Early mortality in multiple myeloma (MM) is usually attributed to combined effects of active disease and comorbid factors. We have studied early deaths in a series of large multicenter trials to assess direct causes of death, their predictability, and whether current management strategies have reduced their frequency. A total of 3,107 newly diagnosed patients entered onto United Kingdom Medical Research Council MM trials from 1980 to 2002 were studied. Trial files, final clinical summaries, and postmortem reports were analyzed. Death within 60 days of trial entry occurred in 299 patients (10%). Logistic regression modeling identified beta 2-microglobulin, performance status, and age as the most important predictors of early death, but only with 61% sensitivity and 73% specificity. Forty-five percent of deaths were attributable to infection, which was often associated with bone pain (particularly thoracic pain) and delay in presenting to medical care. Neutropenia was present at diagnosis in only 11 of the 135 deaths from infection. Renal failure was present in 28% of early deaths and was linked to light-chain MM, hypercalcemia, dehydration, and nonsteroidal anti-inflammatory drugs. There was no time related reduction in the percentage or nature of early deaths in 1,550 patients older than 65 years receiving similar therapy between 1982 and 2002. A tenth of patients die within 60 days of diagnosis of MM. Infection and renal failure are the main direct causes of early mortality, which cannot be accurately predicted by presenting prognostic features. All patients should be considered at high risk of death during induction therapy.
