Dr. Watanakunakorn requests a chart review based on our study's identification of hospitalized patients treated with trimethoprim-sulfamethoxazole and suggests the generation of a daily patient list through pharmacy records. These patients' hospital charts were reviewed prospectively for the data noted in the Methods section. Laboratory data were not available in all patients on a daily basis, as noted by the "n" values provided in our figures. The phrase "when available" should follow "daily" in the Methods section. Eighty of 380 patients treated with trimethoprim-sulfamethoxazole were included in the study. As stated in the Methods section, patients were excluded on the basis of the following criteria: fewer than 5 days of therapy with trimethoprim-sulfamethoxazole, unstable or increasing potassium level, or unstable renal function before treatment with trimethoprim-sulfamethoxazole. Finally, only 25 controls, identified by pharmacy records to be receiving an antibiotic other than trimethoprim-sulfamethoxazole, were required for us to show both a clinically and statistically significant difference in the development of hyperkalemia when comparing the treatment and control groups. The control group was included mainly to ensure that serum potassium levels did not increase in hospitalized patients over time. Only a relatively small number of patients (n = 25) was required to prove this point.

Postelnick and colleagues take exception to our inclusion of nonsteroidal anti-inflammatory drugs and heparin as potassium-altering medications. We strongly disagree. These agents act through many mechanisms to impair renal potassium excretion^[1]. Suppression of aldosterone production probably plays the most important role^[1]. Heparin also impairs aldosterone synthesis, regardless of the dose administered^[2]. Because aldosterone is one of the most important regulatory hormones of distal nephron potassium secretion, excluding these medications from analysis seems inappropriate.

Dr. Thomas asks whether more severe hyperkalemia occurred in study patients treated with the combination of trimethoprim-sulfamethoxazole and an angiotensin-converting enzyme inhibitor. We share his concern, especially in elderly patients with multiple risk factors for hyperkalemia (such as renal failure or pyelonephritis). Analysis of our data, however, showed no further risk for hyperkalemia when each medication was considered separately, although our relatively small patient sample may be inadequate to fully rule out this possibility.

Dr. Parker appropriately notes the importance of volume and solute delivery to the distal nephron as a factor affecting renal potassium secretion. Unfortunately, our chart review cannot provide the data needed to adequately answer his question. Our matched control group, however, showed a decrease in serum potassium level, suggesting that trimethoprim-sulfamethoxazole was integral to the development of hyperkalemia, regardless of this factor.

Ougorets and colleagues propose that an endocrinologic evaluation should be done in patients who develop hyperkalemia related to trimethoprim-sulfamethoxazole therapy, given that future overt adrenal insufficiency may develop. Although this approach may be reasonable in high-risk patients, such as those infected with HIV^[3,4], we believe that this approach would not have a high yield or be cost-effective in most patients.