real risk.¹⁴

A second challenge came from the Burden of Proof analysis published in Nature Medicine in 2022. Lescinsky and colleagues applied a meta-regression approach that relaxes conventional log-linearity assumptions and formally incorporates between-study heterogeneity into the uncertainty estimate. They reported only weak evidence of association between unprocessed red meat consumption and colorectal cancer, breast cancer, type 2 diabetes and ischaemic heart disease, and no evidence of association with ischaemic stroke.¹⁵

A third challenge is methodological rather than empirical. A 2024 analysis fitted least-assumption dose–response models to the same studies used by the Nordic Nutrition Recommendations 2023 and found no statistically significant association between 350 g/week of unprocessed red meat and colorectal cancer (RR 1.04, 95% CI 0.99–1.09), consistent with the least restrictive models of both the Burden of Proof study and WCRF itself.¹⁶ Their central argument is that the association becomes significant at that intake level only under restrictive modelling assumptions such as monotonicity or linearity — in other words, that model assumptions rather than empirical data are driving the threshold at which guidelines are set.¹⁶ Across the twenty-one cohorts analysed, fifteen (65%) reported no association at any consumption level examined.¹⁶

Against these, more recent comprehensive meta-analyses continue to find positive associations. A 2025 meta-analysis searching the literature through November 2024 concluded that high consumption of red and processed meat is significantly associated with increased risk of colorectal, colon and rectal cancer, reinforcing existing recommendations.¹⁷

How should a reader weigh this? A defensible summary is:

• The evidence for processed meat and colorectal cancer is strong, consistent, dose-responsive and mechanistically coherent. It survives most methodological challenges.
• The evidence for unprocessed red meat and colorectal cancer is real but weaker, more heterogeneous, and more sensitive to modelling choices. The direction of effect is consistent; the magnitude and the threshold are not well established.
• Both are observational. Randomised trials of decades-long dietary exposure with cancer endpoints do not exist and will not exist. Residual confounding by the dietary and lifestyle patterns that accompany high meat intake — lower fibre, lower vegetable intake, higher alcohol, higher body weight, lower physical activity, higher smoking prevalence — cannot be fully excluded, though the major cohorts adjust extensively for these.
• The mechanistic evidence is, by contrast, laboratory-based and much less ambiguous. That is a genuine strength of the case, and it is the primary reason IARC placed red meat in 2A despite only limited epidemiological evidence.

3.4 Absolute versus relative risk

Relative risks of 1.12 or 1.18 are easy to sensationalise. They are far less alarming once converted to absolute terms.

Lifetime colorectal cancer risk in most high-income countries is approximately 4–5%. A relative risk increase of 18% applied to a 5% baseline gives roughly 5.9% — an absolute increase on the order of one percentage point over a lifetime for someone eating an additional 50 g of processed meat every single day for decades. That is a real effect at population scale, where a one-point shift across hundreds of millions of people represents a very large number of cancers. It is a modest effect at individual scale.

This distinction is not a rhetorical trick to minimise the finding. It is the correct way to interpret it, and it is why the appropriate response is moderation and preparation change rather than alarm. It is also why the mechanistic and culinary sections that follow are the practically useful part of this article: the achievable reductions in carcinogen formation through cooking technique are proportionally much larger than the risk differences being argued over in the epidemiological literature.

---

4. The Chemistry of Meat Carcinogenesis

There is no single "cancer molecule" in meat. There are at least five mechanistically distinct pathways, three of which are created by cooking and two of which are intrinsic to the meat itself. They differ in their precursors, their reaction conditions, their target organs and — importantly for the practical sections — in whether cooking technique can do anything about them.

Pathway	Origin	Created by cooking?	Principal target	Modifiable by preparation?
Heterocyclic aromatic amines (HCAs)	Muscle creatine + amino acids + sugars	Yes	Colorectum, prostate, breast, pancreas	Yes — substantially
Polycyclic aromatic hydrocarbons (PAHs)	Pyrolysis of fat and fuel	Yes	Colorectum, lung, stomach	Yes — substantially
N-nitroso compounds (NOCs)	Nitrite curing; endogenous nitrosation	Partly	Colorectum, stomach	Partly
Heme iron and lipid peroxidation	Intrinsic to red muscle	No	Colorectum	Partly (co-ingested inhibitors)
Neu5Gc / xenosialitis	Intrinsic to mammalian tissue	No	Carcinomas generally (hypothesised)	No

4.1 Heterocyclic Aromatic Amines

4.1.1 What they are

Heterocyclic aromatic amines are a family of nitrogen-containing polycyclic compounds formed when muscle tissue — meat, poultry or fish — is heated. They are among the most potent mutagens ever identified in the human food supply. In bacterial mutagenicity assays their potency is extraordinary: reviews place them on the order of one hundred to two thousand times more mutagenic per unit mass than aflatoxin and benzo[a]pyrene respectively.¹⁸ They are present in cooked meat at nanogram-per-gram concentrations, with typical domestic cooking producing total HCA levels in the range of roughly 0.1 to 50 ng/g.¹⁸

They divide into two broad structural classes:

Polar (aminoimidazoazaarene) HCAs, formed at ordinary cooking temperatures of roughly 150–300 °C. These are the ones that matter most for domestic cooking:

• PhIP — 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine. Usually the most abundant HCA by mass in well-done meat.
• MeIQx — 2-amino-3,8-dimethylimidazo[4,5-f]quinoxaline. Generally second most abundant, and comparatively enriched in pork, beef and fish.¹⁸
• 4,8-DiMeIQx — 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline
• IQ and MeIQ — the imidazoquinolines
• Harman and norharman — β-carbolines; abundant but weakly mutagenic in themselves, though they can act as co-mutagens

Non-polar (pyrolytic) HCAs, formed above roughly 300 °C by direct amino acid pyrolysis. These include AαC, MeAαC, Trp-P-1 and Trp-P-2. Their formation is largely independent of creatinine availability, and they are the signature of severe charring rather than of ordinary browning.¹⁹

A general ordering in fried, grilled and broiled meat, concentrated in the surface layer, is PhIP > MeIQx > DiMeIQx > IQ, MeIQ, with the highest concentrations of all found in pan residues and grill scrapings.²⁰

The IARC has evaluated several of these individually: PhIP, MeIQ and MeIQx are classified as possibly carcinogenic to humans (Group 2B), and IQ as probably carcinogenic (Group 2A).

4.1.2 The precursors

HCA formation requires three classes of precursor to be present together and heated:²¹

1. Creatine or creatinine — a non-proteinogenic amino acid derivative stored in muscle as creatine phosphate and used as an energy buffer. On heating, creatine cyclises to creatinine, and creatinine supplies the 2-aminoimidazole ring that defines the polar HCAs.¹⁹ Remove creatinine and imidazoquinoline and imidazoquinoxaline formation is disrupted.¹⁹
2. Free amino acids or dipeptides — most importantly phenylalanine (which supplies the phenyl ring of PhIP intact), glycine, threonine and others.
3. Reducing sugars — glucose principally, entering via the Maillard reaction.

This precursor requirement is the single most important fact in the whole of this section, because creatine is found only in animal muscle. Plant proteins contain essentially none. It also means the creatine content of the specific muscle matters, and that grilled plant proteins — beans, groundnut, seitan — generate negligible quantities of the polar HCAs no matter how hard they are grilled.

4.1.3 The formation mechanism

The classical pathway, first proposed by Jägerstad and colleagues in the early 1980s and refined repeatedly since, runs through the Maillard reaction.²²

The Maillard reaction begins when a reducing sugar condenses with an amino group to give a Schiff base, which undergoes Amadori rearrangement. Subsequent degradation of the Amadori product, together with Strecker degradation of amino acids, generates a pool of reactive small molecules: pyridines, pyrazines, aldehydes, and — critically — free radical intermediates. It is postulated that the amino-imidazo portion of the HCA derives from creatine, while the remaining ring system is assembled from these Strecker and Maillard fragments.²³

The radical component is central. Work by Kikugawa and others established that pyrazine cation radicals and carbon-centred radicals generated during the sugar–amino acid Maillard reaction participate directly in producing imidazoquinoxaline-type mutagens.²⁴ Heated model mixtures of glucose with glycine, or glucose with glycine and creatinine, generate detectable unstable carbon-centred radicals.²⁴ Creatine, pyrazines or pyridines, and aldehydes are then assumed to condense to give the IQ-type compounds, with formaldehyde and acetaldehyde serving as one-carbon and two-carbon donors that determine methylation pattern.²⁰

This radical dependence is the key to mitigation. Phenolic antioxidants intercept these radical intermediates. That single fact explains why rosemary extract, grape seed extract, beer, wine, honey, turmeric, garlic and onion all suppress HCA formation, and it converts a scattered list of folk kitchen practices into one coherent chemical strategy.

PhIP formation follows an analogous but distinct route. Isotope-labelling work by Felton's group showed that heating phenylalanine with creatine yields PhIP, that the phenyl ring of phenylalanine is incorporated intact, that the 3-carbon and the amino nitrogen of phenylalanine are incorporated, and that the 1-nitrogen, the methyl carbon and the amino nitrogen of creatine are each incorporated into the product.²⁵ Adding glucose to phenylalanine and creatine increases PhIP yield — but glucose in large molar excess relative to the other precursors is inhibitory.²⁵

That last observation is counterintuitive and worth stating plainly, because it contradicts a common assumption that sugary marinades must be worse. Skog and colleagues found that meat's endogenous sugar content is only about half that of creatine and amino acids, and that raising sugar levels decreased HCA synthesis, an effect attributed to interference with imidazoquinoxaline formation via attack on creatine.²⁶ Glucose, fructose, sucrose and lactose behaved similarly.²⁶ This is part of why honey marinades perform well (Section 6.4).

More recent mechanistic work has extended the picture beyond the Maillard reaction alone. MeIQx synthesis is now understood to involve lipid oxidation and other non-enzymatic pathways as well; Hidalgo, Lavado-Tena and Zamora determined a detailed route in which creatinine reacts with acrolein — a lipid oxidation product — to form an initial adduct that proceeds through imines and enamines, undergoing rearrangement, dehydration and aromatisation to yield MeIQx.²⁷ Lipid oxidation is therefore not a side issue but a contributing pathway, which is a further argument for antioxidant marinades and for trimming excess fat.

4.1.4 Conditions that drive formation

The variables that govern HCA yield are well characterised: temperature, processing time, pH, precursor concentration, amino acid composition, meat type, water activity, storage time of the raw material, and the equipment used.¹⁸²⁸ Two dominate.

Temperature. Higher temperatures increase HCA formation, steeply and reliably.²⁸²⁹ Elevated temperature accelerates dehydration of the meat surface, accelerates conversion of creatine to creatinine, and intensifies the Maillard reaction; all three push in the same direction.³⁰ A controlled air-frying study illustrates the magnitude: total content of eight HCAs in whole chicken air-fried at 200 °C for 40 minutes was 2.56 µg/kg, approximately 2.5 times the 1.03 µg/kg found at 160 °C for 80 minutes — despite the lower-temperature sample being cooked for twice as long.³⁰ Temperature beat time by a wide margin.

The HCA profile also shifts with temperature. In that study AαC, MeAαC, PhIP, harman and norharman were routinely detected, while MeIQx appeared only at higher cooking temperatures.³⁰

Heat transfer geometry. HCA formation increases with cooking methods that transfer heat directly and efficiently from source to food.²⁹ A thin steak in direct contact with a 250 °C cast-iron surface receives far more aggressive surface heating than the same steak suspended in 160 °C convected air, even if both reach the same internal temperature.

Searing. Searing before finishing has a large and specific effect. When beef steak was seared and then air-fried at 200 °C for 13 minutes, total HCA content reached 2.19 µg/kg — 2.6 times the 0.84 µg/kg found in non-seared steak cooked at the same temperature for 18 minutes.³⁰ PhIP was particularly affected, reaching 1.92 µg/kg in the seared sample versus 0.63 µg/kg without searing, and Trp-P-2 was detected only in the seared sample.³⁰

This is an unwelcome finding for anyone who values crust, and it deserves an honest statement: the Maillard chemistry that produces the flavour of a good steak is the same chemistry that produces heterocyclic amines. They cannot be fully decoupled. What can be done is to shift the balance — achieving browning at lower surface temperatures for shorter periods, in the presence of radical scavengers, and with reduced surface creatine. Section 6 sets out how.

Fat and cooking medium. Fat is not neutral. Olive oil and corn oil nearly doubled MeIQx yield in a model system heated for 30 minutes at 180 °C, and oxidised deep-frying oil altered yields further.²¹ Lipid oxidation products feed directly into the acrolein–creatinine pathway described above.²⁷

Salt. Adding sodium chloride reduces cooking water loss, decreases transport of soluble precursors to the meat surface, and results in lower HCA formation.³¹ This is a genuine effect but a limited lever, given the independent cardiovascular and gastric-cancer arguments against high sodium.

4.1.5 Metabolic activation: how an HCA becomes a DNA adduct

An HCA as eaten is not itself the mutagen. It is a pro-carcinogen requiring host metabolic activation, and this two-step activation is where individual susceptibility enters.

Using PhIP as the worked example:

1. Phase I N-oxidation. Cytochrome P450 1A2 (CYP1A2), predominantly hepatic, hydroxylates the exocyclic amino group to give N-hydroxy-PhIP. CYP1A2 activity varies several-fold between individuals and is induced by, among other things, charbroiled meat itself and by cigarette smoke.
2. Phase II esterification. N-hydroxy-PhIP is then esterified by N-acetyltransferase 2 (NAT2) via O-acetylation, or by sulfotransferase 1A1 (SULT1A1) via O-sulfonation, to give an unstable N-acetoxy or N-sulfonyloxy ester.
3. Nitrenium ion formation and DNA adduction. That ester spontaneously heterolyses to a highly reactive arylnitrenium ion, which attacks DNA preferentially at the C8 position of guanine, generating the adduct dG-C8-PhIP.
4. Mutation. If the adduct escapes nucleotide excision repair before replication, it produces characteristic mutations — for PhIP, single-base deletions in repetitive sequences and G→T transversions are typical.

The same general scheme applies to MeIQx and the other aminoimidazoazaarenes. The practical implication is that HCA risk is not determined by intake alone but by intake multiplied by activation capacity. Individuals who are rapid NAT2 acetylators with high CYP1A2 activity generate more reactive ester per unit HCA ingested. Section 11 returns to this.

Additional biology beyond DNA adduction is relevant. PhIP has been reported to have estrogenic activity, which is one proposed reason for associations with breast and prostate rather than only gastrointestinal cancer. HCAs are also detoxified through glutathione conjugation, glucuronidation and sulfation, so the balance of activating and detoxifying enzymes — much of which is influenced by other dietary constituents, notably cruciferous vegetables — is what ultimately determines adduct burden.

4.2 Polycyclic Aromatic Hydrocarbons

4.2.1 What they are and where they come from

PAHs are hydrophobic compounds of two or more fused aromatic rings, built solely of carbon and hydrogen. Around two hundred distinct PAH compounds are recognised.³² Regulatory attention focuses on a subset, with benzo[a]pyrene (BaP) as the classic index compound and the group of four ("PAH4": benzo[a]pyrene, benz[a]anthracene, benzo[b]fluoranthene and chrysene) used as a marker system in European food regulation.

Light PAHs of two or three rings are relatively volatile and of low toxicity; heavy PAHs of four or more rings — including fluoranthene, pyrene, benz[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, indeno[1,2,3-cd]pyrene, dibenz[a,h]anthracene and benzo[ghi]perylene — are more stable and considerably more toxic.³³ Benzo[a]pyrene is an IARC Group 1 human carcinogen.

Unlike HCAs, PAHs in grilled food are not principally formed within the meat. They are formed in the fire and deposited on the meat from smoke. This distinction drives the entire mitigation strategy.

4.2.2 Formation pathways

Several mechanisms operate together during grilling:³⁴³⁵

1. Pyrolysis of dripping fat. Melted fat and cellular juices fall onto the heat source and undergo pyrolysis — thermal decomposition in the near-absence of oxygen. This produces volatile organic fragments which, on incomplete combustion, recombine into PAHs including benzo[a]pyrene.³⁶ These rise in the smoke plume and adsorb onto the meat surface. Fat drippings are described in the literature as playing a crucial role in PAH generation.³⁶
2. Incomplete combustion of fuel. Charcoal, wood and lighter fluid burning with insufficient oxygen produce PAHs directly.³⁵
3. In-food pyrolysis. At temperatures above roughly 200 °C, direct pyrolysis of the food's own fat, protein and carbohydrate contributes.³⁷
4. Radical recombination chemistry. At the molecular level, small radicals build aromatic rings through hydrogen-abstraction–carbon-addition (HACA) sequences and through Diels–Alder cycloadditions, the same pathways that generate soot in flames.³⁴

The dominant variables are therefore fuel type, grilling geometry, temperature, cooking duration and — above all — fat content and whether drippings can reach the heat source.³⁴³⁶

4.2.3 The magnitude of the effect

The difference between direct and indirect grilling is not marginal. Cooking with direct charcoal fire has been shown to increase benzo[a]pyrene levels by five times or more compared with indirect techniques, attributed to proximity of food to the heat source, elevated temperature, and the smoke and tar produced by fat dripping onto the fire.³⁸

Conversely, infrared, electric and pan grilling substantially prevent BaP formation, consistent with the general principle that avoiding direct combustion exposure reduces contamination.³⁶ Reviews report that maintaining stable combustion after visible flames have subsided, and removing dripping fat and charcoal smoke, both decrease PAH content.³⁶

The corollary is that a high-fat cut is a worse choice than a lean one for reasons that have nothing to do with saturated fat and everything to do with drip volume. Studies deliberately select high-fat cuts such as pork neck precisely because they maximise PAH formation.³⁶

It should also be noted that PAH exposure from grilling is not confined to the eater. Reviews of charcoal grilling identify meaningful occupational exposure among people who grill professionally, via inhalation of the smoke.³⁴ Ventilation is a genuine, underappreciated part of the risk picture.

4.2.4 Metabolic activation

Benzo[a]pyrene follows a well-characterised three-step activation:

1. CYP1A1 and CYP1B1 oxidise BaP to benzo[a]pyrene-7,8-epoxide.
2. Epoxide hydrolase converts this to the 7,8-dihydrodiol.
3. A second P450 oxidation gives benzo[a]pyrene-7,8-diol-9,10-epoxide (BPDE), the ultimate carcinogen.

BPDE reacts with the exocyclic amino group of guanine to form the N²-dG-BPDE adduct. If unrepaired, this produces the G→T transversions that are the molecular signature of PAH exposure — the same signature found in the TP53 mutation spectrum of smokers' lung tumours.

The aryl hydrocarbon receptor (AhR) adds a feed-forward loop: PAHs bind AhR, which upregulates CYP1A1 transcription, increasing the very enzymes that activate them.

4.3 Heme Iron and Lipid Peroxidation

This pathway is qualitatively different, because it is intrinsic to red meat and cannot be cooked away. It is also, arguably, the strongest mechanistic explanation for why red meat specifically — rather than poultry or fish, which also form HCAs and PAHs when grilled — is the meat category associated with colorectal cancer.

4.3.1 The three actions of heme

Heme is the iron-porphyrin prosthetic group of myoglobin, present at far higher concentration in red muscle than in white. Undigested heme passes into the colonic lumen, where it acts in three ways:³⁹⁴⁰

Direct cytotoxicity. Heme, or a "cytotoxic heme factor" generated from it, damages colonic surface epithelium, promoting apoptosis of surface cells and compensatory hyperproliferation of the crypt.⁴¹ Sustained compensatory proliferation is itself a tumour-promoting state: more cell divisions means more opportunity for replication of damaged templates and more chance of fixing a mutation.

Catalysis of nitrosation. Heme iron catalyses the endogenous formation of N-nitroso compounds in the gut lumen (Section 4.4).³⁹

Catalysis of lipid peroxidation. Heme iron catalyses peroxidation of dietary polyunsaturated fatty acids, generating a cascade of reactive aldehydes. The principal genotoxic end-products are 4-hydroxynonenal (4-HNE) and malondialdehyde (MDA).³⁹⁴⁰

4.3.2 Why 4-HNE matters

4-HNE is an α,β-unsaturated aldehyde that reacts with DNA bases to form exocyclic propano adducts, principally at guanine. These adducts are mutagenic and, significantly, there is evidence that 4-HNE-derived damage preferentially affects codons in the APC tumour suppressor gene.⁴⁰ Since biallelic APC inactivation is the canonical initiating event in the great majority of sporadic colorectal cancers, a dietary factor that promotes APC mutation has an unusually direct route to colorectal carcinogenesis.

4-HNE additionally acts as a signalling molecule, forming protein adducts, depleting glutathione, and — at the concentrations found in the lumen after a heme-rich meal — selecting for the survival of cells that have already acquired resistance to apoptosis. Bastide and colleagues framed this as heme iron generating a luminal environment that selects for pre-initiated cells.⁴²

Bastide's earlier meta-analysis of heme iron intake and colorectal cancer, published in Cancer Prevention Research, together with the associated mechanistic review, established the model that has dominated the field since: heme catalyses both apparent total N-nitroso compound (ATNC) formation and lipid peroxidation end-product formation, and this partially explains the promoting effect of red and processed meat on colorectal cancer.³⁹

4.3.3 The inhibitors — and why this matters practically

This is the most actionable finding in the entire heme literature. The catalytic effects of heme can be inhibited by trapping the heme. Calcium and chlorophyll both do this.³⁹ Calcium precipitates heme and bile acids in the lumen; chlorophyll, structurally a magnesium-porphyrin closely analogous to heme, appears to form a complex that blocks heme's catalytic activity.

Separately, endogenous ATNC formation is inhibited by vitamins C and E, and polyphenols appear to inhibit lipid peroxidation.³⁹ Red wine polyphenols have been shown to prevent absorption of cytotoxic lipid peroxidation products in humans.³⁹ Dietary antioxidants inhibit metmyoglobin-induced peroxidation of linoleic acid, with the site of action differing between aqueous and lipid phases.³⁹

Translated into food: eating red meat with green leafy vegetables (chlorophyll), with dairy or another calcium source, and with vitamin C-rich foods and polyphenol-rich foods is not a folk superstition. It is a mechanistically grounded intervention targeting a specific catalytic step. Section 6.8 develops this.

4.4 N-Nitroso Compounds

4.4.1 Two sources

N-nitroso compounds reach the colon by two routes.

Exogenous. Nitrite and nitrate are added to cured meats as preservatives — they inhibit Clostridium botulinum, fix the pink colour by forming nitrosylmyoglobin, and contribute the characteristic cured flavour. In the presence of secondary amines and heat, nitrite generates nitrosamines directly in the product. This is why processed meat carries a Group 1 classification while unprocessed red meat does not: curing introduces an entire additional carcinogen pathway that steak does not have. Meat processing by curing and smoking results in the formation of carcinogenic chemicals including N-nitroso compounds and PAHs.²

Endogenous. This is the more important route for unprocessed red meat, and it was established by the Cambridge group led by Sheila Bingham roughly two decades ago, who first showed that volunteers consuming large quantities of red meat excrete correspondingly high quantities of nitroso compounds in faeces.⁴³

4.4.2 The endogenous mechanism

Heme leads to enhanced formation of nitroso compounds in the gastrointestinal tract.⁴³ The species formed are collectively measured as apparent total N-nitroso compounds (ATNC), and comprise principally S-nitrosothiols and nitrosyl heme (FeNO), along with N-nitrosation products of peptide-derived amines.⁴²⁴³

The chemistry: nitrogen oxides, derived from dietary nitrate reduced by oral and gut bacteria and from nitric oxide produced by inflammatory cells, react with thiols and with heme iron. These nitrosated species act as nitrosating agents in turn, transferring the nitroso group to amines and amides to yield N-nitroso compounds.

4.4.3 The DNA lesions

The postulated genotoxic step is alkylation of guanine at the O⁶ position, producing the promutagenic lesions O⁶-methylguanine and O⁶-carboxymethylguanine.⁴³ If these are not repaired in time — the relevant repair protein being O⁶-methylguanine-DNA methyltransferase, MGMT, which is a suicide enzyme consumed stoichiometrically and therefore saturable — they mispair during replication, producing G→A transitions.

This is mechanistically elegant, because G→A transitions at O⁶-alkylguanine sites are precisely the mutation class enriched in KRAS codon 12 and in TP53 in a subset of colorectal tumours. The endogenous formation of nitroso compounds has also been demonstrated in faecal samples from patients with inflammatory bowel disease, which is consistent with the nitric-oxide-driven component of the mechanism.⁴³

4.4.4 Practical implications

Because endogenous nitrosation is heme-catalysed and is inhibited by ascorbate and tocopherol,³⁹ the mitigation strategy overlaps with that for lipid peroxidation. For processed meat, the mitigation is simpler and blunter: eat less of it. There is no preparation technique that removes nitrosamines from bacon.

4.5 Neu5Gc and Xenosialitis

This is the newest and most speculative of the five pathways, and also the one that most elegantly answers a question the others cannot.

4.5.1 The problem it solves

Heterocyclic amines, PAHs and even N-nitroso compounds are also found in cooked fish and poultry — foods that are not associated with colorectal cancer risk in the epidemiology.⁴⁴ Heme iron partly answers this, since white meat and fish contain far less. But an additional hypothesis attributes red meat's specific effect to a molecule that is present in mammalian tissue and absent from birds and most fish.

4.5.2 The biology

Most mammals express two sialic acids on their cell surfaces: N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc). Humans express only Neu5Ac, because we lost a functional CMAH gene — the enzyme that converts Neu5Ac to Neu5Gc — at some point in hominin evolution.⁴⁴

Despite being genetically unable to synthesise it, Neu5Gc is detectable on human epithelial and endothelial surfaces, and at higher levels in malignant tissue.⁴⁵ With no alternative biosynthetic route available, the only possible source is dietary, and food surveys identify red meat as the prominent source.⁴⁵

Metabolic incorporation of dietary Neu5Gc into human tissue makes this glycan the first known example of a xeno-autoantigen — a foreign molecule that becomes displayed on one's own cells.⁴⁵ Most humans carry circulating anti-Neu5Gc antibodies, predominantly IgG, directed against a spectrum of Neu5Gc-containing glycans.⁴⁶ These appear in infants at around six months of age, coinciding with dietary introduction of Neu5Gc, and reach adult levels by the first birthday.⁴⁶ Interestingly, they are probably not induced by direct gut exposure but by commensal bacteria such as Haemophilus influenzae, which scavenge Neu5Gc and display it in their lipooligosaccharides, effectively immunising the host.⁴⁶

The resulting antigen–antibody interaction on the surface of one's own tissues is hypothesised to generate chronic low-grade inflammation, termed xenosialitis, which could contribute to carcinogenesis.⁴⁵

4.5.3 The experimental evidence

The key experiment used Cmah-null mice, which like humans cannot make Neu5Gc. Mice were immunised to generate anti-Neu5Gc antibodies and then fed either a Neu5Gc-rich diet or a matched Neu5Ac-rich control diet. Almost half of the Cmah-null mice that were both fed Neu5Gc and immunised developed hepatocellular carcinomas, including at least one with lung metastasis, in striking contrast to all other groups.⁴⁷ Immunohistochemistry confirmed Neu5Gc incorporation into the tumours.⁴⁷ The combination of dietary Neu5Gc and circulating antibody specifically enhanced carcinoma rates only in Neu5Gc-deficient animals.⁴⁷

In humans, an analysis using samples from the Nurses' Health Studies, the Health Professionals Follow-up Study and EPIC found that total anti-Neu5Gc antibody levels were associated with colorectal cancer risk, and proposed these antibodies as candidate biomarkers for red-meat-associated disease.⁴⁶ More recent work has proposed that dietary Neu5Gc promotes colorectal cancer specifically through upregulation of Wnt signalling.⁴⁴

4.5.4 How much weight to give it

This hypothesis is biologically compelling and it explains the red-versus-white meat discrepancy better than any competitor. But the direct human evidence remains associative, the mouse model is genetically artificial by necessity, and Neu5Gc is present in dairy products as well as red meat, which complicates the epidemiological picture. It should be regarded as a strong hypothesis under active investigation rather than settled mechanism.

It is also, importantly, not modifiable by cooking. Neu5Gc is a structural glycan; grilling technique does not touch it. If this pathway proves to be quantitatively important, the only lever is total red meat quantity.

4.6 Advanced Glycation End Products, TMAO and the Microbiome

Three further mechanisms warrant briefer treatment.

Dietary advanced glycation end products (dAGEs). The same Maillard chemistry that produces HCAs also produces stable glycation adducts such as N^ε-carboxymethyllysine. Dry-heat, high-temperature cooking of animal foods produces the highest dietary AGE loads. AGEs signal through the receptor RAGE, driving NF-κB activation, oxidative stress and chronic inflammation. The epidemiological evidence linking dAGEs specifically to cancer is considerably weaker than for the pathways above, and dAGE content correlates strongly with cooking method, making it hard to separate from HCA and PAH exposure. Treat it as a plausible contributor rather than an established one.

Trimethylamine-N-oxide (TMAO). Gut bacteria metabolise dietary L-carnitine and choline — both abundant in red meat — to trimethylamine, which hepatic flavin monooxygenase 3 oxidises to TMAO. The strongest evidence for TMAO concerns atherosclerosis rather than cancer, though associations with colorectal cancer risk have been reported. Notably, the microbial capacity to produce TMA from carnitine is itself diet-dependent and is markedly lower in long-term vegetarians, which is a nice illustration of how dietary pattern shapes metabolic response to a single meal.

The microbiome and bile acids. High intakes of red meat and fat increase hepatic bile acid secretion. Colonic bacteria with 7α-dehydroxylase activity convert primary bile acids into secondary bile acids, principally deoxycholic and lithocholic acid, which are cytotoxic to colonocytes, induce oxidative DNA damage and promote proliferation. Meat-heavy, fibre-poor dietary patterns also shift the colonic fermentation profile away from saccharolytic (producing protective butyrate) toward proteolytic (producing hydrogen sulfide, ammonia, phenols and p-cresol). Since fibre intake and meat intake are strongly inversely correlated in real populations, some portion of the "meat effect" in observational studies is plausibly a fibre-deficiency effect. This is a confounding problem for the epidemiology and simultaneously a practical opportunity: increasing fibre alongside meat is a mechanistically sound intervention.

4.7 Acrylamide — Relevant, But Not Primarily a Meat Issue

Acrylamide forms from the reaction of the amino acid asparagine with reducing sugars at temperatures above roughly 120 °C. It is classified by IARC as probably carcinogenic to humans (Group 2A) on the basis of animal data.

Acrylamide is predominantly a plant food issue — potatoes, bread, coffee, biscuits — because asparagine is far more abundant in those substrates. It is mentioned here for three reasons: it appears in the air-frying literature alongside HCAs and PAHs;³⁰ higher temperatures and searing increase acrylamide as well as HCAs, particularly in beef;³⁰ and it is relevant to soya specifically, because yaji contains groundnut cake and is often applied with sugar-containing bouillon, both of which supply asparagine and reducing sugars to a surface held well above 120 °C. A heavily spice-crusted, well-charred skewer is a plausible acrylamide source in a way that a plain steak is not. This has not, to the author's knowledge, been measured in suya, and it is a further research gap.

4.8 What Is Probably Not the Mechanism

Three commonly cited culprits deserve qualification.

Saturated fat. Despite decades of assumption, the evidence that saturated fat per se causes cancer is weak. The WCRF/AICR evaluations do not identify saturated fat as a cancer-causing dietary constituent. Fat matters in this context primarily as a substrate: as a source of drippings that pyrolyse into PAHs,³⁶ and as a source of polyunsaturated fatty acids for heme-catalysed peroxidation.³⁹

Animal protein as such. Very high protein intakes stimulate the IGF-1/mTOR axis, and IGF-1 is associated with several cancers. But this applies to protein generally, not to red meat specifically. This pathway does not distinguish beef from fish or beans, and it cannot explain the site-specific epidemiology.

Meat "rotting in the colon." This popular claim has no basis in physiology. Protein digestion and absorption are essentially complete in the small intestine at normal intakes. The colonic effects described in Section 4.6 are real but concern bile acids, undigested heme and fermentation substrate balance, not putrefying meat.

---

5. Cooking Method: A Quantitative Comparison

Having established the chemistry, we can now rank cooking methods rationally rather than by intuition. The two governing questions for any method are:

1. How hot does the meat surface get, and for how long? This determines HCA yield.
2. Can fat and juices reach an ignition source, and can combustion smoke reach the food? This determines PAH yield.

A method can be bad on one axis and good on the other. Charcoal grilling is bad on both. Pan-frying is bad on HCAs but good on PAHs. Sous vide followed by a brief sear is good on both. Understanding the two axes independently is what allows sensible trade-offs.

5.1 Method-by-method assessment

Charcoal grilling over direct flame — highest combined risk.
Direct charcoal fire can increase benzo[a]pyrene by five-fold or more relative to indirect techniques, driven by proximity to the heat source, high temperature, and smoke and tar from fat dripping onto the fire.³⁸ Surface temperatures over direct coals routinely exceed 250 °C, well into the high-yield HCA regime, and the pyrolytic non-polar HCAs begin to appear above 300 °C. This is the worst-case method on both axes simultaneously.

Gas grilling over direct flame — high risk, marginally better.
Gas burns more completely than charcoal, contributing fewer PAHs from the fuel itself, but the dominant PAH source is dripping fat pyrolysing on hot elements or flavouriser bars, and that mechanism is unchanged.³⁶ Flare-ups — the visible flames that erupt when fat ignites — are the single highest-yield PAH event in domestic cooking.

Broiling (overhead grill) — moderate to high.
Very high radiant surface temperature drives HCA formation. PAH exposure is lower than with under-fired grilling because drippings fall away from rather than onto the heat source, unless the pan itself becomes hot enough to smoke.

Pan-frying — high HCA, low PAH.
Direct metal-to-meat contact transfers heat very efficiently, and efficient direct heat transfer increases HCA formation.²⁹ Pan residues carry among the highest HCA concentrations found anywhere in cooked food,²⁰ which is why making gravy from the fond is a meaningfully high-exposure practice. Cooking oil compounds the problem: olive and corn oil nearly doubled MeIQx yield in a model system at 180 °C.²¹

Air frying — moderate, and highly dependent on settings.
The air-frying data are useful because they are recent and well controlled. Total HCAs rose 2.5-fold going from 160 °C/80 min to 200 °C/40 min in chicken, and searing before air-frying at 200 °C raised total HCAs 2.6-fold in beef steak.³⁰ Air frying is not inherently low-carcinogen; it is low-carcinogen only if run at moderate temperature without a searing step.

Roasting and baking — moderate to low.
Oven air temperatures of 160–200 °C are lower than grill surface temperatures, and heat transfer through air is far less efficient than through metal contact or radiation. Surface temperature stays well below the peak values of grilling. Drippings collect in a pan rather than pyrolysing.

Stewing, braising, poaching, steaming, boiling — lowest.
Water-based methods cannot exceed 100 °C at the meat surface. HCA formation is negligible and PAH formation is zero. There is no browning, and therefore no HCA-generating Maillard chemistry. The trade-off is entirely culinary.

Sous vide followed by a brief high-heat sear — low, and the best compromise for steak.
The meat reaches its target internal temperature in a water bath at 50–60 °C, during which zero HCA and zero PAH form. The final sear is then only seconds long, on a dry surface, purely for crust. Total time at high surface temperature is a fraction of that in conventional cooking. This method reduces the high-temperature exposure window from many minutes to under a minute while producing arguably better steak.

Microwaving — very low, and useful as a pre-treatment.
Microwave heating produces minimal browning and minimal HCA formation. Its greatest value is as a pre-treatment step before grilling, discussed in Section 6.5.

5.2 Summary table

Method	Typical surface temp	HCA yield	PAH yield	Overall	Notes
Charcoal, direct flame	>250 °C	Very high	Very high	Worst	BaP up to 5× indirect38
Gas grill, direct flame	230–280 °C	Very high	High	Very poor	Flare-ups dominate PAH
Broiling	250 °C+	High	Low–moderate	Poor	Drippings fall away
Pan-frying	180–230 °C	High	Very low	Poor	Fond/gravy is high-exposure20
Charcoal, indirect + drip guard	150–200 °C	Moderate	Low	Moderate	Large improvement over direct36
Air frying, 180 °C, no sear	180 °C	Moderate	Very low	Moderate	Temperature-sensitive30
Roasting/baking	160–200 °C	Low–moderate	Very low	Good
Sous vide + brief sear	55 °C then brief	Low	Very low	Very good	Minimal high-temp window
Braising/stewing	≤100 °C	Negligible	None	Best	No browning
Microwave	≤100 °C	Negligible	None	Best	Poor culinary result alone

5.3 Doneness

Doneness is a proxy for cumulative surface thermal exposure, and HCA content tracks it closely. Because higher temperatures and longer cooking times both increase HCA production,²⁸ well-done meat carries substantially more HCA than medium-rare meat cooked by the same method.

This produces an unusual convergence: the doneness level most steak cooks prefer on culinary grounds — medium-rare, roughly 52–57 °C internal — is also the lower-HCA choice. This is a rare case where the enjoyable option and the lower-risk option coincide.

Two caveats. First, food safety sets a floor. Whole-muscle steak is safe at lower internal temperatures because contamination is a surface phenomenon and the surface is seared, but ground beef must reach 71 °C (160 °F) throughout, because grinding distributes surface bacteria through the interior. Do not apply steak logic to burgers. Second, immunocompromised individuals, pregnant women, young children and older adults should follow standard food safety guidance regardless of the HCA arithmetic. Acute foodborne illness is a certain and immediate risk; HCA exposure is a probabilistic and distant one.

5.4 Thickness and geometry

A thick cut has a favourable surface-to-volume ratio: it needs a given quantity of surface browning to develop crust, but that browned surface represents a smaller fraction of the total mass. A thin cut requires nearly the same absolute browning spread over a much greater relative surface area.

Practically, a single 2.5 cm (1 inch) steak generates less HCA per gram than the same weight cut as three thin minute steaks. The same logic argues against thin, high-surface-area formats generally.

---

6. The Low-Carcinogen Steak Protocol

Everything in Sections 4 and 5 converges here. The interventions below are ordered by the strength of the supporting evidence and the size of the achievable effect. Nothing in this section requires giving up steak.

6.1 Intervention 1 — Choose lean cuts and trim visible fat

Mechanism: fat drippings are the principal PAH source, pyrolysing on the heat source and depositing PAHs on the food via smoke.³⁶ Fat also supplies polyunsaturated substrate for heme-catalysed peroxidation³⁹ and feeds the acrolein–creatinine pathway to MeIQx.²⁷

Practice: trim external fat before cooking rather than after. Choose sirloin, tenderloin, flank or round over heavily marbled ribeye when grilling over a fire. Note this argument is about PAH chemistry, not about saturated fat and heart disease — it applies specifically when there is an ignition source below the meat.

Expected effect: large on PAHs; moderate on HCAs.

6.2 Intervention 2 — Use indirect heat and a physical barrier

Mechanism: minimising direct contact between meat and flame limits pyrolysis of the organic components of the meat.³⁸ Removing dripping fat and charcoal smoke decreases PAH formation.³⁶

Practice:
- Bank coals to one side and cook the meat on the other; move it over the coals only for the final brief sear.
- Raise the grate. Every centimetre of distance reduces both radiant intensity and smoke deposition.
- Use a drip barrier — a foil tray, a plancha, a grill mat or a cast-iron griddle plate — so that fat never reaches the coals. Commercial grill designs that incorporate a grease drainage system to prevent drips are specifically noted in the literature for reducing PAH formation.³⁶
- Let flames die down and cook over stable, glowing coals rather than active flame; maintaining stable combustion after flames subside reduces PAH content.³⁶
- If a flare-up occurs, move the meat away immediately. Do not spray with water, which raises ash and smoke.

Expected effect: direct-fire cooking can produce five-fold or more benzo[a]pyrene relative to indirect;³⁸ this is the single largest available reduction on the PAH axis.

6.3 Intervention 3 — Lower the temperature, extend the time

Mechanism: temperature dominates HCA formation more than time does. Doubling cooking time while dropping from 200 °C to 160 °C produced 2.5-fold less total HCA, not more.³⁰

Practice: target a moderate cooking temperature and accept a longer cook. For grilling, this means cooking over the cooler zone until the internal temperature is a few degrees short of target, then applying high heat only briefly at the end — the reverse-sear method. For air frying, prefer 160–180 °C over 200 °C.

Expected effect: roughly 60% reduction in total HCAs in the controlled air-frying comparison.³⁰

6.4 Intervention 4 — Marinate with a polyphenol-rich marinade

This is the best-evidenced and most cost-effective intervention available, and the effect sizes are remarkable.

Mechanism: three distinct actions operate together.
1. Radical scavenging. Phenolic compounds intercept the pyrazine and carbon-centred radicals that are obligate intermediates in HCA formation.²⁴ This is the primary mechanism.
2. Surface cooling and moisture retention. A wet surface layer evaporatively buffers the meat surface, keeping it nearer 100 °C for longer and delaying entry into the high-yield temperature regime.
3. Sugar-mediated inhibition. Sugar in molar excess relative to creatine and amino acids inhibits mutagen formation,²⁵²⁶ which is why sweet marinades perform better than intuition suggests.

The quantitative evidence:

Marinade	Matrix	Reported reduction	Ref
Beer, 6 h	Pan-fried beef	PhIP ~88%; MeIQx ~40%	48
Red wine, 6 h	Pan-fried beef	PhIP ~88%; MeIQx ~40%	48
Beer + herbs	Pan-fried beef	~90% of total HCAs	49
Rosemary extract, 10–20% ethanolic, 0.5%	Beef patties	MeIQx up to 92%; PhIP up to 85%	50
Rosemary extract, 2%	Grilled chicken / beef shawarma	PhIP 62%; PhIP and Trp-P-1 up to 100%	50
Grape seed and rosemary extract	Fried beef patties	57% and 90% at highest concentration	51
Turmeric	Air-fried beef steak	69.4% total HCAs	30
Milk or beer	Air-fried chicken	up to 60.6%	30
Gelam (stingless bee) honey	Grilled beef satay	95.1%	52
Brown sugar, olive oil, cider vinegar, garlic, mustard, lemon juice, salt	Grilled beef, 20 min	56 → 1.7 ng/g total HCAs (~97%)	53

Several nuances are worth noting. Beer marinades outperformed white wine marinades, and adding herbs produced a superior inhibitory effect, reducing around 90% of HCAs.⁴⁹ Herbs explained roughly 30% of PhIP inhibition — while alcohol itself increased PhIP formation, meaning the benefit of beer and wine marinades comes from their polyphenols despite, not because of, their ethanol.⁴⁹ Interestingly, no correlation was observed between the radical-scavenging activity of the marinades and total HCA formation in that study,⁴⁹ whereas a separate study found antioxidant capacity did significantly correlate with HCA reduction.⁵¹ The mechanism is therefore not purely a bulk-antioxidant effect; specific compounds matter.

One caution: marinades are not uniformly protective across all HCAs. A study of consumer-tested marinades found that all reduced MeIQx and DiMeIQx (though only under indirect heat), and PhIP under both grilling methods — but a Dijon-based marinade produced surprisingly high harman content.⁵⁰ Effects are compound-specific.

A practical evidence-based marinade. Combining the best-supported elements:

• Acid base: 60 ml red wine, beer, or cider vinegar with lemon juice
• Oil: 30 ml olive oil (use sparingly — oil increases MeIQx in model systems²¹)
• Polyphenol herbs: 2 tbsp finely chopped fresh rosemary and thyme, plus oregano
• Alliums: 3 crushed garlic cloves, 1 tbsp grated onion
• Turmeric: 1 tsp (69.4% reduction reported in air-fried beef³⁰)
• Sugar: 1 tbsp honey or brown sugar
• Salt: 1 tsp (reduces migration of soluble precursors to the surface³¹)
• Black pepper: spicing with pepper is among the practices associated with reduced HCA formation⁵⁰

Marinate at least 1 hour; the beer and wine data used 6 hours,⁴⁸ and there is no evidence of diminishing returns before that point. Pat dry before cooking — a wet surface will steam rather than sear and will also carry marinade sugars that can burn. Discard the used marinade; do not baste with it late in cooking, because it will have accumulated surface juices.

Expected effect: 50–95% reduction in HCAs. This is the highest-yield intervention available.

6.5 Intervention 5 — Microwave pre-treatment

Mechanism: brief microwaving before grilling drives out water carrying dissolved creatine, creatinine, amino acids and sugars. Discarding that liquid physically removes HCA precursors from the meat before the high-temperature step. Since creatinine is required for imidazoquinoline and imidazoquinoxaline formation,¹⁹ removing it caps the yield.

Practice: microwave the steak for 60–90 seconds, pour off and discard the released liquid, then grill. This reduces final grilling time as well, compounding the benefit.

Expected effect: substantial reductions in HCA content have been reported for this technique across multiple studies. The culinary cost is real — some texture and juiciness are lost — which is why this ranks below marinating despite good evidence.

6.6 Intervention 6 — Turn frequently

Mechanism: frequent turning prevents any one surface from accumulating sustained high temperature, and keeps the surface temperature oscillating rather than climbing. It also cooks more evenly, shortening total cook time.

Practice: turn every minute or so rather than once. This contradicts conventional grilling advice about leaving the meat undisturbed to develop crust; the compromise is frequent turning during the bulk of the cook, with a final undisturbed period for crust.

Expected effect: moderate but consistent reductions reported.

6.7 Intervention 7 — Remove char, and never eat the drippings

Mechanism: HCAs and PAHs are surface phenomena, concentrated in the browned and blackened layer. Grill scrapings and pan residues contain the highest measured concentrations of HCAs.²⁰

Practice:
- Cut away visibly blackened areas before eating. This does not remove the browned crust, only the carbonised black portions.
- Do not make gravy or pan sauce from the fond. If you must have a sauce, deglaze with wine and discard the first deglazing, or build the sauce separately.
- Clean the grill between uses. Accumulated carbonised residue on the grate transfers directly to the next piece of meat.

Expected effect: moderate. It removes the highest-concentration fraction of the exposure.

6.8 Intervention 8 — What you eat with the steak

This targets the intrinsic pathways that cooking cannot touch, and it is under-appreciated.

Chlorophyll. Green leafy vegetables trap heme, inhibiting its catalytic activity.³⁹ Spinach, kale, chard, rocket, broccoli, parsley. A large green salad or a side of greens with steak is a mechanistically targeted intervention, not a nutritional platitude.

Calcium. Calcium also traps heme and precipitates bile acids in the colonic lumen.³⁹ Dairy, fortified plant milks, tinned fish with bones, or calcium-set bean curd.

Vitamin C and vitamin E. Endogenous N-nitroso compound formation is inhibited by vitamins C and E.³⁹ Peppers, citrus, tomatoes, berries; nuts and seeds for tocopherols.

Polyphenols. Polyphenols inhibit lipid peroxidation,³⁹ and red wine polyphenols have been shown to prevent absorption of cytotoxic lipid peroxidation products in humans.³⁹ Berries, tea, coffee, cocoa, olive oil, herbs. (This is a statement about polyphenols, not an endorsement of alcohol, which is itself a Group 1 carcinogen.)

Cruciferous vegetables. Broccoli, cabbage, Brussels sprouts, watercress and rocket contain glucosinolate-derived isothiocyanates that induce phase II detoxification enzymes, including glutathione S-transferases, which conjugate and eliminate activated HCA and PAH metabolites. Cruciferous intake also modulates CYP1A2 activity.

Fibre and resistant starch. Fibre dilutes luminal carcinogens, speeds transit, binds bile acids and feeds butyrate-producing bacteria. Butyrate is the primary energy substrate of colonocytes and has direct anti-proliferative and pro-differentiation effects. Since high-meat diets are typically low-fibre diets, this is where much of the practical benefit lies.

Expected effect: unquantified in humans but mechanistically well grounded, and it targets the pathways that no cooking technique can reach.

6.9 Intervention 9 — Quantity and frequency

Ultimately, dose matters more than technique. Limiting red meat to about 350–500 g cooked weight per week — roughly three moderate portions — is the WCRF/AICR recommendation,⁷ corresponding to about 700–750 g raw.⁷ Processed meat should be little if any.⁷

For someone eating well within that range, the marginal benefit of perfect grilling technique is small. For someone eating red meat daily, technique matters considerably more — and so does the quantity itself.

6.10 Intervention 10 — Ventilate

Grilling smoke carries PAHs, and reviews identify meaningful occupational exposure among people working over charcoal grills.³⁴ Grill outdoors or under effective extraction, stand upwind, and do not lean over the smoke plume. This is a genuine inhalation exposure that is entirely separate from what ends up on the plate.

6.11 The complete protocol, assembled

A worked example: reverse-seared marinated sirloin

1. Select a lean sirloin about 2.5 cm thick. Trim all external fat.
2. Marinate 4–6 hours in the polyphenol marinade in Section 6.4.
3. (Optional, highest-protection variant) Microwave 60–90 seconds and discard the released liquid.
4. Pat the surface completely dry.
5. Cook by indirect heat — the cool side of the grill, or a 130 °C oven, or a 55 °C sous vide bath — until the internal temperature is about 5 °C below target.
6. Move to high heat for 45–60 seconds per side only, turning once or twice. Use a barrier plate or plancha if grilling over fire.
7. Rest 5 minutes. Trim any blackened areas.
8. Serve with a large green salad, a cruciferous vegetable, a source of calcium, and something vitamin-C rich. Do not make gravy from the pan residue.

Compared with a marbled ribeye grilled over open flame to well-done and served with pan gravy, this protocol plausibly reduces total HCA and PAH exposure by an order of magnitude, based on the multiplicative effect of interventions each independently documented at 50–95%. It also produces, in most people's judgement, a better steak.

---

7. Soya / Suya: What It Is and How It Is Made

Everything in Sections 4 to 6 applies to soya, because soya is grilled red meat. But it applies with different weightings, and several soya-specific factors have no equivalent in Western barbecue. This section describes the process; Sections 8 and 9 quantify what it produces.

7.1 The traditional process

A typical tsire preparation runs as follows:

1. Cut. Boneless beef — often from the hindquarter, but in commercial practice frequently whatever is affordable — is sliced very thin, into strips or into pieces around 2.5–5 cm across.⁵⁴ Some vendors pound the meat flat to increase tenderness and surface area.
2. Skewer. The strips are threaded onto wooden or metal sticks, spread out rather than packed, so that heat reaches the whole surface.
3. Spice. The meat is coated with dry yaji — groundnut cake powder, chilli, ginger, black pepper, clove, salt and bouillon.⁵⁷
4. Oil. Vegetable oil is brushed onto the meat before and repeatedly during grilling.⁵⁴⁵⁶
5. Grill. The skewers are arranged around or over a glowing charcoal fire, often leaned against a frame surrounding the fire, and turned periodically.⁵⁶ Domestic recipes specify grill or broiler temperatures around 230 °C.⁵⁴
6. Re-spice and serve. More yaji is added after grilling, and the meat is served with raw onion, tomato and pepper sauce.⁵⁵

Home versions adapt this to an oven broiler, which changes the risk profile considerably — see Section 10.

7.2 The four features that distinguish soya from steak

Thinness. This is the single most important structural difference. Section 5.4 explained why a thick cut generates less carcinogen per gram: the browned surface represents a smaller fraction of total mass. Soya inverts this. Thin strips or pounded sheets have a very high surface-to-volume ratio, and heterocyclic amines and PAHs are overwhelmingly surface phenomena.²⁰ Gram for gram, a thin soya strip is close to being all surface. The same weight of beef eaten as a thick steak and as soya will not carry the same carcinogen load even if grilled identically.

Repeated oil basting. Brushing oil onto meat suspended over live coals does two things, both bad. It supplies additional fat to drip onto the fire, and fat dripping onto the heat source is the principal route by which PAHs are generated and deposited back onto the food.³⁶ It also supplies lipid for oxidation, feeding the acrolein–creatinine pathway to MeIQx.²⁷ Recall that olive and corn oil nearly doubled MeIQx yield in a model system heated at 180 °C.²¹

Direct contact with combustion fumes. Traditional West African grilling places the meat in direct contact with wood or charcoal combustion fumes, and this is explicitly identified in the analytical literature as the reason for high PAH contamination in these products.⁵⁸ There is usually no drip barrier, no lid, and no separation between food and fire.

Long total grilling time. Because soya is cooked to a dry, well-done texture and is often held near the fire between sales, cumulative thermal exposure is high. Both temperature and time drive HCA formation,²⁸ and while thin meat cooks quickly, commercial practice frequently involves extended holding.

7.3 What soya has going for it

It would be a distorted account to list only the negatives, because soya has one genuine and substantial advantage over an unmarinated Western steak: it is coated in a heavy layer of polyphenol-rich spice before it ever meets the fire.

Ginger, clove, black pepper, red pepper, garlic and onion are precisely the classes of ingredient shown across the mitigation literature to suppress HCA formation. Turmeric produced a 69.4% reduction in air-fried beef; garlic and rosemary also suppressed formation.³⁰ The use of spices before grilling, and specifically spicing with garlic and onion, with pepper, and with other phenolic-containing spices, is associated with decreased HCA formation.⁵⁰ Clove bud oil appears in the same literature as an effective inhibitor.⁵⁰

Yaji is, in effect, an indigenous antioxidant HCA-mitigation system that predates the science by several centuries. The traditional preparation was not designed with nitrenium ions in mind, but the chemistry is the chemistry.

The problem is that this genuine benefit operates on only one of the two axes. Spice rubs act on HCA formation inside the meat. They do very little about PAHs arriving from outside in the smoke — and PAHs are where the measured African data look worst.

---

8. What Has Actually Been Measured in African Grilled Meat

This is the part of the article where general food-chemistry principles give way to numbers measured on the actual product, in the actual market, using the actual fuel. The findings are not reassuring, and they are also not uniform — which itself is informative, because variation between vendors means the risk is controllable.

8.1 Polycyclic aromatic hydrocarbons in Nigerian suya

The South-West Nigeria survey. A study of 100 commercial grilled meat (suya) samples and 100 smoked catfish samples collected purposively from five major cities in South-West Nigeria between July and October 2019 found varying concentrations of five carcinogenic PAHs: benzo[a]pyrene, benz[a]anthracene, benzo[b]fluoranthene, benzo[ghi]perylene and dibenz[a,h]anthracene.⁵⁹ The authors reported concentrations of carcinogenic PAHs in commercial suya above the EU maximum permissible level of 2.0 µg/kg, and attributed this to poor process control — specifically high grilling temperature and the longer grilling times associated with traditional suya processing.⁵⁹ They concluded this constitutes a public health concern for consumers.⁵⁹

Notably, the same study measured lead, cadmium, mercury, chromium and arsenic in the same samples and found all of them well below WHO permissible limits.⁵⁹ Heavy metals are not the problem here. The combustion products are.

The traditional wood-fire finding. A study using microwave-assisted saponification, solid-phase extraction and HPLC with spectrofluorometric detection examined smoked and grilled meat and fish products commonly consumed in Nigeria. Samples processed using traditional systems involving a wood fire were found to be heavily contaminated with benzo[a]pyrene at levels ranging from 2.4 to 31.2 µg/kg wet weight.⁵⁸ The paper attributes this directly to smoking or grilling carried out with direct contact with wood combustion fumes.⁵⁸

For context, the EU maximum for benzo[a]pyrene in smoked meat is in the range of 2–5 µg/kg depending on the category. The upper end of that Nigerian range is roughly six to thirteen times the limit. It is also an order of magnitude above the levels reported in controlled European grilling studies, where none of the samples exceeded legal BaP limits.³⁶

The Port Harcourt study. Research on traditionally roasted foods in Port Harcourt using GC-FID against fifteen PAH standards detected a mean of 7.23 µg/g of benz[a]anthracene in suya beef, alongside benzo[a]pyrene at 2.41 µg/g and benzo[b]fluoranthene at 4.51 µg/g in roasted mackerel.⁶⁰ Thirteen further PAHs were present in the smoked fish. The authors concluded that regular consumption of suya is risky and may expose consumers to cancer.⁶⁰

A methodological caution is warranted here: these figures are reported in micrograms per gram, three orders of magnitude above the µg/kg figures in the other studies. If taken at face value they would be extraordinarily high. Readers and republishers should treat the units in this particular paper carefully and verify against the primary source; the qualitative conclusion — that suya carries a substantial PAH burden — is consistent across studies regardless.

The Warri study. A 2025 analysis of 24 samples (12 suya, 12 grilled fish) across 13 sampling stations in Warri Metropolis, using HPLC-FID against the 16 EPA priority PAHs, found total 16-PAH concentrations ranging from 0.03 to 2.48 mg/kg. In the suya samples, five of the 13 sampling stations recorded benzo[a]pyrene above the EU limit.⁶¹ The most abundant single PAH in suya was pyrene, at 0.99 mg/kg at one location, and the highest toxicity-equivalent concentration in suya was for benzo[b]fluoranthene.⁶¹

The finding that five of thirteen stations exceeded the limit, and eight did not, is the most practically useful result in this whole section. The risk is vendor-dependent, not inherent to the dish.

The Lagos and Abeokuta study. An assessment of 48 protein food samples from Lagos and Abeokuta established high cancer risk for benzo[a]pyrene, with cumulative cancer-risk values for some smoked fish and crayfish samples exceeding the priority risk level of 1.0 × 10⁻⁴.⁶²

8.2 Heterocyclic amines in West African grilled meat

Direct HCA measurements on suya are scarcer than PAH measurements, but a Burkinabè study provides the most relevant available data and its design is unusually clean.

Researchers screened eight polar and apolar heterocyclic aromatic amines — IQ, Trp-P-1, Trp-P-2, AαC, PhIP, MeAαC, 4,8-DiMeIQx and MeIQx — by HPLC in 29 samples of flamed chicken and 66 samples of braised chicken collected in Ouagadougou.⁶³ The contrast between cooking methods was stark:

	Apolar HAAs	Polar HAAs
Flamed chicken	32.66 ± 10 ng/g	3.48 ± 10.39 ng/g
Braised chicken	2.70 ± 9.67 ng/g	0.92 ng/g
Ratio	~12×	~3×

Apolar HAAs were approximately twelve times more abundant, and polar HAAs approximately three times more abundant, in flamed than in braised chicken.⁶³

Two things follow. First, this is a real-world African confirmation of the temperature and heat-transfer principles set out in Section 4.1.4, measured on street food rather than in a laboratory model system. Second, and more importantly, the apolar HAAs dominated. Apolar HAAs — AαC, MeAαC, Trp-P-1, Trp-P-2 — are the pyrolytic class formed above roughly 300 °C by direct amino acid pyrolysis, and their formation is largely independent of creatinine.¹⁹ Their dominance is the chemical signature of direct flame contact and severe surface charring, not of ordinary browning.

This matters for mitigation, because the yaji spice coating acts most effectively on the Maillard-radical pathway that generates the polar HCAs. It offers much less protection against pyrolytic apolar HCA formation, which is governed by whether flames touch the meat.

8.3 Why soya scores worse than a Western steak

Assembling the evidence, soya as commonly prepared is exposed on every variable that matters:

Variable	Effect on carcinogen yield	Soya as typically prepared
Direct flame contact	BaP up to 5× or more38	Yes, standard
Fat/oil dripping onto fire	Principal PAH source36	Yes, repeated oil basting
Drip barrier	Reduces PAHs36	Rarely used
Surface-to-volume ratio	Surface carries the carcinogens20	Very high — thin strips
Cooking temperature	Dominant HCA driver30	High; ~230 °C or above
Doneness	HCA tracks doneness28	Well-done by convention
Cumulative time near fire	Time increases HCA28	Often extended holding
Fuel	Wood fumes → high BaP58	Charcoal or wood, direct
Antioxidant spice coating	50–95% HCA reduction484950515253	Yes — the one strong protective factor
Accompanying vegetables	Heme/NOC inhibition39	Onion, tomato, pepper — genuinely helpful

The last two rows are why soya is not simply a worst-case food. Yaji is a serious mitigating agent on the HCA axis, and the traditional accompaniment of raw onion, fresh tomato and pepper sauce supplies exactly the vitamin C and polyphenols that inhibit endogenous nitrosation and lipid peroxidation.³⁹ The traditional meal is better designed than it looks.

The problem is concentrated almost entirely on the PAH axis, and PAHs are the thing that spice rubs cannot fix. Section 10 addresses this directly, and the good news is that the interventions required are cheap, physical and immediately implementable.

---

9. Yaji: The Double-Edged Spice

Yaji deserves a section of its own, because it is simultaneously the best and the worst thing about soya from a cancer standpoint. Most popular writing on suya and health ignores it entirely, which is a significant omission.

9.1 The protective side

The composition — groundnut cake, red pepper, black pepper, ginger, clove, salt and bouillon⁵⁷ — is a dense concentration of phenolic antioxidants. Every relevant class in the HCA-mitigation literature is represented:

• Clove (Syzygium aromaticum) is among the most polyphenol-dense spices known, dominated by eugenol. Clove bud oil appears in the inhibitor literature.⁵⁰
• Ginger contains gingerols and shogaols, both effective radical scavengers.
• Black pepper and Ashanti pepper contribute piperine and associated phenolics; spicing with pepper is associated with decreased HCA formation.⁵⁰
• Red pepper / chilli contributes capsaicinoids and carotenoids.
• Garlic and onion, where used, are directly documented as reducing HCA formation,⁵⁰ and garlic suppressed HCA formation in air-fried beef.³⁰
• Groundnut contributes tocopherols and resveratrol.

Given that turmeric alone achieved 69.4% suppression in air-fried beef,³⁰ that rosemary extract reached 85–92%,⁵⁰ and that herb-containing marinades reduced total HCAs by around 90%,⁴⁹ it is entirely plausible that a heavy yaji coating provides substantial HCA protection. To the author's knowledge this has not been directly quantified for yaji specifically, and it is an obvious and worthwhile research gap: a controlled comparison of yaji-coated versus uncoated beef grilled under identical conditions would be a straightforward and valuable study.

9.2 The aflatoxin problem

This is the finding that changes the shape of the whole discussion, and it has nothing to do with grilling.

Aflatoxin B1 is an IARC Group 1 human carcinogen. It is produced by Aspergillus flavus and Aspergillus parasiticus, it contaminates groundnuts, maize and stored spices in warm humid conditions, and it causes hepatocellular carcinoma. Its mechanism is well characterised: metabolic activation to aflatoxin-8,9-epoxide, formation of an N7-guanine adduct, and a characteristic G→T transversion at codon 249 of TP53 — the R249S mutation that is a molecular fingerprint of aflatoxin exposure in liver tumours across West Africa and East Asia. Critically, aflatoxin acts multiplicatively with chronic hepatitis B infection, and hepatitis B prevalence is high across much of West Africa. The combined risk in a co-exposed person is far greater than either exposure alone.

Yaji's principal bulk ingredient is groundnut cake, and groundnut is the classic aflatoxin substrate. The published surveys are alarming.

The Port Harcourt suya spice survey. 150 suya spice mix samples were collected from three locations in Port Harcourt and analysed for fungal contaminants and aflatoxin B1. Approximately 94% of samples were deemed unacceptable for consumption on fungal counts, and 53–100% were unsafe on aflatoxin B1 content against national and international standards.⁶⁴ Aspergillus flavus was the only aflatoxin-producing species isolated, and it was isolated from all 150 samples.⁶⁴

The Ibadan suya spice study. Vended suya spices from five locations in Ibadan were compared against laboratory-prepared control spice. Aspergillus niger, A. flavus, A. parasiticus, A. ochraceus, Fusarium sp., Rhizopus stolonifer, yeast and Trichoderma koningii were all isolated from market samples, and aflatoxins were detected in varying proportions in every location — except the control, which had no detectable aflatoxin.⁶⁵ The measured profile included aflatoxin G1 as the dominant fraction at 32.76%, with individual location values reaching 18.63 µg/kg, alongside detectable B2 and G2 fractions.⁶⁶ The authors attributed contamination to airborne flora, poor sanitary conditions among handlers and producers, and inadequate storage.⁶⁵

Kuli-kuli itself. The groundnut cake ingredient has been studied independently. Analyses of kuli-kuli from Ibadan markets found aflatoxin B1 concentrations ranging from 13 to 2,824 µg/kg.⁶⁷ Earlier work found aflatoxin B1 in all but two of the samples analysed from four major Ibadan markets.⁶⁷ For scale, many national regulatory limits for aflatoxin B1 in food sit at 2–20 µg/kg. The upper end of that range exceeds a 20 µg/kg limit by more than a hundredfold.

Why the control sample matters. In the Ibadan study, laboratory-prepared yaji contained no detectable aflatoxin.⁶⁵ In a separate study of ginger-based yaji, the laboratory-produced sample was uncontaminated while market samples were not.⁶⁴ This is not a problem with yaji as a recipe. It is a problem with sourcing, drying, handling and storage. That is the most important practical conclusion in this section, and it is an encouraging one: the contamination is preventable at every point, and researchers have repeatedly recommended better handling and storage, public enlightenment, and encouragement of industrial production under good manufacturing practice to enable compliance.⁶⁴⁶⁵

9.3 Putting the two risks in proportion

A blunt comparison is useful.

The PAH and HCA exposure from grilled meat is a probabilistic, cumulative, modest-effect-size risk, comparable in magnitude to the red meat associations debated in Section 3, and mostly directed at the colorectum.

Aflatoxin B1 is a Group 1 carcinogen with a known molecular mechanism, a specific mutational signature, a defined target organ, and multiplicative interaction with a viral infection that is endemic in the region. Its dose-response for hepatocellular carcinoma is steep.

If a reader in West Africa takes one action from this article, sourcing clean, properly dried, properly stored, ideally commercially produced yaji from a reputable source is very likely a larger single risk reduction than any change in grilling technique. Home-prepared yaji from freshly bought, visibly sound, properly dried ingredients — discarding any groundnut that is discoloured, shrivelled, mouldy or off-smelling — is the accessible version of this.

Sorting visibly damaged groundnuts is a well-established and effective aflatoxin reduction measure. It does not eliminate the toxin, since aflatoxin is heat-stable and survives roasting and grilling entirely, but it substantially lowers the load.

9.4 Salt, bouillon and the other additives

Yaji contains table salt and monosodium glutamate in the form of bouillon cubes as two of its seven standard ingredients.⁵⁷ Two observations:

Salt is an independent risk factor for gastric cancer through mucosal damage and facilitation of Helicobacter pylori colonisation. Suya is a high-sodium food, additional yaji is typically served alongside for dipping, and consumption is often frequent. This is a genuine, separate concern from the grilling chemistry. Section 4.1.4 noted that salt reduces HCA formation by limiting migration of soluble precursors to the surface³¹ — but that is a poor trade at the sodium levels involved.

Monosodium glutamate is not a carcinogen. Regulatory bodies including the FDA and JECFA have repeatedly assessed it as safe at normal dietary intakes, and the evidence linking MSG to cancer does not exist. Its contribution here is sodium, not glutamate.

Microbiological quality is outside the scope of a cancer article but should be flagged, since it is the more immediate hazard: multiple Nigerian studies have documented poor microbiological quality in both suya and yaji arising from handling and environmental exposure.⁶⁴⁶⁵ Acute foodborne illness from a poorly handled suya is a far more probable near-term harm than any of the carcinogens discussed here.

---

10. The Lower-Risk Soya Protocol

The interventions below preserve the dish. None of them require abandoning charcoal, skewers, yaji or the character of the food. They are ordered by expected effect size.

10.1 For the home cook

1. Break the fat-to-fire pathway. This is the highest-value single change, because it addresses the axis where African measurements look worst. Options, in descending order of effectiveness:
- Grill on a solid plancha, griddle plate or perforated grill pan placed over the coals, so no oil or fat reaches the fire.³⁶
- Bank the coals to one side and grill by indirect heat, moving skewers over the coals only briefly at the end. Direct charcoal fire can produce five times or more the benzo[a]pyrene of indirect technique.³⁸
- Use a drip tray beneath the skewers.
- Raise the grate to increase distance from the coals.

2. Cut back the oil basting. Oil brushed onto meat over open coals feeds the PAH pathway directly.³⁶ Use the minimum needed to keep the meat from drying, apply it before rather than during grilling, and apply it with a brush rather than pouring. If the meat needs moisture during cooking, a water-based or acidic baste is a better choice than oil.

3. Wait for the flames to die down. Grill over stable, glowing coals, never over active flame. Maintaining stable combustion after flames subside reduces PAH content.³⁶ Given that the Burkina Faso data showed apolar pyrolytic HCAs — the direct-flame signature — dominating in flame-grilled meat,⁶³ this matters on both axes.

4. Use clean fuel. Hardwood lump charcoal or clean hardwood only. Never use treated timber, painted or varnished wood, plywood, chipboard, scrap building material, rubber, tyres or plastic to start or feed the fire. These generate PAHs, dioxins and heavy metals far in excess of anything discussed in this article, and the practice is not uncommon where fuel is scarce. Kerosene and lighter fluid should be fully burnt off before the meat goes on.

5. Cut thicker. This runs against tradition and against texture preference, but the surface-to-volume argument is straightforward: thin strips are almost entirely surface, and carcinogens are concentrated at the surface.²⁰ Cutting to 1.5–2 cm rather than paper-thin measurably reduces exposure per gram. A reasonable compromise is a moderately thick cut, cooked slightly less.

6. Keep the yaji, and make it good. The spice coating is protective on the HCA axis.³⁰⁵⁰ Apply it generously before grilling. But source or make the yaji carefully — see point 7, which is more important than everything above it.

7. Control the aflatoxin risk in the yaji. This is probably the single largest risk reduction available.
- Buy groundnuts and kuli-kuli fresh, in small quantities, from sources with visible turnover.
- Sort by hand. Discard any nut that is shrivelled, discoloured, mouldy, damaged or off-smelling.
- Store spices dry, sealed, and away from heat and humidity; buy small and use quickly.
- Prefer commercially produced yaji from a manufacturer operating under good manufacturing practice where available — laboratory-prepared and controlled samples in the published studies were consistently free of detectable aflatoxin.⁶⁴⁶⁵
- Do not assume roasting or grilling destroys aflatoxin. It does not. It is heat-stable.

8. Turn frequently and don't let it char black. Frequent turning limits sustained high surface temperature. Trim off blackened areas before eating; the charred fraction carries the highest concentrations.²⁰

9. Consider the oven-broiler version. Cameroonian home preparations commonly use the oven broiler.⁵⁵ This eliminates the fuel-combustion PAH source almost entirely, since there are no drippings falling into a fire and no wood or charcoal smoke. It costs some smoke flavour and gains a great deal on the PAH axis. A middle path is to finish briefly over coals after cooking through in the oven — the reverse-sear logic of Section 6.11 applied to soya.

10. Eat it the traditional way. Raw onion, fresh tomato, pepper sauce and lime alongside supply vitamin C and polyphenols, which inhibit endogenous nitrosation and lipid peroxidation.³⁹ This is not a modern addition to be improved upon; it is already correct. Add more vegetables rather than fewer, and consider adding a green leafy element for chlorophyll, which traps heme.³⁹

11. Moderate frequency. The WCRF figure of roughly 350–500 g cooked red meat per week applies to soya as much as to steak.⁷ Soya is often eaten as an evening snack alongside a full meal, which makes it easy to exceed intake targets without registering it as a meat portion.

12. Stand upwind. Grilling smoke is an inhalation exposure, and reviews specifically note occupational PAH exposure among people who grill professionally.³⁴

10.2 For commercial vendors and mai suya

The Warri finding that five of thirteen stations exceeded the EU benzo[a]pyrene limit while eight did not⁶¹ is the key datum: vendor practice, not the dish, determines the exposure. The following are low-cost, high-impact changes.

Change	Cost	Effect
Perforated metal plate or grill pan between meat and coals	Low, one-off	Blocks the dominant PAH pathway36
Grease drainage or drip tray	Low, one-off	Documented PAH reduction36
Raise grill height above coals	Nil	Reduces radiant intensity and smoke deposition
Grill only over settled, glowing coals	Nil	Reduces PAHs and pyrolytic HCAs3663
Reduce oil basting frequency	Negative (saves money)	Reduces drip-derived PAHs36
Clean hardwood charcoal only; no scrap, tyres or treated wood	Low	Eliminates a severe contamination source
Source yaji from a reputable supplier; sort and store groundnut properly	Low	Addresses a Group 1 carcinogen64ˑ67
Clean the grill and frame regularly	Nil	Removes accumulated carbonised residue20
Shorter holding time near the fire	Nil	Reduces cumulative thermal exposure28

There is also a commercial argument. Meat grilled over settled coals with a barrier plate, basted lightly, and cut slightly thicker is juicier, loses less weight to rendering, and tastes less acrid. Lower carcinogen formation and better product are, in this case, the same intervention.

10.3 A regulatory and public health note

Several of the Nigerian studies cited here end with the same recommendation, and it is worth amplifying: researchers have called for public enlightenment on contamination risks and for government encouragement of industrial yaji production under good manufacturing practice, precisely because controlled preparation eliminated the aflatoxin problem in their samples.⁶⁴⁶⁵ There is a straightforward, achievable food-safety intervention available here that does not require anyone to stop eating suya.

Equally, none of the PAH mitigation measures in Section 10.2 requires new technology. A metal plate and a drip tray are the intervention. The gap is awareness, not resources.

---

11. Substitution and Dietary Context

Soya is red meat, and the Section 3 evidence applies: roughly 350–500 g cooked red meat weekly is the WCRF/AICR recommendation.⁷ Practical adjustments in a West African context:

Rotate the protein. Grilled chicken and fish soya carry no heme iron and no Neu5Gc, removing two of the five pathways in Section 4 entirely. They still form HCAs — the Burkina Faso data are on chicken⁶³ — and they still pick up PAHs from smoke, so cooking method still matters. But the intrinsic red-meat pathways are gone.

Braise more, flame less. The single cleanest finding in the African literature is that braised chicken carried roughly one-twelfth the apolar HAA content of flamed chicken.⁶³ West African cuisine is rich in stews, soups and braises — egusi, ogbono, pepper soup, groundnut stew — cooked in water at or below 100 °C, where HCA formation is negligible and PAH formation is zero. Shifting the balance between grilled and stewed meat, rather than eliminating grilled meat, captures most of the benefit.

Legumes and vegetables. Cowpea, bambara nut, moin-moin, akara, and leafy greens such as ugu, bitterleaf, ewedu and waterleaf supply fibre and chlorophyll. Fibre dilutes luminal carcinogens and feeds butyrate producers; chlorophyll traps heme.³⁹ In populations where meat intake is rising and fibre intake falling, this is where much of the practical benefit lies.

Kilishi and processed forms. Kilishi is dried and often re-grilled, meaning extended cumulative thermal exposure and, depending on preparation, an additional drying and storage window in which mould can develop on the spice paste. It should be treated as the higher-exposure form of the three.

---

12. Individual Susceptibility and Gene–Diet Interaction

Population-average risk estimates conceal substantial individual variation. Several sources of variation are well characterised.

NAT2 acetylator status. N-acetyltransferase 2 has common polymorphisms producing rapid, intermediate and slow acetylator phenotypes, with markedly different frequencies across populations. Because NAT2 catalyses the O-acetylation step that converts N-hydroxy-HCA into the unstable ester that decomposes to the DNA-binding nitrenium ion, rapid acetylators generate more reactive metabolite per unit HCA ingested. Studies of well-done meat intake and colorectal cancer have reported the strongest associations in individuals who are both rapid NAT2 acetylators and high CYP1A2 metabolisers, which is exactly what the mechanism predicts.

CYP1A2 activity. Varies several-fold between individuals, is inducible by charbroiled meat itself and by cigarette smoking, and is inhibited by some dietary constituents. High activity means more N-hydroxylation, the obligatory first activation step.

GSTM1 and GSTT1 null genotypes. Roughly half of people of European ancestry carry a homozygous deletion of GSTM1, eliminating that glutathione S-transferase. Since glutathione conjugation is a principal detoxification route for activated PAH and HCA metabolites, null individuals may clear them less efficiently — and this is also the genotype in which cruciferous vegetable intake shows the largest protective associations, since isothiocyanates induce the remaining GST isoforms.

MGMT capacity. O⁶-methylguanine-DNA methyltransferase repairs the O⁶-alkylguanine lesions produced by N-nitroso compounds.⁴³ It is a suicide protein, consumed stoichiometrically, so its capacity is finite and saturable. Individuals with lower MGMT expression, whether by genotype or by promoter methylation, are less able to reverse these lesions.

Hepatitis B status. This is the most important interaction in the West African context and it belongs here rather than in a footnote. Aflatoxin B1 and chronic hepatitis B infection act multiplicatively on hepatocellular carcinoma risk, not additively. Someone with chronic HBV who is also exposed to aflatoxin-contaminated yaji is in a far higher risk category than either exposure alone would suggest. Hepatitis B vaccination and treatment status is therefore directly relevant to how seriously an individual should take Section 9.2.

Iron status. People with haemochromatosis or otherwise elevated iron stores have a different risk calculus around heme iron entirely, and should be managing iron intake under medical supervision.

Routine clinical genotyping for these variants is not currently justified, and no guideline recommends it. The practical implication is more modest: population-level advice is an average, and individual response varies more than the average suggests. This is an argument for humility about precise thresholds, not for ignoring the direction of the evidence.

---

13. Practical Synthesis and Risk Hierarchy

13.1 What matters most, ranked

Tier 1 — Largest effect, strongest evidence
1. Source clean yaji and clean groundnut. Aflatoxin B1 is a Group 1 carcinogen causing liver cancer, it acts multiplicatively with hepatitis B, it is heat-stable and survives grilling, and published surveys found 53–100% of commercial suya spice samples unsafe on aflatoxin content while laboratory-prepared controls contained none.⁶⁴⁶⁵ For a West African reader this is very probably the largest single available risk reduction in this entire article, and it has nothing to do with the fire.
2. Limit processed meat to little or none. The only Group 1 food classification in the meat discussion, with convincing evidence and no preparation-based mitigation.⁷
3. Keep total red meat, including soya, within about 350–500 g cooked weight per week.⁷
4. Stop fat and oil reaching the fire. A barrier plate, a drip tray, or indirect heat. Direct charcoal fire can produce five times or more the benzo[a]pyrene of indirect technique,³⁸ and Nigerian surveys repeatedly found commercial suya exceeding EU benzo[a]pyrene limits.⁵⁹⁶¹
5. Use clean fuel only. Hardwood charcoal or clean hardwood. Never treated timber, scrap building material, rubber, tyres or plastic.

Tier 2 — Large effect, good evidence
6. Grill over settled glowing coals, never over live flame. Traditional wood-fire methods produced benzo[a]pyrene at 2.4–31.2 µg/kg,⁵⁸ and flame-grilling produced roughly twelve times the apolar heterocyclic amine content of braising.⁶³
7. Keep and apply the yaji generously — it is a genuine antioxidant HCA inhibitor, in the same class as the spice and herb marinades documented at 50–95% reduction.³⁰⁵⁰
8. Reduce oil basting during grilling.³⁶
9. Cook at moderate temperature for longer rather than high temperature briefly.³⁰
10. Prefer medium doneness to well-done and heavily charred, subject to food safety.²⁸

Tier 3 — Moderate effect, sound mechanism
11. Cut thicker rather than paper-thin; carcinogens are concentrated at the surface.²⁰
12. Trim visible fat before cooking.³⁶
13. Eat it with the traditional accompaniments — raw onion, tomato, pepper, lime — and add green leafy vegetables. These inhibit heme catalysis and endogenous nitrosation.³⁹
14. Increase fibre and legume intake.
15. Remove blackened areas; clean the grill and frame between uses.²⁰
16. Turn frequently.

Tier 4 — Worthwhile, smaller or less certain effect
17. Shift some grilled meat consumption to braised and stewed preparations.⁶³
18. Rotate chicken and fish soya alongside beef, removing the heme and Neu5Gc pathways.
19. Grill outdoors or with good extraction; stand upwind of the smoke.³⁴
20. Watch total sodium, given the salt and bouillon content of yaji.⁵⁷

13.2 The soya summary

• Soya is grilled red meat, and everything in Sections 4 to 6 applies to it.
• It is exposed on almost every variable that increases carcinogen formation: thin cut, high surface-to-volume ratio, direct flame, repeated oil basting, no drip barrier, well-done convention, and wood or charcoal smoke in direct contact with the food.
• Nigerian measurements bear this out. Commercial suya has repeatedly been found above the EU benzo[a]pyrene limit,⁵⁹⁶¹ and traditional wood-fire preparation produced levels up to 31.2 µg/kg.⁵⁸
• But five of thirteen Warri sampling stations exceeded the limit and eight did not.⁶¹ The risk is a function of vendor practice, not of the dish.
• Yaji is a real and substantial protective factor on the heterocyclic amine axis, and an under-recognised one. It is also the principal aflatoxin vector, and that is the more serious problem.
• The interventions that matter most are physical and cheap: a barrier between fat and fire, clean fuel, settled coals, and clean spice.

13.3 What this does not say

It does not say steak or soya causes cancer in any deterministic sense. It does not say a barbecue is dangerous. It does not say anyone should stop eating suya. It says that meaningful, well-characterised chemistry is at work, that the magnitude of the population effect is modest, that the levers for reducing exposure are large and cheap, and that a person who eats soya three times a week from a vendor using a barrier plate, settled coals, clean fuel and properly stored spice is in a substantially different position from one eating charred, oil-basted, thin-cut meat from an open wood fire coated in mouldy groundnut powder. Both are eating the same dish.

---

14. Limitations of the Evidence Base

Intellectual honesty requires stating these plainly.

Observational design. Every human study cited on meat and cancer, and on soy and cancer, is observational. Randomised controlled trials of decades-long dietary exposure with cancer incidence endpoints are not feasible and will not be conducted.

Dietary measurement error. Food frequency questionnaires are imprecise. They typically capture "beef" without capturing doneness, cooking method, fat trimming or marination — precisely the variables this article argues are most important. Studies that have attempted to capture cooking method rely on recall of grilling practice over years, which is very poor data.

Confounding. High red meat intake correlates with lower fibre, lower vegetable intake, higher body weight, higher alcohol, lower physical activity, higher smoking prevalence and lower socioeconomic status. Cohorts adjust for these, but adjustment is imperfect and residual confounding is likely. No cohort study has, to the author's knowledge, examined suya consumption specifically in relation to cancer incidence; the epidemiology here is imported from red meat generally and the local evidence is exposure measurement rather than outcome data.

Model dependence. The 2024 least-assumption analysis makes a genuinely uncomfortable point: whether an association is statistically significant at 350 g/week depends heavily on whether monotonicity and linearity are assumed, and 65% of the individual cohorts reported no association at any consumption level examined.¹⁶ Guideline thresholds are partly artefacts of modelling choices.

Mechanistic-to-human extrapolation. The chemistry is solid, but the doses used in animal carcinogenicity studies of HCAs and PAHs vastly exceed human dietary exposure. Extrapolating from a rodent fed milligram quantities of PhIP to a human consuming nanogram quantities involves assumptions about low-dose linearity that are debatable.

Publication and interest bias. This field has funding from both the meat industry and from organisations with plant-based advocacy positions. Read declarations of interest on primary papers.

Limitations specific to the African data. The Nigerian and Burkinabè studies cited here are the best available and they are genuinely valuable, but they carry real constraints. Sample sizes are modest. Sampling is often purposive rather than random, so the studies characterise selected vendors rather than a representative national picture. Analytical methods vary — GC-FID, HPLC-FID and HPLC with fluorescence detection are not equally sensitive or equally selective — and reported units are inconsistent across papers, with at least one study reporting in µg/g where others report µg/kg.⁶⁰ Recovery rates in one study ranged from 52.9% to 71.3%.⁶¹ Cross-study comparison should therefore be done cautiously, and the direction of the findings should be trusted more than the absolute magnitudes.

No direct HCA data on suya. The heterocyclic amine figures cited are from Burkinabè grilled chicken,⁶³ not from Nigerian beef suya. The mechanism transfers straightforwardly, but the specific numbers do not. Similarly, the protective effect of yaji on HCA formation is inferred from the broader spice and marinade literature³⁰⁵⁰ rather than measured on yaji itself. Both are obvious and tractable research gaps.

Recency. The evidence base moves. Readers should treat the quantitative figures here as accurate to the literature available at time of writing and check for subsequent systematic reviews.

---

15. Conclusions

1. Processed meat is the clearer problem in the general literature. It carries a Group 1 classification, convincing evidence for colorectal cancer, an additional nitrosamine pathway that unprocessed meat lacks, and no meaningful mitigation other than eating less.

2. Unprocessed red meat, including soya, carries a real but modest and contested risk. The mechanistic case — heme iron catalysing lipid peroxidation and endogenous nitrosation, cooking-generated HCAs and PAHs, and possibly Neu5Gc-driven xenosialitis — is strong. The epidemiological effect size is small and sensitive to methodology.

3. Cooking method is a bigger lever than most people realise, and soya sits at the wrong end of it. Thin cuts, direct flame, repeated oil basting, no drip barrier, well-done convention and direct exposure to wood combustion fumes stack every unfavourable variable together. Nigerian measurements show the result: benzo[a]pyrene above EU limits in commercial suya,⁵⁹⁶¹ and up to 31.2 µg/kg in traditionally wood-fired products.⁵⁸

4. But the variation between vendors is the whole story. Five of thirteen Warri stations exceeded the EU benzo[a]pyrene limit; eight did not.⁶¹ That is not a statement about suya. It is a statement about how any given person makes it, and it means the problem is solvable with a metal plate, a drip tray and clean charcoal.

5. Yaji is better than it gets credit for, and worse. As a spice mixture it is a dense source of exactly the phenolic compounds that suppress heterocyclic amine formation — an indigenous mitigation system that anticipates the food chemistry by centuries. As a groundnut-based product handled and stored in warm humid conditions, it is the leading aflatoxin vector in the dish.

6. The aflatoxin finding should reorder people's priorities. Aflatoxin B1 is a Group 1 carcinogen with a defined mechanism, a specific TP53 mutational signature, a defined target organ, and multiplicative interaction with hepatitis B, which is endemic across much of West Africa. Surveys have found 94% of suya spice samples unacceptable on fungal counts and 53–100% unsafe on aflatoxin B1,⁶⁴ with groundnut cake itself measured from 13 to 2,824 µg/kg.⁶⁷ Laboratory-prepared controls contained none.⁶⁴⁶⁵ This is a sourcing, drying and storage problem, and it is fixable. Any article about suya and cancer that discusses only smoke has buried the lead.

7. None of this requires giving up soya. The interventions are physical, cheap and immediately available: a barrier between the fat and the fire, clean fuel, settled rather than flaming coals, a slightly thicker cut, less oil, generous good-quality spice, and the traditional onion, tomato and pepper alongside. Several of them improve the product.

8. The strongest single recommendation remains dietary pattern, not any single food. Moderate red meat, minimal processed meat, thoughtful grilling, clean spice, abundant vegetables and fibre, more braising and less flaming. Every mechanism in Section 4 is addressed by that pattern.

---

Appendix A: A Note on Soya Bean, for Disambiguation

Because this article uses "soya" in its West African sense, a brief note is warranted for readers who arrive expecting the legume.

Soya bean (Glycine max) and soya (grilled meat) are entirely unrelated. The soya bean is a legume rich in isoflavones — genistein, daidzein and glycitein — which bind estrogen receptors and were for two decades suspected of promoting hormone-sensitive cancers. That concern arose from cell-culture and rodent data and has been substantially overturned by human evidence. The resolution lies in receptor subtype selectivity: isoflavones bind preferentially to ERβ, which is generally anti-proliferative, rather than to ERα, which drives proliferation in breast and uterine tissue. At dietary concentrations they behave as selective estrogen receptor modulators rather than as estrogens.

Large prospective cohorts, including the Shanghai Breast Cancer Survival Study, have found soya food intake associated with reduced breast cancer recurrence and mortality, with benefit plateauing at moderate intakes and with no evidence of interference with tamoxifen. Major cancer organisations now regard soya foods as safe for breast cancer survivors.

Two further points of relevance to a West African readership. First, this applies to soya foods — tofu, soya milk, tempeh, edamame — not to high-dose isolated isoflavone supplements, which are a different exposure. Second, soya bean flour and soya-based products are themselves subject to the same aflatoxin storage risks discussed in Section 9.2 wherever they are dried and stored in warm humid conditions.

If there is sufficient reader interest, the soya bean question warrants its own article rather than an appendix.

---

16. Reviewed Literature: Annotated Bibliography

This section reviews every source cited in the article, grouped by theme, with a short assessment of what each contributes and how much weight it should carry. Reference numbers correspond to the numbered list in Section 16.

Editorial note for publication. Author lists, volume numbers and page ranges should be verified against the primary source before publication, particularly for the small number of entries below where the citation is given title-first. DOIs and PubMed/PMC identifiers are supplied where available to make verification straightforward.

16.1 Authoritative evaluations and guidelines

[1] Bouvard V, Loomis D, Guyton KZ, et al. Carcinogenicity of consumption of red and processed meat. Lancet Oncol. 2015.
The formal summary of the IARC Working Group's conclusions, published simultaneously with the October 2015 announcement. This is the primary citable source for the Group 1 / Group 2A classifications and for the widely quoted 18%-per-50g figure for processed meat and colorectal cancer. Authoritative; should be cited in preference to press coverage.

[2] IARC Working Group. Red Meat and Processed Meat. IARC Monographs Vol. 114. Lyon: IARC; 2018.
The full monograph. Roughly five hundred pages covering exposure data, human cancer studies, animal bioassays and mechanistic data. The definitive source for definitions of red and processed meat and for the statement that curing and smoking generate NOCs and PAHs while high-temperature cooking generates HCAs and PAHs. Dense but freely available.

[5] IARC Monograph 114, General Remarks section.
Important for two specific disclaimers that are routinely overlooked: the Working Group did not itself evaluate heme iron, HCAs, NOCs or PAHs in relation to meat consumption, and cooking was expressly excluded from the definition of "processing". Anyone arguing about what the classification does or does not cover should read this section first.

[6] WCRF/AICR. Diet, Nutrition, Physical Activity and Cancer: A Global Perspective. Third Expert Report. 2018.
The other pillar of the evidence base alongside IARC. Uses a graded evidence framework (convincing / probable / limited-suggestive) and generates quantitative dose–response estimates. Judged processed meat a convincing cause of colorectal cancer and red meat a probable cause.

[7][9][10] WCRF and AICR recommendation and evidence pages.
The public-facing summaries of the Third Expert Report. Source for the 350–500 g cooked weight per week figure, the raw-to-cooked conversion, and the limited-suggestive findings for nasopharyngeal, lung and pancreatic cancer. Practical rather than technical.

[8] Nordic Nutrition Recommendations 2023, red meat chapter.
Useful because it restates the WCRF dose–response figures compactly (12% per 100 g/day combined red and processed meat; 10% per 50 g/day unprocessed red meat) and then sets a stricter threshold of 350 g/week, incorporating sustainability alongside health. The stricter threshold is precisely what reference [16] contests.

[11] Adherence to WCRF/AICR 2018 recommendations and cancer risk: prospective cohort studies.
Moves the discussion from single foods to dietary pattern. Each one-point increment in adherence score associated with 3–4% reduced total cancer risk. The finding that over 90% of participants failed the plant-food and meat recommendations quantifies the prevention gap.

16.2 The dissenting and methodological literature

[12] Johnston BC, Zeraatkar D, Han MA, et al. (NutriRECS). Ann Intern Med. 2019.
The most prominent challenge to the consensus. Applied GRADE methodology rigorously and concluded evidence certainty was low to very low. Widely criticised on grounds of framing and of applying a clinical-trial evidence standard to nutritional epidemiology, but the underlying methodological critique is serious and should not be dismissed as industry advocacy.

[13] Zeraatkar D, Han MA, Guyatt GH, et al. Ann Intern Med. 2019.
The companion systematic review on all-cause mortality and cardiometabolic outcomes. Same methodology, same conclusions.

[14] WCRF. "What's the beef? Conflicting recommendations for meat and cancer risk."
WCRF's public rebuttal. Not peer-reviewed, and openly institutional in tone, but useful for understanding how the two camps frame each other's work.

[15] Lescinsky H, Afshin A, Ashbaugh C, et al. Health effects associated with consumption of unprocessed red meat: a Burden of Proof study. Nat Med. 2022;28:2075–82.
Methodologically the most sophisticated entry in the debate. Relaxes log-linearity assumptions and formally incorporates between-study heterogeneity into uncertainty. Found only weak evidence for colorectal and breast cancer and none for ischaemic stroke. Should be read by anyone who wants to understand why the effect size is contested.

[16] "Guidelines to restrict consumption of red meat to under 350 g/wk... are not consistent with health evidence." Nutrition. 2024. PMID 38492553.
A pointed methodological analysis fitting least-assumption dose–response models to the same underlying studies. Its central finding — that 15 of 21 cohorts reported no association at any consumption level, and that significance at 350 g/week depends on assuming monotonicity or linearity — is uncomfortable and should be engaged with rather than ignored. Check the funding declaration.

[17] Comprehensive meta-analysis of red and processed meat and colorectal cancer risk. 2025. PMC12181564.
The most recent large synthesis, searching through November 2024, reaching conclusions supporting the consensus position. Provides the counterweight to [15] and [16].

[3] Gallus S, Bosetti C. Int J Cancer. 2016.
A short but valuable correction of media over-interpretation, pointing out that shared IARC group membership does not imply shared potency. Worth citing whenever the cigarette comparison appears.

[4] Straif K. The IARC Monographs Vol. 114 (presentation).
Not peer-reviewed literature, but useful for the scale of the evidence base reviewed: >700 epidemiological studies on red meat, >400 on processed meat, >400 on mechanisms.

16.3 Heterocyclic aromatic amines: formation chemistry

[22] Jägerstad M, Reuterswärd AL, Olsson R, et al. Creatin(ine) and Maillard reaction products as precursors of mutagenic compounds. Food Chem. 1983.
The foundational paper. Established the creatine–amino acid–sugar precursor triad and the Maillard route. Everything downstream builds on this.

[21] Skog K, Jägerstad M, et al. Factors affecting the formation and yield of heterocyclic amines. Mutat Res. 1995. PMID 8844791.
Model-system work identifying IQx, MeIQx, DiMeIQx, TriMeIQx and PhIP and quantifying precursor effects. Also the source of the finding that olive and corn oil nearly double MeIQx yield at 180 °C. Notably honest about unidentified mutagenic fractions remaining in the system.

[25] Zöchling S, Murkovic M. Formation of the heterocyclic aromatic amine PhIP: identification of precursors and intermediates. Food Chem. 2002.
Reviews the isotope-labelling evidence (Taylor, Fultz, Morris, Knize & Felton 1988) showing exactly which atoms of phenylalanine and creatine end up in PhIP. Also the source of the counterintuitive glucose-excess inhibition finding. Excellent mechanistic detail.

[24] Kikugawa K. Free radical intermediates in Maillard-derived mutagen formation. 1999.
Establishes the radical mechanism — pyrazine cation radicals and carbon-centred radicals. This is the paper that makes antioxidant mitigation make sense, and it is the intellectual bridge between Sections 4.1 and 6.4 of this article.

[27] Hidalgo FJ, Lavado-Tena CM, Zamora R. 2021, as reviewed in Crit Rev Food Sci Nutr 2025.
Modern mechanistic work showing MeIQx formation via creatinine reacting with acrolein, a lipid oxidation product, proceeding through imines and enamines to aromatisation. Extends the picture beyond the Maillard reaction alone and gives a chemical rationale for controlling lipid oxidation.

[18] Heterocyclic Aromatic Amines in Meat: Formation, Isolation, Risk Assessment, and Inhibitory Effect of Plant Extracts. Foods. 2021. PMC8307633.
A comprehensive and accessible review. Source for typical domestic HCA ranges (0.1–50 ng/g), the relative potency comparison against aflatoxin and benzo[a]pyrene, and species-specific HCA profiles. Good entry point to the field.

[19] Heterocyclic Amine Formation and Mitigation in Processed Meat and Meat Products: A Mini-Review. J Food Prot. 2021.
Concise and clear on the creatine-to-creatinine conversion, the requirement of creatinine for imidazoquinoline and imidazoquinoxaline formation, and the creatinine-independence of non-polar HCA formation. This last point is the basis for the claim that plant proteins do not generate polar HCAs.

[20] Smith JS. Formation and Inhibition of Heterocyclic Amines in Cooked Meat. AMSA Reciprocal Meat Conference; 2011.
Conference material rather than a peer-reviewed paper, but it reproduces the Knize & Felton (2005) concentration data showing that pan residues and grill scrapings carry the highest HCA levels — the empirical basis for the advice against making gravy from fond.

[23] NTP. Heterocyclic Amines (Selected): PhIP, MeIQ, MeIQx. Report on Carcinogens, 15th Edition.
Government toxicological monograph. Authoritative on structures, formation conditions and carcinogenicity classification.

[28][29] Knize MG et al. 1994; Skog K et al. 1995; Layton DW et al. 1995.
The classic dose–response work establishing that higher temperature and longer time increase HCA yield, and that direct or efficient heat transfer increases it further. Cited throughout the mitigation literature.

[31] Persson E, Sjöholm I, Skog K. 2003 (and Smith, Ameri & Gadgil 2008).
Source for the salt effect: reduced cooking water loss, reduced transport of soluble precursors to the surface, lower HCA formation.

16.4 Cooking method and quantitative mitigation

[30] Kwon et al. Mitigation of heterocyclic amines, polycyclic aromatic hydrocarbons, and acrylamide in air-fried chicken and beef: effects of cooking methods and marinades. Food Sci Biotechnol. 2025.
The most useful single paper in the practical sections of this article. Modern analytical methods (UHPLC-QqQ and GC-MS), ten HCAs, four PAHs and acrylamide quantified, with controlled variation of temperature, time, searing, marination and spices. Source for the 2.5-fold temperature effect, the 2.6-fold searing effect, turmeric at 69.4% reduction and milk/beer marinades at up to 60.6%. Recent, well controlled, and directly applicable to domestic cooking.

[48] Melo A, Viegas O, Petisca C, Pinho O, Ferreira IMPLVO. Effect of beer/red wine marinades on the formation of heterocyclic aromatic amines in pan-fried beef. J Agric Food Chem. 2008;56(22):10625–32.
The single most-cited marinade study, and deservedly so: controlled temperature and time, HPLC-DAD/FLD quantification, unmarinated reference samples. PhIP reduced approximately 88% and MeIQx approximately 40% after six hours. Also honest about high variability for AαC (7–77%).

[49] Viegas O, et al. Inhibitory effect of antioxidant-rich marinades on the formation of heterocyclic aromatic amines in pan-fried beef. J Agric Food Chem. 2012;60(24):6235–40.
Important refinement. Beer outperformed white wine; adding herbs gave a superior effect, reducing around 90% of total HCAs; herbs explained roughly 30% of PhIP inhibition; alcohol itself increased PhIP formation; and no correlation was found between bulk radical-scavenging activity and HCA reduction. The last two findings are the ones most often omitted from popular summaries.

[50] Puangsombat K, Smith JS. Inhibition of heterocyclic amine formation in beef patties by ethanolic extracts of rosemary. J Food Sci. 2010;75(2):T40–7.
Systematic comparison of extraction solvents, finding 10% and 20% ethanolic rosemary extracts most effective (MeIQx up to 92%, PhIP up to 85%), attributed to synergy between rosmarinic acid, carnosol and carnosic acid. The associated literature (Awney & Sindi 2010) reports up to 100% inhibition of PhIP and Trp-P-1 in grilled chicken at 2% rosemary extract.

[51] Gibis M, Weiss J. Antioxidant capacity and inhibitory effect of grape seed and rosemary extract in marinades on the formation of heterocyclic amines in fried beef patties. Food Chem. 2012.
Reductions of 57% and 90% at the highest extract concentrations. Two findings raise its practical value: antioxidant capacity did significantly correlate with HCA reduction here (contrasting with [49]), and sensory evaluation of grape-seed-marinated patties scored excellently — mitigation without a palatability penalty.

[52] The Inhibitory Effects of Heterotrigona itama Honey Marinades on the Formation of Carcinogenic Heterocyclic Amines in Grilled Beef Satay. Molecules. 2020. PMC7504569.
Reductions of 95.1%, 88.5%, 85.7% and 57.2% for gelam, starfruit, acacia and Apis honeys respectively, versus a table-sugar control, with PLS regression showing inhibition correlated with honey antioxidant activity. Demonstrates that a sweet marinade is not automatically a worse marinade.

[53] Salmon CP, Knize MG, Felton JS. Effects of marinating on heterocyclic amine carcinogen formation in grilled chicken. Food Chem Toxicol. 1997.
Source of the striking figures for a simple household marinade (brown sugar, olive oil, cider vinegar, garlic, mustard, lemon juice, salt): total detectable HCAs reduced from 56 to 1.7 ng/g at 20 minutes grilling, 158 to 10 at 30 minutes, and 330 to 44 at 40 minutes. Note the absolute values rising steeply with grilling time even in the marinated samples — marination reduces, it does not abolish.

16.5 Polycyclic aromatic hydrocarbons

[34] Mechanisms, Exposure, and Reduction Strategies of Polycyclic Aromatic Hydrocarbons in Charcoal-Grilled Meat. Meat Technology. 2025.
Recent, comprehensive, and the source for the HACA and Diels–Alder ring-building mechanisms as well as for occupational exposure among people who grill professionally. Good synthesis of the mitigation options.

[36] Evaluation of Polycyclic Aromatic Hydrocarbons (PAHs) in Pork Meat Cooked with Two Different Methods. Molecules. 2025;30:1886.
Well-designed controlled comparison using deliberately high-fat pork neck, cooking to a fixed 72 °C core temperature, with UHPLC-fluorescence quantification. Source for the role of fat drippings and pyrolysis, for the finding that infrared, electric and pan grilling avoid BaP formation, and for the value of grease drainage systems and stable post-flame combustion. Notably, none of the samples exceeded legal BaP limits — a useful corrective to alarmism.

[38] Lee et al. 2016; Anjum et al. 2019, as reviewed in the grilling-procedure literature.
The source of the headline five-fold-or-more increase in benzo[a]pyrene with direct charcoal versus indirect technique. This is the largest single mitigation effect available in the PAH literature.

[35] Factors and reduction strategies of PAHs formation in charcoal-grilled food products. J Life Sci Agric Technol. 2023.
Reviews the five contributing reactions — incomplete combustion, pyrolysis, Maillard, lipid oxidation and degradation — and is appropriately candid that the mechanism is not fully resolved.

[37] The Presence of Polycyclic Aromatic Hydrocarbons in Grilled Beef, Chicken and Fish, with Dietary Exposure and Risk Assessment. 2020. PMC7492177.
Provides actual measured Σ16PAH contamination levels across five heat-treated meat products (4.42–7.26 µg/kg) plus a dietary exposure assessment. Useful for grounding the discussion in real concentrations rather than relative effects.

[32][33] ATSDR 1995; Singh et al. 2016; Hamidi et al. 2016; Ewa & Danuta 2017.
Background chemistry: the ~200 recognised PAH compounds, and the light-versus-heavy toxicity distinction that justifies focusing on four-plus-ring compounds.

16.6 Heme iron, lipid peroxidation and N-nitroso compounds

[39] Bastide NM, Pierre FHF, Corpet DE. Heme iron from meat and risk of colorectal cancer: a meta-analysis and a review of the mechanisms involved. Cancer Prev Res. 2011;4(2):177–84.
The key paper for the heme hypothesis, combining a meta-analysis with a mechanistic synthesis. Most practically valuable for the inhibitor findings: heme catalysis blocked by calcium and chlorophyll; endogenous ATNC formation inhibited by vitamins C and E; polyphenols inhibiting lipid peroxidation; red wine polyphenols preventing absorption of cytotoxic peroxidation products in humans. Section 6.8 of this article is essentially derived from this reference.

[42] Bastide NM, et al. A Central Role for Heme Iron in Colon Carcinogenesis Associated with Red Meat Intake. Cancer Res. 2015;75(5):870–9.
Experimental follow-up. Sets out the three-mechanism framework (NOCs, HCAs, heme) and provides the methodological detail for measuring ATNC, faecal heme and TBARS. Advances the model that heme creates a luminal environment selecting for pre-initiated cells.

[40][41] The role of heme iron molecules derived from red and processed meat in the pathogenesis of colorectal carcinoma. Crit Rev Oncol Hematol. 2018.
Source for the "cytotoxic heme factor", surface epithelial apoptosis with compensatory hyperplasia, and — importantly — the link between heme-driven lipid peroxidation, DNA adducts and APC mutation. The APC connection is what makes this mechanism specifically colorectal.

[43] Red Meat-Derived Nitroso Compounds, Lipid Peroxidation Products and Colorectal Cancer. Foods. 2019;8(7):252. PMID 31336781.
A focused and well-written review. Traces the field from Bingham's original faecal ATNC work through to the identification of S-nitrosothiols and nitrosyl heme as the principal species, and the O⁶-methylguanine and O⁶-carboxymethylguanine lesions as the proposed genotoxic endpoint. Appropriately careful in distinguishing demonstrated from postulated steps.

16.7 Neu5Gc and xenosialitis

[45] Samraj AN, Pearce OMT, Läubli H, et al. A red meat-derived glycan promotes inflammation and cancer progression. PNAS. 2015;112(2):542–7.
The central paper. Establishes Neu5Gc as a dietary xeno-autoantigen, documents its enrichment in carcinomas, and reports the Cmah-null mouse experiment in which nearly half of Neu5Gc-fed, antibody-bearing animals developed hepatocellular carcinoma. Elegant experimental design, though necessarily reliant on a genetically artificial model.

[46] Polyclonal human antibodies against glycans bearing red meat-derived Neu5Gc... total antibody levels are associated with colorectal cancer risk. PLoS One. 2018. PMID 29912879.
The human-population arm. Uses samples from the Nurses' Health Studies, the Health Professionals Follow-up Study and EPIC. Also documents the developmental appearance of anti-Neu5Gc antibodies at around six months of age and the proposed bacterial route of immunisation via Haemophilus influenzae.

[44] Dietary intake of the red meat-derived glycan Neu5Gc fuels colorectal cancer through up-regulation of Wnt signaling. 2025.
The most recent extension, proposing a specific signalling mechanism. Also states clearly the argument that motivates the whole hypothesis: HCAs, PAHs, NOCs and heme are also found in fish and poultry, which are not linked to colorectal cancer risk.

[47] Samraj AN, et al. AACR abstract LB-156. Cancer Res. 2014;74(19 Suppl).
Conference abstract preceding [45]. Cited for completeness; the full paper supersedes it.

16.8 Soya / suya: description and preparation

[68] Suya: Traditional West African Skewers. 196 Flavors. and [69] Cameroonian Soya. Precious Core.
Culinary rather than scientific sources, cited only for the descriptive facts they are competent to establish: the regional distribution of the dish across Ghana, Nigeria, Niger, Cameroon and Sudan; the naming convention (suya in Nigeria, soya in Cameroon, brochettes in French); typical cut dimensions; the yaji ingredient list as used domestically; the oil-basting step; and the ~230 °C grill or broiler temperature specified in home recipes. Not cited for any health claim.

[70] Okhuebor S, Izevbuwa O. A review of street grilled meat (suya) in Benin City, Nigeria: a potential public health risk. 2020.
The best available structured description of commercial suya production, including the three principal forms (tsire, kilishi, balangu), the use of innards alongside muscle meat, and the composition of the coating. Frames suya explicitly as a public health question, which is the correct starting point.

[71] Chemical Composition and Biological Effects of Yaji: A Popular Nigerian Composite Suya Sauce. Trends Appl Sci Res. 2019;14:215–25.
The definitive reference for yaji composition: the seven standard ingredients, the Hausa etymology, and its use across beef, goat, pork and chicken preparations. Also reviews the biological effects literature on individual constituents. This is the paper to cite for what yaji actually is.

16.9 PAH, HCA and aflatoxin measurements in African grilled meat and spice

[72] Determination of polycyclic aromatic hydrocarbons (PAHs) in commonly consumed Nigerian smoked/grilled fish and meat. Food Addit Contam Part A. 2009;26(7).
Methodologically the strongest of the Nigerian PAH papers — microwave-assisted saponification with simultaneous extraction, SPE cleanup, HPLC separation and spectrofluorometric detection. Source for the headline finding that traditional wood-fire products carried benzo[a]pyrene at 2.4–31.2 µg/kg wet weight, and for the explicit attribution to direct contact with wood combustion fumes.

[73] A Study on Polycyclic Aromatic Hydrocarbon and Heavy Metal Concentrations of Commercial Grilled Meat (Suya) and Smoked Catfish from South-West Nigeria. Polycyclic Aromatic Compounds. 2021;42(6).
The largest survey cited here: 100 suya and 100 smoked catfish samples across five South-West cities. Found five carcinogenic PAHs above the EU 2.0 µg/kg limit and attributed this to high grilling temperature and long grilling time. Valuable for its negative finding too — lead, cadmium, mercury, chromium and arsenic were all well below WHO limits, which usefully rules heavy metals out of the picture.

[74] Genotoxicity and carcinogenicity of traditionally roasted meat using indicator PAHs, Port Harcourt, Nigeria.
Reports benz[a]anthracene at a mean of 7.23 µg/g in suya beef, with benzo[a]pyrene and benzo[b]fluoranthene in roasted mackerel. Units caution: figures are reported per gram rather than per kilogram, three orders of magnitude above comparable studies. Verify against the primary source before quoting the absolute values; the qualitative conclusion is consistent with the rest of the literature.

[75] Effect of polynuclear aromatic hydrocarbon levels in suya and barbecue fish on cancer risk index in Warri Metropolis, Southern Nigeria. J Agric Food Environ. 2025;6(1):5–12.
The most useful paper for practical purposes, because it reports between-vendor variation: five of thirteen sampling stations exceeded the EU benzo[a]pyrene limit and eight did not. Full 16-EPA-PAH panel by HPLC-FID with stated LOQ and recovery rates (52.9–71.3%), which is more methodological transparency than several of the others offer.

[76] Levels and health risk assessment of polycyclic aromatic hydrocarbons in protein foods from Lagos and Abeokuta, Southwestern Nigeria. 2019.
Adds formal cancer risk index calculation, finding cumulative risk above the 1.0 × 10⁻⁴ priority level for some smoked fish and crayfish samples. Note the internal inconsistency in reported congener magnitudes; treat the risk-index framework as the contribution rather than the absolute concentrations.

[77] Assessment of heterocyclic aromatic amines contents in flamed and braised chicken in Burkina Faso. 2022. PMC9803208.
The single most valuable African data point in this article, and the only direct HCA measurement on West African street-grilled meat cited here. Ninety-five samples, eight polar and apolar HAAs by HPLC, with a clean cooking-method contrast: apolar HAAs approximately twelve times and polar HAAs approximately three times higher in flamed than braised chicken. The dominance of the apolar (pyrolytic, >300 °C, creatinine-independent) fraction is the chemical signature of direct flame contact and is what makes this study mechanistically interesting rather than merely descriptive.

[78] Potential public health risks associated with suya spice mix in Port Harcourt, Nigeria. Toxicol Environ Health Sci. 2020;12.
The key aflatoxin paper. 150 suya spice samples, three locations. Approximately 94% unacceptable on fungal counts; 53–100% unsafe on aflatoxin B1 against national and international standards; Aspergillus flavus isolated from all 150 samples and the only aflatoxin producer found. Also reports that laboratory-produced yaji was uncontaminated — the finding that makes this a solvable problem rather than an inherent one.

[79][80] Jonathan SG, Adeniyi MA, Asemoloye MD. Fungal Biodeterioration, Aflatoxin Contamination, and Nutrient Value of "Suya Spices". Scientifica. 2016;2016:4602036. PMC4820623.
Five Ibadan locations compared against an aseptically prepared laboratory control. Isolated Aspergillus niger, A. flavus, A. parasiticus, A. ochraceus, Fusarium sp., Rhizopus stolonifer, yeast and Trichoderma koningii. Aflatoxins detected in every market sample and in none of the control. Reports the fractional breakdown (G1 dominant at 32.76%, with location values to 18.63 µg/kg). Attributes contamination to airborne flora, handler hygiene and storage — all modifiable.

[81] Aflatoxin in Nigerian groundnut cake (kuli-kuli): market surveys. (Including Ezekiel CN, et al. 2012.)
Establishes the contamination level in yaji's bulk ingredient rather than in the finished spice. Aflatoxin B1 in Ibadan kuli-kuli reported from 13 to 2,824 µg/kg; earlier Ibadan market work found aflatoxin B1 in all but two samples. Given that typical regulatory limits sit at 2–20 µg/kg, the upper end of this range is extreme. Author list and exact values should be verified against the primary papers before republication, as these figures are reported across several overlapping studies.

16.10 Overall assessment of the evidence base

Weighing the corpus as a whole:

Strongest evidence: the formation chemistry of HCAs and PAHs, and the effect of cooking variables on their yield. This is laboratory science with controlled conditions, direct measurement and good replication. The mitigation figures in Section 6 rest on firm ground.

Strong evidence: processed meat and colorectal cancer. Convergent epidemiology, dose–response, mechanistic plausibility and formal evaluation by two independent expert bodies.

Moderate evidence: unprocessed red meat and colorectal cancer. Direction consistent, magnitude contested, sensitive to modelling assumptions.[15][16]

Strong evidence: aflatoxin B1 as a cause of hepatocellular carcinoma, and its multiplicative interaction with hepatitis B. This is settled science with a known mechanism and a specific mutational signature. What is less certain is the quantitative exposure attributable to yaji specifically, since the surveys measure contamination in the spice rather than intake in consumers.

Moderate evidence: the exposure measurements in African grilled meat. The direction is consistent across independent studies and locations — commercial suya carries a meaningful PAH burden, and flame-grilling generates far more heterocyclic amines than braising. The absolute magnitudes are less reliable, given small samples, purposive sampling, varied analytical methods and at least one unit inconsistency.⁶⁰

Emerging or hypothesis-stage: Neu5Gc and xenosialitis; dietary AGEs; TMAO. Biologically interesting, not yet actionable.

Unmeasured: the protective effect of yaji on HCA formation. Strongly plausible by analogy with the spice and marinade literature,³⁰⁵⁰ but never directly quantified. A controlled comparison of yaji-coated and uncoated beef grilled under identical conditions is an obvious, cheap and publishable study that nobody appears to have done.

Weakest: precise numerical thresholds of any kind. The 350 g and 500 g figures are reasonable public-health heuristics derived from linear extrapolation, not empirically demonstrated inflection points.[16]

---

17. References

---

Publication Notes for aruksworld.com

Suggested pull quotes

"The Maillard chemistry that produces the flavour of a good steak is the same chemistry that produces heterocyclic amines. They cannot be fully decoupled — but the balance can be shifted dramatically."

"Five of thirteen Warri sampling stations had suya above the EU benzo[a]pyrene limit. Eight did not. That is not a statement about suya — it is a statement about how any given person makes it."

"Yaji is the best and the worst thing about suya: a centuries-old antioxidant system that suppresses heterocyclic amines, and the main route by which a Group 1 liver carcinogen reaches the plate."

"Processed meat and plutonium are both IARC Group 1. They are in the same category because the evidence that each can cause cancer is comparably strong — not because the magnitudes are remotely comparable."

Suggested section breaks for serialisation, if publishing as a series rather than a single long-form piece:
- Part 1: Sections 1–3 (What the evidence actually says)
- Part 2: Section 4 (The chemistry of meat carcinogenesis)
- Part 3: Sections 5–6 (Cooking method and the low-carcinogen steak protocol)
- Part 4: Sections 7–8 (Soya: what it is, and what has been measured in Nigeria and Burkina Faso)
- Part 5: Sections 9–10 (Yaji, aflatoxin, and the lower-risk soya protocol) — this is the strongest standalone piece and the most shareable
- Part 6: Sections 11–17 (Substitution, susceptibility, synthesis, limitations and references)

Suggested internal graphics
1. Reaction scheme: creatine → creatinine + Maillard radical intermediates → PhIP / MeIQx
2. Diagram: PAH formation from dripping fat, showing pyrolysis → smoke → surface deposition
3. Bar chart: HCA yield by cooking method
4. Bar chart: marinade reduction percentages from the Section 6.4 table
5. Photo sequence or illustration: direct-flame suya grilling vs barrier-plate grilling, annotated with the PAH pathway
6. Map or chart: benzo[a]pyrene levels across Nigerian sampling locations, showing which exceeded the EU limit
7. Comparison chart: apolar vs polar HAAs in flamed vs braised chicken (Burkina Faso data)
8. Infographic: how to sort groundnut for yaji — what to discard and why

A note on tone for a West African readership. This article should not read as an attack on suya, and it has been written carefully to avoid that. Suya is a culturally significant food, a livelihood for a large number of people, and — in its traditional serving form with onion, tomato and pepper — better designed than most Western barbecue. The findings here are about practice, not about the dish, and every recommendation is affordable. If the piece lands as "stop eating suya" it will be both wrong and ignored. If it lands as "put a plate between the fat and the fire, and buy your yaji from someone who stores it properly," it may actually change something.

Suggested calls to action. Consider ending with two concrete asks: (1) readers should ask their regular mai suya what fuel he uses and whether he grills over flames or settled coals; (2) readers should buy yaji in small quantities from a source with visible turnover, and sort their own groundnut if making it at home.