Acute myocardial infarction (AMI) is a leading cause of sudden death and poses diagnostic challenges in forensic pathology when gross or histopathological ischemic changes are subtle. Traditional postmortem AMI diagnosis relies on morphological findings, but biomarkers have been evaluated to improve diagnostic certainty. Periostin, a matricellular extracellular matrix protein involved in tissue remodeling and repair, has been implicated in cardiac remodeling after AMI and has emerged as a potential circulating biomarker in clinical studies. This study aimed to evaluate serum periostin levels in forensic autopsy cases and to provide baseline postmortem data assessing its utility for AMI diagnosis.
Eighty medicolegal autopsy cases performed between 01/04/2015 and 31/08/2025 were included and categorized as AMI (n = 46), fatal asthma (n = 16), or traumatic deaths (n = 18). AMI cases required identification of a culprit coronary artery with coronary thrombus or thin-cap fibroatheroma and histopathological evidence of myocardial ischemic changes across sampled myocardial regions (LVAW, LVLW, LVPW, IVS, RVW). Cases with only coronary stenosis without myocardial abnormalities (presumed fatal arrhythmia) were excluded. Asthma diagnosis relied on specific findings and premortem symptoms; traumatic deaths served as the reference group. Individuals with coronary atherosclerosis or MI were excluded from asthma and trauma groups. Cases with conditions likely to influence periostin (eg, atopic dermatitis, neoplasm, fibrotic systemic disease) were screened out.
Demographics: AMI 46 (6 women, 40 men), mean age 57.28 ± 17.18 years; asthma 16 (4/12), mean age 56.31 ± 16.08 years; trauma 18 (5/13), mean age 58.06 ± 22.51 years. Mean PM intervals were similar across groups.
Blood was collected from the right atrial appendage at autopsy. Samples with marked hemolysis were excluded. Blood cells were separated by immediate centrifugation, serum supernatants stored at −80 °C without repeated freeze–thaw cycles, and periostin measured by enzyme-linked immunosorbent assay (outsourced to SRL, Inc., Tokyo, Japan).
Data are reported as median and interquartile range (IQR) unless noted. Skewed variables, including periostin concentrations, were log-transformed for parametric testing when applicable. Normality was assessed by Shapiro–Wilk test. Group comparisons were performed with t-tests or one-way ANOVA with Bonferroni adjustment. Correlations used Pearson’s coefficient. Receiver operating characteristic (ROC) analysis assessed diagnostic utility; statistical significance thresholds were p < 0.05, < 0.01, and < 0.001 as reported.
Median serum periostin values were: AMI 728.50 ng/mL (IQR 465.25–1139.75), asthma 410.50 ng/mL (IQR 253.00–756.00), and trauma 386.50 ng/mL (IQR 229.75–532.50). One-way ANOVA with Bonferroni adjustment showed significantly higher periostin in AMI versus asthma (p = 0.03015) and AMI versus trauma (p = 0.00048). No significant difference was observed between asthma and trauma (p = 0.98659).
To evaluate postmortem stability, periostin in the trauma group (mean PM interval 34.50 ± 21.10 h; range 11–84 h) was correlated with PM interval. No significant correlation was identified (r = −0.058, p = 0.82), indicating minimal association between PM interval within the studied range and measured serum periostin.
Within the AMI cohort, periostin did not correlate significantly with age (r = 0.128, p = 0.398), body mass index (r = −0.0581, p = 0.701), or heart weight (r = −0.0793, p = 0.6). Median periostin values showed no statistically significant differences by sex (women vs men), history of diabetes mellitus, dyslipidemia, or hypertension when compared across groups as reported.
Serum periostin did not differ significantly by culprit coronary artery: LAD (n = 32) median 694.5 ng/mL (IQR 415.75–1191), LCx (n = 6) median 554 ng/mL (IQR 397–891), RCA (n = 8) median 1047 ng/mL (IQR 768.25–1087.75) (p = 0.97–1.00). No significant variation was found by primary infarct region (LVAW, LVLW, LVPW, IVS) or by presence versus absence of myocardial fibrotic scarring attributable to prior MI (re-MI).
ROC analysis using a cut-off of 556 ng/mL produced sensitivity 93.8%, specificity 67.4%, and area under the curve (AUC) 0.814 (95% CI 0.708–0.920). Reported diagnostic indices for that cut-off included positive predictive value 0.939, negative predictive value 0.516, diagnostic accuracy 0.734, positive likelihood ratio 6.065, and negative likelihood ratio 0.367.
The study provides baseline postmortem serum periostin concentrations and shows markedly higher periostin in AMI cases compared with fatal asthma and traumatic deaths. The lack of correlation with PM interval in the trauma group suggests periostin measurements may be reasonably stable across the PM intervals studied. The high sensitivity and AUC indicate potential utility of periostin in supporting postmortem AMI diagnosis, particularly when morphological findings are equivocal. Because periostin did not vary substantially with age, BMI, heart weight, sex, major cardiovascular risk factors, infarct location, or culprit artery in this dataset, it may offer a biomarker signal that is relatively independent of those variables in the sampled population.
Findings are limited to the cases and methods described: right atrial appendage blood sampling, exclusion of hemolyzed samples, ELISA measurement by an external laboratory, and PM intervals up to the reported range. External validation, broader PM intervals, and assessment of other potential confounders were not reported in the source. The authors conclude that serum periostin measurement may aid postmortem evaluation of AMI and provide reference data for forensic investigations.