Kafkas Üniversitesi Veteriner Fakültesi Dergisi Early View
Aflatoxin M1 Detection Frequency in Cheese: A Meta-analysis of Recovered Evidence with a Systematic Literature Update
Mustafa ATASEVER1, Rabia Nurselin KALAYCI1, Halil ÜZÜM1, Meryem AYDEMİR ATASEVER1
1Atatürk University, Faculty of Veterinary Medicine, Department of Food Hygiene and Technology, TR-25240 Erzurum - TÜRKİYE DOI : 10.9775/kvfd.2026.38111 Reported aflatoxin M1 (AFM₁) detection in cheese depends on contamination, sampling and analytical thresholds. We reanalyzed recovered historical evidence and integrated a systematic database update covering 1 August 2023 through 19 August 2026. The revised primary analysis combined one non-overlapping aggregate per study using logit random-effects meta-analysis with restricted maximum likelihood and study-level robust inference. Sensitivity analyses examined the influence of study inclusion and available quality appraisals, while subgroup and moderator analyses explored differences by analytical method and publication year. The primary set comprised 111 study clusters, 11.369 samples and 5.938 positives. Pooled detection frequency was 56.2% (95% CI 47.4-64.6%); the crude sample-weighted fraction was 52.2%. Heterogeneity was substantial (I²=95.4%; 95% prediction interval 3.6-97.8%). The secondary 271-stratum model gave 56.3% (95% CI 47.3-65.0%). Among 141 strata from 53 clusters with usable LOD information, the exploratory year-adjusted association per tenfold higher LOD was OR 0.393 (95% CI 0.154-1.004; P=0.051). A separate concentration synthesis included 12 studies. Historical selection decisions and some appraisals could not be recovered, and regional database coverage remained incomplete. These estimates describe the recovered evidence and should not be interpreted as universal prevalence or proof of complete global retrieval. Keywords : Aflatoxin M1, Cheese, Detection frequency, Evidence reconstruction, Limit of detection, Meta-analysis