Trimethoprim-Sulfamethoxazole Induces Reversible Hyperkalemia

Methods

The hospital records of all HIV-infected patients who were hospitalized and treated for P. carinii pneumonia at The Brookdale Hospital Medical Center from December 1989 to February 1991 were reviewed. Those patients who received Tmp-Smx either intravenously or orally at doses of 20 mg/kg per day (trimethoprim) and 100 mg/kg per day (sulfamethoxazole) (the accepted protocol for the treatment of P. carinii infection) for more than 6 days were studied. Trimethoprim-sulfamethoxazole was dispensed in generic forms. Patients who received potassium supplements or any medication known to alter potassium homeostasis and metabolism or renal function were excluded. Patients who had renal insufficiency, as defined by a serum creatinine level of 186 mol/L or more, and those with significant diarrhea or vomiting were also excluded from the study. All patients who met our inclusion criteria formed the study group. The records of all HIV-infected patients who were hospitalized during the same period but who did not receive Tmp-Smx served as the control group. These patients met the same inclusion and exclusion criteria as the study group. All patients received a normal hospital diet, containing 4 g of sodium and 3 to 3.5 g of potassium.

We recorded the age, sex, risk factors for HIV disease and other infections, and medications taken both before and during hospitalization for each patient who met the study criteria. The serum potassium concentrations during the first 2 weeks of Tmp-Smx therapy were recorded. In addition to serum potassium concentration, the blood urea nitrogen, serum creatinine, sodium, carbon dioxide content, cholesterol, creatine kinase, lactic dehydrogenase, albumin, arterial pH, hemoglobin, and weight were recorded when available. The same measurements were recorded in the control group as in the Tmp-Smx-treated group within 3 days of hospitalization and again as the averaged value observed on the ninth and tenth day of hospitalization (designated as the follow-up period). This interval was chosen because it represented the average period required for the peak serum potassium concentration to be observed in those patients treated with Tmp-Smx.

All results are expressed as a mean standard error or as a mean with a 95% confidence interval (CI). The statistical significance of the difference was determined by paired or unpaired t-test. A P value of less than 0.05 was considered statistically significant.

Results

All patients tested positive for HIV by Western blot assay. Twenty-six patients who were not treated with Tmp-Smx met our study criteria and formed the control group. This group ranged between 22 and 61 years old (mean SE, 36.6 1.6 years). Twenty patients were men and six were women. Sixteen were black, eight were Hispanic, and two were white. Risk factors for HIV in the control group included intravenous drug abuse in 18, homosexuality in 2, multiple sexual partners in 2, and no known risk factors in 4.

Twenty-five HIV-infected patients who received Tmp-Smx met our study criteria (treatment group). This group ranged between 22 and 52 years old (37.0 1.6 years). Twenty-one patients were men and four were women. Fourteen were black, 10 were Hispanic, and 1 was white. Risk factors for HIV included intravenous drug abuse in 15, homosexuality in 2, multiple sexual partners in 4, and no identifiable risk factors in 4. All 25 patients received Tmp-Smx at a dose of 20 mg/kg per day (trimethoprim) and 100 mg/kg per day (sulfamethoxazole) for more than 6 days for the treatment of P. carinii pneumonia. The diagnosis of P. carinii pneumonia was established by bronchoscopy in 18 patients and by clinical criteria in 7. The study drug was administered orally in 10 patients, intravenously in 5, and by both methods in the remaining 10 patients.

Table 1 summarizes the clinical data obtained within 3 days after hospitalization in the control and treatment groups. No statistically significant differences were noted with respect to demographic or laboratory variables, except that the serum lactic dehydrogenase level was significantly higher in patients treated with Tmp-Smx.

In the control group, no significant alterations were noted in any of the laboratory variables during the follow-up period. The serum potassium concentration during the follow-up period was 4.2 0.1 mmol/L, a value similar to the baseline value of 4.3 0.1 mmol/L. In the Tmp-Smx-treated group, the serum potassium level before therapy was 4.1 0.1 mmol/L and increased significantly by 1.1 mmol/L (CI, 0.8 to 1.5 mmol/L) to a peak level of 5.2 0.1 mmol/L 9.8 0.5 days after initiation of Tmp-Smx therapy (Figure 1). Of these 25 patients, 19 had an increase in their serum potassium concentration of 1.0 mmol/L or more from their baseline value. Seven patients had a serum potassium concentration greater than 5.5 mmol/L; three of these patients had a maximum potassium level greater than 6.0 mmol/L, and in one patient the serum potassium reached 7.2 mmol/L. Two of the seven hyperkalemic patients required kayexalate therapy at an average dose of 60 g to correct the hyperkalemia. Serial potassium measurements both during and after discontinuation of Tmp-Smx therapy were available in five patients in whom no potassium-lowering agents were administered. The serum potassium concentration increased progressively during the course of the Tmp-Smx therapy and returned to baseline after discontinuing therapy (Figure 2).

View larger version:

• In this window
• In a new window

Figure 1. Values in 25 patients were obtained at the initiation of therapy and at the time of the peak serum potassium level. The increase in the serum potassium level was statistically significant ( < 0.0001). Changes in serum potassium concentrations with trimethoprim-sulfamethoxazole therapy.P

View larger version:

• In this window
• In a new window

Figure 2. Values were obtained in five patients during therapy with trimethoprim-sulfamethoxazole (solid lines) and after discontinuation of therapy ( ). Therapy began on day 0. Serial changes in serum potassium concentrations.dotted lines

The laboratory results obtained within 1 day of initiating Tmp-Smx therapy and at the time of the peak recorded potassium concentration are summarized in Table 2. In the Tmp-Smx group, both the blood urea nitrogen and serum creatinine levels increased significantly, whereas the serum sodium and chloride concentrations decreased significantly. No other changes were noted during Tmp-Smx therapy; in particular, none of the patients manifested a significant change in carbon dioxide content, arterial blood pH, or body weight. In one patient, the renin-aldosterone system was assessed both before and after Tmp-Smx therapy. Before therapy, the serum potassium concentration was 4.6 mmol/L. The supine plasma renin activity and serum aldosterone level were 8.8 nmol/L per hour and 363.9 pmol/L and increased to 10.0 nmol/L per hour and 625.1 pmol/L, respectively, after the patient assumed an upright position for 2 hours. After 10 days of Tmp-Smx therapy, the serum potassium level increased to 5.7 mmol/L. The supine plasma renin activity and serum aldosterone level increased to 16.0 nmol/L per hour and 914.1 pmol/L and were 15.3 nmol/L per hour and 833.4 pmol/L, respectively, after the patient remained upright for 2 hours.

Discussion

Within the last several years, a spectrum of glomerular, tubulointerstitial, and electrolyte disorders associated with AIDS and with the treatment of HIV-related disorders have been reported [1, 2, 9]. We observed that serum potassium levels increased significantly in patients with AIDS who were treated with high-dose Tmp-Smx therapy; these changes were associated with a decrease in serum sodium concentration and mild elevations in blood urea nitrogen and serum creatinine levels, typically occurring 9 to 10 days after the initiation of therapy. The increase in serum potassium level was usually disproportionate to the degree of azotemia that developed, suggesting that the high-dose Tmp-Smx therapy altered potassium homeostasis. This hypothesis is supported by the temporal relation between the initiation of Tmp-Smx therapy and the progressive increase in the serum potassium concentration and a progressive decline in the serum potassium concentration after discontinuation of therapy (see Figure 2). A control group of HIV-infected patients who did not receive Tmp-Smx had no alteration in the serum potassium concentration during the same period. Finally, other factors that could cause this electrolyte abnormality in these hospitalized patients were not evident, including specific measurements of the renin-aldosterone system in one patient.

To date, Tmp-Smx has been associated with two major forms of nephrotoxicity. Acute interstitial nephritis after the use of Tmp-Smx has been reported [10-15], and trimethoprim alone has been shown to cause mild elevations in the serum creatinine level and decreases in creatinine clearance [16, 17]. Because glomerular filtration rate, as measured by renal clearance of Iodine-131-iothalamate, is not altered by trimethoprim, several authors have speculated that the effect of trimethoprim on creatinine clearance is mediated by the inhibition of tubular secretion of creatinine and not by a reduction in glomerular filtration [17]. Although large collaborative studies have shown a relatively low incidence of adverse reactions (including nephrotoxicity) to Tmp-Smx therapy (8%), none have reported hyperkalemia [18-20]. These studies, however, were done in patients receiving far lower doses of Tmp-Smx (trimethoprim, 320 mg/d; sulfamethoxazole, 1600 mg/d) for the treatment of urinary and upper respiratory tract infections. Indeed, the large doses of Tmp-Smx used for the treatment of P. carinii pneumonia in HIV-infected patients have been associated with a much higher incidence of adverse reactions (44% to 70%) than those noted in the previous studies using standard doses [21, 22]. In a recent report, hyperkalemia was observed in 20% of patients with AIDS who were taking Tmp-Smx for P. carinii pneumonia and is more frequent (53%) in those taking trimethoprim-dapsone, possibly resulting from inhibition of trimethoprim metabolism by dapsone and an increase in the drug levels [23].

Several causes of hyperkalemia can be eliminated as being responsible for the increase in the serum potassium seen with Tmp-Smx therapy. Acidemia, hypoinsulinemia, an increase in plasma osmolality, stimulation of the adrenergic nervous system, or blockade of the adrenergic nervous system all may result in transcellular shifts in potassium and subsequent hyperkalemia. Based on our data, these possibilities can be excluded. No significant alteration in either arterial pH or serum bicarbonate was noted in our patients between the initiation of Tmp-Smx therapy and the time at which the peak serum potassium concentration was observed. Insulin levels, when measured (two patients), were not depressed, nor did clinical or laboratory data suggest that a hypoinsulinemic state had developed. Finally, none of the patients studied was receiving agents known to alter either the or adrenergic nervous systems.

Exogenous or endogenous potassium loads, or both, may result from hemolysis, tissue necrosis, or hypercatabolic states and predispose to the development of hyperkalemia in patients with renal insufficiency. No patient reviewed, however, received an exogenous potassium load, nor showed evidence of an endogenous potassium load; hemoglobin, creatine kinase, and lactic dehydrogenase levels all remained unchanged from the baseline state. Similarly, none of the patients had pseudohyperkalemia associated with marked thrombocytosis or leukocytosis.

Hyperkalemia due to adrenal insufficiency or hyporenin-hypoaldosteronism has been previously described in patients with AIDS [3-5, 7]. None of our patients had clinical evidence of adrenal insufficiency, and, in two patients, the adrenal axis was intact when specifically examined. In addition, the electrolyte abnormalities resolved spontaneously after discontinuation of Tmp-Smx without requiring adrenal hormonal replacement. The renin-aldosterone axis was examined in a patient before and during Tmp-Smx therapy; the serum potassium concentration increased despite elevated baseline and stimulated plasma renin and aldosterone levels. Murphy and colleagues [8] have described an HIV-infected patient in whom both hyperkalemia and hyponatremia developed 9 days after initiation of Tmp-Smx therapy for P. carinii pneumonia. The renin-aldosterone system in this patient was also found to be normal, and the serum potassium concentration returned to normal after discontinuation of Tmp-Smx therapy as in our patients. Similarly, another report has described Tmp-Smx-induced hyperkalemia with an intact renin-aldosterone system in a patient with AIDS in whom 9 -fluorocortisol failed to increase urinary potassium excretion [24]. Thus, it would appear that neither adrenal insufficiency nor hyporenin-hypoaldosteronism is likely to account for the increased potassium observed with Tmp-Smx therapy.

Acute interstitial nephritis with acute renal failure has been reported to be associated with Tmp-Smx therapy; however, it is unlikely that this factor explains the increased serum potassium concentrations, given that none of the usual features of this syndrome were observed in our patients. Furthermore, the mild elevation in the serum creatinine level and clinical course of these patients are not consistent with the acute interstitial nephritis usually associated with Tmp-Smx therapy.

The results of our study do not permit us to unravel the mechanisms whereby Tmp-Smx increases serum potassium concentrations. Recent in vitro and in vivo studies, however, suggest that trimethoprim inhibits renal potassium secretion in a dose-dependent fashion. Specifically, trimethoprim reversibly inhibits sodium channels in the cultured A6 epithelial cell line [24] and impairs potassium secretion in the distal tubule by blocking luminal sodium channels, as shown by a micropuncture study in the rat [25]. As a consequence of impairment of sodium transport in the distal tubule, the lumen-negative potential difference generated by sodium transport would be reduced, thus decreasing the electrical gradient favorable for passive potassium secretion. Thus, trimethoprim may inhibit potassium secretion in a fashion similar to two structurally related cationic diuretics: amiloride and triamterene. Prospective studies are needed, however, to assess the effects of Tmp-Smx on distal tubular function in humans.

In conclusion, therapy with high-dose Tmp-Smx may lead to life-threatening hyperkalemia, particularly 7 to 10 days after the initiation of therapy. The pathogenesis of this electrolyte disorder may be related to inhibition of sodium channels and impaired urinary potassium secretion in the distal nephron.