Thiazolidinedione Safety and Efficacy in Ambulatory Patients Receiving Hemodialysis

Harold J. Manley, Pharm.D., Nicole M. Allcock, Pharm.D.


Pharmacotherapy. 2003;23(7) 

In This Article


This study was a retrospective chart review of patients receiving hemodialysis at our ambulatory hemodialysis clinic (Dialysis Clinic, Inc., Kansas City, MO) who were prescribed either rosiglitazone or pioglitazone from April 2001-May 2002. The adult health sciences investigational review board of the University of Missouri-Kansas City approved the protocol and data collection.

All patients were under the care of the same group of nephrologists. All patient information was obtained from medical charts and the electronic medical information system. Demographic data collected were age, sex, cause and start date of ESRD, comorbid conditions, drug profile, hospitalization dates, and reason for admission. Laboratory values were obtained for hematocrit, iron indexes (transferrin saturation and ferritin), mean corpuscular volume, and A1C; body weight before and after dialysis, and predialysis systolic and diastolic blood pressures also were measured.

Thiazolidinedione-induced anemia was defined as increases in mean erythropoietin dose requirements after controlling for iron deficiency (transferrin saturation, mean corpuscular volume, and serum ferritin) and blood loss. Thiazolidinedione-induced hypertension was defined as a clinically significant increase in blood pressure of ≥ 5 mm Hg or more from baseline after controlling for any change in antihypertensive drug regimen or target dry weight. Hospitalizations for new or worsening chronic heart failure were recorded. Thiazolidinedione-induced changes in water balance were determined by comparing average interdialytic weight gain before and after the start of thiazolidinedione therapy after any change in target dry weight. Change in A1C values were reviewed to evaluate the efficacy of rosiglitazone and pioglitazone after controlling for any change in insulin or any other oral hypoglycemic drug regimen. All monitoring parameters of thiazolidinedione effects were evaluated for 3 months before and after thiazolidinedione therapy was begun.

Patient data were separated into three groups: rosiglitazone therapy only, pioglitazone therapy only, and the composite of all patients who received either rosiglitazone or pioglitazone therapy. Continuous variables (age, duration of ESRD, weight before and after treatment, hematocrit, erythropoietin dose, ferritin, transferrin saturation, mean corpuscular volume, and A1C) were expressed as mean ± SD. Discrete variables (sex, number of hospitalizations for chronic heart failure) were expressed as counts. Discrete and continuous variables were compared by using 2 analysis and paired Student's t test, respectively. Between-group comparisons in weight before and after treatment, hematocrit, erythropoietin dose, A1C, and number of hospital admissions were evaluated using analysis of variance. All statistical tests were two-sided; an a priori level less than 0.05 was considered significant. Analyses were conducted using SPSS software, version 10.0.7 (SPSS Inc., Chicago, IL).