In some forensic autopsies blood is not available, and other matrices are sampled for toxicological analysis. The aims of the present study were to examine whether heroin metabolites can be detected in different post-mortem matrices, and investigate whether analyses in other matrices can give useful information about concentrations in peripheral blood. Effects of ethanol on the metabolism and distribution of heroin metabolites were also investigated. We included 45 forensic autopsies where morphine was detected in peripheral blood, concomitantly with 6-acetylmorphine (6-AM) detected in any matrix. Samples were collected from peripheral blood, cardiac blood, pericardial fluid, psoas muscle, lateral vastus muscle, vitreous humor and urine. Opioid analysis included 6-AM, morphine, codeine, and morphine glucuronides. The 6-AM was most often detected in urine (n = 39) and vitreous humor (n = 38). The median morphine concentration ratio relative to peripheral blood was 1.3 (range 0-3.6) for cardiac blood, 1.4 (range 0.07-5.3) for pericardial fluid, 1.2 (range 0-19.2) for psoas muscle, 1.1 (range 0-1.7) for lateral vastus muscle and 0.4 (range 0.2-3.2) for vitreous humor. The number of 6-AM positive cases was significantly higher (P = 0.03) in the ethanol positive group (n = 6; 86%) compared to the ethanol negative group (n = 14; 37%) in peripheral blood. The distribution of heroin metabolites to the different matrices was not significantly different between the ethanol positive and the ethanol negative group. This study shows that toxicological analyses of several matrices could be useful in heroin-related deaths. Urine and vitreous humor are superior for detection of 6-AM, while concentrations of morphine could be assessed from peripheral or cardiac blood, pericardial fluid, psoas muscle and lateral vastus muscle.