The reliability of measuring mast cell tryptase in post-mortem blood to diagnose anaphylactic deaths has been questioned because moderate elevation of tryptase can be seen also in control cases. Very high tryptase concentrations have been recorded even in a few control cases with known cause of death such as myocardial infarction or trauma. Aiming to compare findings we measured tryptase in 193 cases: 176 with known cause of death, 10 unexplained deaths and seven anaphylactic or anaphylactoid deaths (AADs). Using binary logistic regression we calculated the sensitivity and specificity of the tryptase test at different cut-off values and found 10 micrograms/l to be optimal, the sensitivity being 86% and the specificity 88%. Traumatic deaths (n = 23), sudden infant death syndrome (n = 40) or deaths after heroin-injection (n = 22) showed elevated tryptase values in 35%, 35% and 32%, respectively, and in 40% of the unexplained deaths (n = 10), which was higher than expected (12%). Heart blood tryptase level was elevated in 22% of the controls and femoral blood tryptase in 10%. No correlations were seen with age or post-mortem delay. It is concluded that tryptase measurements are useful in confirming death from AAD, and that blood should be sampled from the femoral vessels. In unexplained deaths tryptase measurement is a useful indicator, but the diagnosis is not to be based on the test alone.