Regarding Dr. Nishioka's first concern, we doubt that we substantially overestimated the importance of older age as a risk factor for death. Sudanese staff estimated the ages of patients and commonly considered factors in addition to physical appearance (for example, for a woman, the ages and number of her children and the timing of her marriage [that is, menarche] with respect to a political or military event).

In our multivariable analysis of the risk for death among adults, we avoided making assumptions about the relation between age and death by treating age as a series of independent binary variables instead of a continuous variable. Each age category was then independently compared with the referent group (patients 18 to 24 years of age); all adults 45 years of age and older were included in the same category. As shown in the second column of the Table 1, the approximate risk ratio for death progressively increases as age increases. This relation persists if we deduct an arbitrary 5 years from the ages of all adults who died (third column) or of all markedly malnourished adults (regardless of whether they lived) who had a body mass index less than 14 kg/m² (fourth column).

With respect to Dr. Nishioka's second concern, we wish we could have assessed leukocyte and neutrophil counts and included these data in our multivariable analysis of risk for death. Although anemia is not a perfect surrogate for neutropenia among patients with visceral leishmaniasis (hemoglobin values and leukocyte and neutrophil counts are sometimes differentially affected by leishmaniasis and comorbid illnesses), we agree that anemic patients are also typically neutropenic (and thrombocytopenic). Secondary infections (and complications caused by bleeding) are known to result in death among patients who have visceral leishmaniasis. Therefore, our patients were asked daily about various symptoms, and members of our staff were quick to provide antibiotic therapy.

Given the limitations of our clinical and laboratory evaluations, we intentionally avoided making deductions about pathophysiology and causality with respect to the adverse outcomes that we studied and the risk factors (such as anemia) that we identified. The extent to which severe anemia directly or indirectly contributes to illness and death is debatable. However, ascertainment of hemoglobin values, which can easily be done even in such settings as southern Sudan, provides clinically useful information about the severity of visceral leishmaniasis. Thus, it is a means of identifying patients who should be more carefully monitored.