Serum lactate dehydrogenase iso-enzyme 1 (S-LD-1) was determined in 24 patients with testicular germ cell tumours who died during follow-up. Serum samples were obtained at the start of chemotherapy or late into the clinical course for those without evidence of a tumour. Seven of the eight patients with tumour-associated death had S-LD-1 > 150 U l-1 (median 260 U l-1 range 90-905 U l-1, as did two of the six who died without evidence of a tumour (median 134 U l-1 range 89-128 U l-1) (P = 0.03, Mann-Whitney U-test, two-tailed). The remaining 10 patients had previously been reported as to prediction of S-LD-1 for the prognosis, and were evaluated only as to the reproducibility of the S-LD-1 determination. S-LD-1 was determined in two serum samples obtained concomitantly from eight of these 10 patients: the difference between the two S-LD-1 determinations was median 6% of the average of the two S-LD-1 determinations (25-75% range 3-17%). Our study showed an unforeseeable high level of imprecision of the assay system. Nevertheless, S-LD-1 determinations in the regular monitoring of patients with testicular germ cell tumours may be useful for predicting tumour-associated deaths.