We agree with Seligman and Gross that the wasting syndrome is an important clinical condition among HIV-infected persons and that death certificate data have limitations that hinder analysis of deaths associated with the wasting syndrome. One cannot ascertain from death certificate data alone the diagnostic criteria for the "HIV wasting syndrome" as described in the surveillance case definition for acquired immunodeficiency syndrome: "findings of ... weight loss of more than 10% of baseline body weight plus either chronic diarrhea ... or chronic weakness and documented fever" [1]. Such nonspecific signs and symptoms are infrequently listed on death certificates. In addition, it is difficult to know whether severe weight loss resulted from HIV infection directly, from an HIV-related opportunistic infection or cancer, or from a condition unrelated to HIV.

Despite these limitations, it may be useful to examine the percentage of documented HIV-related deaths associated with severe weight loss. Using the same methods as those in our analysis of infectious diseases and cancers among persons dying of HIV infection [2], we examined trends in wasting as defined by one or more of the codes 261, 263.9, 783.2, and 799.4 in the International Classification of Diseases, Ninth Revision. These codes represent such entities as nutritional marasmus, severe calorie deficiency, severe malnutrition not otherwise specified, unspecified protein-calorie malnutrition, abnormal weight loss, and cachexia. The percentage of documented HIV-related deaths associated with any of these conditions has more than doubled from 3.9% (394 of 10 001) in 1987 to 10.8% (2618 of 24 230) in 1992. This trend did not vary by sex or race. The increase in the prevalence of severe weight loss preceding HIV-related deaths may be at least partly due to reductions in the percentage of HIV-related deaths caused by pneumocytosis and cryptococcosis [2]. Deaths from these acute infectious diseases were negatively associated with wasting. Between 1987 and 1992, wasting was reported among only 3.8% of HIV-related deaths associated with pneumocystosis, 4.0% of those associated with cryptococcosis, 9.4% of those not related to pneumocystosis, and 8.5% of those not related to cryptococcosis. Prevention of death from these infectious diseases has probably prolonged survival among HIV-infected persons and resulted in more of them dying of diseases more commonly associated with wasting, such as nontuberculous mycobacteriosis (for example, Mycobacterium avium complex). Wasting was reported among 12.3% of HIV-related deaths in persons with nontuberculous mycobacteriosis compared with 7.8% of deaths in persons without mycobacteriosis.