Adverse clinical consequences of hyperglycemia from total parenteral nutrition exposure during hematopoietic stem cell transplantation.

Immunocompromised hematopoietic stem cell transplant (HSCT) recipients frequently receive total parenteral nutrition (TPN), a dextrose-based solution that may exacerbate the infectious risks associated with hyperglycemia. This study assessed the incidence of hyperglycemia (glucose level>or=110 mg/dL) and its effect on clinical outcomes in TPN versus non-TPN recipients who received HSCTs. A retrospective cohort of 357 adults who were admitted for initial autologous or allogeneic transplantation at 2 university-affiliated centers was examined. To discern the temporality of outcomes, "before" and "after" comparisons were made by using actual infusion times for TPN patients and using timeframes based on mean hospital days before ("before") or during ("after") parenteral infusion for non-TPN patients. Patients demonstrated similar demographic and clinical characteristics when analyzed by institution, feeding, and donor-type strata, and 57% received TPN. After attempts to equilibrate disease acuity were employed, the proportion of hyperglycemic days was equivalent before but significantly greater after in patients exposed versus unexposed to TPN (87.5% versus 8.3%, respectively; P<.001). Using logistic regression, the likelihood of infection doubled (odds ratio [OR], 2.2; 95% confidence interval [CI], 1.4-3.5) after adjustment for donor type, diagnosis, age, gender, ethnicity, institution, mucositis, and obesity. This association was only slightly attenuated when patients with infections before were removed (OR, 1.9; 95%, CI, 1.1-3.3), steroid recipients were eliminated (OR, 2.1; 95% CI, 1.2-3.4), and when patients with nonablative regimens were excluded (OR, 2.1; 95% CI, 1.3-3.5), but was considerably higher for patients who were classified as normal or underweight (body mass index<or=25 kg/m2; n=118; OR, 4.3; 95% CI, 1.7-10.6). In addition, the effect of TPN became insignificant when glucose was added as an independent variable, thus symbolizing their collinear relation. Parenteral nutrition recipients versus nonrecipients also developed significantly greater requirements for red cell (P=.001) and platelet transfusions (P=.001) after and significant delays in granulocyte and platelet engraftment times for autologous (P=.01) and allogeneic (P=.02) subjects. The broad use of TPN in patients undergoing initial HSCT was associated with profound hyperglycemia, resultant greater morbidity, and questionable efficacy in this adult, well-nourished cohort.