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


Diabetes mellitus is a common cause and comorbid condition in patients with ESRD. Treatment of diabetes in these patients may necessitate administration of oral hypoglycemic agents. Concerns over safety and efficacy arise when these agents are administered because this population is at increased risk for adverse drug events.[2]

Overall, we found that thiazolidinedione therapy is safe and effective for patients receiving hemodialysis. After 3 months of therapy, A1C values declined by 0.61%; this finding is consistent with that reported for the general population.[15,16,17,18] Nonsignificant changes were observed in interdialytic weight, mean corpuscular volume, transferrin saturation, ferritin, and erythropoietin dose ( Table 1 ).

Although no changes in erythropoietin dose requirements were noted, rosiglitazone reduced hematocrit values; this was statistically, but not clinically, significant. This finding is perhaps due to intravascular volume expansion based on the significant increase in interdialytic weight gain observed in these patients.

A randomized, double-blind trial compared rosiglitazone with placebo in more than 950 patients over 26 weeks.[15] The frequency of edema was dose related; 5.2% of patients taking rosiglitazone 4 mg/day and 6.4% of those taking 8 mg/day experienced edema compared with only 1.6% of those in the placebo group. This study documented anemia as an adverse effect as well. Overall, hemoglobin level decreased an average of 0.5-0.9 g/dl, and hematocrit decreased an average of 1.6-2.5% in the rosiglitazone groups, a statistically significant change compared with the placebo group (p≤0.0001).

A second randomized, double-blind, placebo-controlled study of rosiglitazone evaluated 493 patients.[16] Edema was noted in 6.0% of patients taking rosiglitazone 2 mg twice/day, 10.7% of those taking rosiglitazone 4 mg twice/day, and 1.9% of those in the placebo group. Hemoglobin levels decreased 0.6 g/dl from baseline in the 2-mg rosiglitazone group and 1.0 g/dl in the 4-mg group; mean decreases in hematocrit were 0.8% (p=0.0001) and 2.1% (p<0.0001) in the 2- and 4-mg rosiglitazone groups, respectively.[16]

A third trial compared patients taking a sulfonylurea plus rosiglitazone with those taking a sulfonylurea plus placebo over 26 weeks.[17] The study involved 574 patients enrolled in 60 centers in Europe. In addition to a sulfonylurea, patients took rosiglitazone 1 mg, rosiglitazone 2 mg, or placebo twice/day. Both hemoglobin and hematocrit values decreased from baseline (p<0.0001) in both rosiglitazone treatment groups. Hemoglobin decreased by 0.39 and 0.66 g/dl and hematocrit by 1.52% and 2.34% in the rosiglitazone 1- and 2-mg groups, respectively.

In a 26-week randomized, double-blind trial of pioglitazone versus placebo, 408 patients received pioglitazone 7.5, 15, 30, or 45 mg/day or placebo.[18] Of those taking pioglitazone, 3.6% experienced edema versus none in the placebo group. All of the pioglitazone-treated groups experienced a dose-related decrease in hemoglobin and hematocrit values. The largest decreases occurred in the 45-mg treatment group, with decreases in hemoglobin and hematocrit of 0.74 g/dl and 1.3%, respectively.

We determined that thiazolidinedione administration was associated with reductions in both systolic and diastolic blood pressure 3 months after the start of therapy. This is consistent with the findings of a recent study that examined the effect of rosiglitazone on insulin resistance and blood pressure in 24 patients who had essential hypertension but not diabetes.[19] Two weeks after discontinuation of treatment with an antihypertensive, a euglycemic-hyperinsulinemic clamp was placed for glucose infusion, ambulatory blood pressure monitoring, and blood tests for cardiovascular risk factors. The patients were then given oral rosiglitazone 4 mg twice/day and their usual antihypertensive drugs (but not angiotensin- converting enzyme inhibitors) for 16 weeks, and baseline tests were repeated.

Significant decreases occurred in mean 24-hour blood pressures (systolic 138 ± 2 vs 134 ± 2 mm Hg, p<0.02; diastolic 85 ± 2 vs 80 ± 2 mm Hg, p<0.0001).[19] The decline in systolic blood pressure was correlated with the improvement in insulin sensitivity (r=0.59, p<0.005). The authors concluded that rosiglitazone treatment in patients who have hypertension but not diabetes improves insulin sensitivity and reduces systolic and diastolic blood pressure. Although the findings of this study[19] and our results demonstrate a statistically significant reduction in blood pressure, a change of 5-6-mm Hg may not be clinically significant. Additional research in this area is warranted to truly determine the clinical outcomes associated with this blood pressure reduction.

The retrospective nature of our study is a limitation. We had to assume that the electronic medical records were accurate. However, we felt confident in their accuracy given our results in a previous study investigating this issue.[20] Of 2709 drugs reviewed in that study, only 113 (4.1%) drug record discrepancies were identified.

We also assumed that patients adhered to their prescribed therapy. Unfortunately, we did not have refill records for the prescribed thiazolidinedione. Although we cannot guarantee patient compliance, the observed reduction in A1C suggests that this may be the case. Finally, we did not measure salt and fluid intake of the patients. However, we did use a surrogate marker -- interdialytic weight gain -- for salt and fluid intake. Given that interdialytic weight gain did not change from beginning to end of the study period, we can conclude that salt and fluid intake also did not change over that period. In addition, all of our study patients had anuria and thus could not excrete any additional fluid by way of the kidneys.