Hyperglycemia in Patients with Acute Myeloid Leukemia Is Associated with Increased Hospital Mortality.

Patients with acute myeloid leukemia (AML) who receive therapy for their disease are at significant risk for infections. One risk factor for infections is hyperglycemia which can often be observed in medically ill patients. In post-operative critically ill patients, intensive insulin therapy has been shown to reduce the incidence of nosocomial infections. (Van den Berghe, 2001) In one study of patients with acute lymphoblastic leukemia, hyperglycemia during induction therapy was associated with more infections and shorter disease free survival. (Weiser, 2004) We hypothesized that hyperglycemia is associated with worse outcomes in AML due to increased incidence and severity of infectious complications. We performed a retrospective cohort study to determine if there was an association between hyperglycemia and hospital mortality in hospitalized patients with AML. Four-hundred and sixty-four adult patients were identified through a search of our administrative databases as having the diagnosis of acute myeloid leukemia over a three year period. These individuals experienced 1,004 admissions to the hospital. The occurrence of increasing hyperglycemia, defined as the proportion of glucose assessments above normal (>110 mg/dL), during all hospital days was associated with increased odds of hospital mortality [OR 1.24, 95% CI 1.13–1.36, p<0.001) after adjusting for co-variates like age, gender, diabetes and other co-morbidities. In addition, the odds of developing sepsis [OR 1.08, 95% CI 1.02–1.14, p=0.007], severe sepsis [OR 1.16, 95% CI 1.08–1.26, p<0.001), respiratory [OR 1.23, 95% CI 1.07–1.40, p=0.003] or renal failure [OR 1.14, 95% CI 1.01–1.30, p=0.04] were increased with increasing hyperglycemia. Because duration of neutropenia is known to increase the risk of infectious complications, we included leukopenia (WBC<1000) as a possible mediator in the risk-adjusting models. Independently, leukopenia was associated with the development of severe sepsis, however inclusion of this co-variate did not appreciably change the relationship between hyperglycemia and sepsis [OR 1.07, 95% CI 1.01–1.16, p=0.01], severe sepsis [OR 1.18, 95% CI 1.09–1.28, p<0.001] or mortality [OR 1.24, 95% CI 1.12–1.35, p<0.001]. Finally, after adjusting for the diagnosis of severe sepsis, hyperglycemia remained significantly associated with mortality, but the point estimate was reduced (OR from 1.24 to 1.16) suggesting that the association between abnormal glucose regulation and hospital mortality may be partially mediated by the development of severe sepsis. Based on these findings we are performing a more formal analysis of the relationship between hyperglycemia and the morbidity and mortality associated with treatment of patients with AML in a prospective study.