Methods

Top Methods Results Discussion Author & Article Info References

Study in Humans

All patients receiving high-dose trimethoprim (20 mg/kg per day), with either sulfamethoxazole or dapsone, at the Yale-New Haven Hospital during a 4-month period, were included. As a part of the clinical management of these patients, serum measurements of sodium, potassium, and creatinine levels were recorded before, during, and after trimethoprim treatment. In some patients, increased serum potassium levels (>5.0 mmol/L) were identified while patients were receiving trimethoprim. In a group of these patients, we reviewed the clinical course and recommended further evaluation to search for causes of hyperkalemia. This evaluation included measurements of serum glucose, renin, aldosterone, and cortisol levels as well as osmolality; measurements of urinary sodium, potassium, chloride, glucose, and protein levels as well as osmolality; and examination of the urinary sediment. The tubule fluid/plasma concentration ratio for potassium in the cortical collecting duct (transtubular potassium gradient) [11, 12] was calculated from urine and serum values for potassium and osmolality as (K)_urine/[K]_serum)/([Osm]_urine/[Osm]_serum). The ability of cosyntropin to stimulate cortisol secretion was determined in patients whose cortisol level was < 552 nmol/L (<20 µg/dL).

Studies in Rats

Male Sprague-Dawley rats, allowed free access to standard rat chow and tap water up to the time of the experiment, were anesthetized before surgical exposure of one kidney, as previously described [13]. The ureter was cannulated.

Intravenous Infusion of Trimethoprim

A salt solution (140 mmol/L sodium chloride, 4 mmol/L potassium chloride) was infused at 15 mL/h per kg body weight, and 45 minutes was allowed for equilibration after surgical preparation was completed. Subsequently, six 30-minute urine collections were obtained. The first 90 minutes was period I (collections 1 to 3), and the subsequent 90 minutes was period II (collections 4 to 6). The control and experimental groups of animals differed only in that at 90 minutes (after period I), 0.64 g/L of trimethoprim was added to the intravenous infusate (the trimethoprim infusion rate was 9.6 mg/h per kg body weight) in the experimental group and was maintained throughout period II. Sodium, potassium, and chloride concentrations were measured in urine [13]. The urine flow rate was measured, and the excretion rates for fluid, sodium, potassium, and chloride were calculated.

Microperfusion of Distal Tubules

Microperfusion experiments were done as described previously [13, 14]. Distal tubules were perfused with an artificial tubule fluid with or without trimethoprim (the composition of the perfusion solutions is given in Table 1. A perfusion pipette was positioned at the upstream end of the tubule, and a collection pipette was positioned at the downstream end. The perfusion pump was set to deliver 15 nL/min. A paired design was used as follows: After the first tubule fluid sample was collected, both the collection pipette and the perfusion pipette were removed: Then a second perfusion pipette containing a different solution [the order of solutions was alternated] and a second collection pipette were used to collect a second tubule fluid sample. The volume of collected samples was measured. Sodium, potassium, chloride, and inulin concentrations in perfused and collected fluids were measured, as described previously [14]. Osmolality and pH of bulk solutions were measured [14]. The perfusion rate was calculated from the collection rate and the inulin concentrations. Net fluid transport was calculated as the difference between perfusion and collection rates. Net transport rates for sodium, potassium, and chloride levels by each distal tubule were determined. Transepithelial voltage across the wall of the late distal tubule was measured, as previously described [15]. In experiments designed to test the effect of different concentrations of trimethoprim on transepithelial voltage, a higher perfusion rate (30 compared with 15 nL/min) was used to minimize changes in luminal ion composition along the length of the perfused tubule.

Statistical Analysis

Results were analyzed using the t-statistic. A P value of less than 0.05 (95% CI) was statistically different.

Results

Top Methods Results Discussion Author & Article Info References

Human Studies

Records from 30 consecutive inpatients who were treated with high-dose trimethoprim at Yale-New Haven Hospital between 10 July and 18 November 1992 were reviewed. No patients were excluded from this study. All patients were HIV positive and were treated for presumed or confirmed P. carinii pneumonia. Twenty-three of the patients were treated with trimethoprim-sulfamethoxazole, and seven were treated with trimethoprim-dapsone. On average, the length of the trimethoprim treatment period was 5.3 ± 2.79 days (mean ±SD; range, 1 to 13 days). Figure 1 shows that the serum potassium concentration increased by 0.6 mmol/L (CI, 0.29 to 0.95 mmol/L) during treatment with trimethoprim. When the drug was discontinued, the potassium concentration decreased to pretreatment values. In 15 of 30 patients (50%), the serum potassium concentration was more than 5.0 mmol/L on at least 1 day during trimethoprim treatment. Severe and potentially life-threatening hyperkalemia (potassium > 6.0 mmol/L) occurred in three patients (10%). The serum creatinine concentration was slightly higher during trimethoprim treatment than during recovery (mean difference, 15.9 mmol/L; CI, 5.3 to 26.5 mmol/L [0.18 mg/dL; CI, 0.06 to 0.30 mg/dL]). None of the patients were taking nonsteroidal anti-inflammatory drugs, converting-enzyme inhibitors, or potassium-sparing diuretics.

View larger version (28K): [in this window] [in a new window]   Figure 1. Effect of trimethoprim on serum potassium concentration in patients with AIDS. Values are for individual patients, averaged for available potassium measurements in the 4 weeks before the initiation of therapy (Before), throughout the period (1 to 13 days) of therapy (During), and during the 2 weeks after trimethoprim was discontinued (After). Lines connect values in individual patients. Missing lines indicate that the data were not collected during that period. Mean values for all patients ± SE for each period are 4.2 ± 0.09 mmol/L, 4.8 ± 0.14 mmol/L, and 4.3 ± 0.09 mmol/L.

Renal and adrenal function were evaluated in seven patients during hyperkalemia (potassium > 5 mmol/L). The mean serum potassium concentration was 5.9 ± 0.9 mmol/L, and the urinary potassium concentration was 11.3 ± 5.8 mmol/L (mean ±SD). Oliguria was not present in any of the patients, and the serum creatinine concentration was not increased above baseline (mean difference, 17.7 mmol/L; CI, 9.72 to 43.3 mmol/L [0.2 mg/dL; CI, –0.11 to 0.49 mg/dL]). The transtubular potassium gradient calculated for the cortical collecting duct [11, 12] was 1.9 ± 1.1 (mean ±SD). This value was low (expected range was 6 to 11) for a plasma potassium concentration of >5 mmol/L. In three patients, the transtubular potassium gradient was calculated both during and after treatment with trimethoprim. After discontinuation of trimethoprim, the transtubular potassium gradient increased to normal values in all three patients (mean difference, 4.5; CI, 1.4 to 7.5). The urinary sodium concentration was 103 ± 65.7 mmol/L, and there was mild hyponatremia (132 ± 2.8 mmol/L). The plasma cortisol (497 ± 152 nmol/L, supine [18.0 ± 5.5 µg/dL, supine]); renin (0.667 ± 0.25 ng/[L x s], supine [2.4 ± 0.9 ng/mL per hour, supine]); and aldosterone (535 ± 264 pmol/L, supine [19.3 ± 9.5 ng/dL, supine]) levels were all high normal or increased during hyperkalemia. In two patients with borderline serum cortisol levels (221 and 469 nmol/L [8 and 17 µg/dL]), cosyntropin stimulation test results were normal. Glucose levels were all within the normal range.

Rat Studies

The effect of trimethoprim on potassium and sodium excretion rates in the whole kidney is shown in Figure 2. Compared with the control group, trimethoprim decreased potassium excretion by 572 nmol/min (CI, 299 to 845 nmol/min). The reduction in potassium excretion during trimethoprim infusion was 40% (CI, 21% to 60%) of the control rate measured in period II. Although sodium excretion in the control group increased with time, trimethoprim caused a twofold larger increase in sodium excretion. The difference between these changes, 1192 nmol/min (CI, 240 to 2142), was statistically significant. The increase in sodium excretion during trimethoprim infusion was 46% (CI, 9% to 83%) of the control rate measured in period II. There was no effect of trimethoprim on urine flow or chloride excretion rate (Appendix 1).

View larger version (23K): [in this window] [in a new window]   Figure 2. The scales depict the change () in ion excretion rate with time (period II minus period I) in control and experimental (Trimethoprim) animals. Bars and vertical lines indicate means and 95% CIs. Panel A. Trimethoprim statistically inhibited renal potassium excretion when compared with control animals. The absolute potassium excretion rates that were used to calculate these values were: for the control animals in periods I and II, 1369 nmol/min (CI, 1173 to 1565 nmol/min) and 1415 nmol/min (CI, 1222 to 1608 nmol/min); for the trimethoprim animals: 1334 nmol/min (CI, 1093 to 1575 nmol/min) and 807 nmol/min (CI, 565 to 1049 nmol/min). Panel B. Trimethoprim statistically increased renal sodium excretion when compared with control animals. The absolute sodium excretion rates that were used to calculate these values were: for the control animals in periods I and II, 1450 nmol/min (CI, 800 to 2100 nmol/min) and 2592 nmol/min (CI, 1703 to 3481 nmol/min); for the trimethoprim animals, 1491 nmol/min (CI, 956 to 2026 nmol/min) and 3824 nmol/min (CI, 2627 to 5021 nmol/min).

Table 1 gives flow rates, lumen ion concentrations, and ion transport rates from in-vivo microperfusion of distal tubules of rats. Figure 3 shows that 1 mmol/L of trimethoprim inhibited net potassium secretion by 59% (CI, 26% to 92%). The rate of net sodium absorption did not decrease. Rates of net chloride and water absorption were also not affected.

View larger version (23K): [in this window] [in a new window]   Figure 3. Net potassium transport during perfusion of 14 distal tubules with control and trimethoprim (TMP) solutions. Lines connect measurements in the same tubule. Black circles and vertical lines indicate means and CIs. Positive values indicate absorption; negative values indicate secretion.

Figure 4 shows the effect of trimethoprim concentration on the transepithelial voltage in the distal tubule when perfused at a constant high flow rate of 30 nL/min. Maximum inhibition of transepithelial voltage occurred at approximately 1 mmol/L of trimethoprim. In paired experiments at this dose, trimethoprim inhibited the voltage by 66% (CI, 44% to 85%). The concentration of trimethoprim needed to inhibit the transepithelial voltage by 50% is approximately 0.1 mmol/L.

View larger version (14K): [in this window] [in a new window]   Figure 4. Dose-response curve for trimethoprim effect on transepithelial voltage in the rat distal tubule. Values are means and CIs. Paired tests of the effect of four concentrations of trimethoprim versus the control solution (no trimethoprim) were done. The mean voltage measured with the control solution (lumen versus blood) ranged from –31.6 to –37.2 mV. The effect of each dose of trimethoprim on the distal tubule voltage is plotted as percentage change compared with its own control voltage. The maximum effect of luminal trimethoprim was reached at approximately 1 mM.

Discussion

Top Methods Results Discussion Author & Article Info References

Hyperkalemia has been reported in 16% of hospitalized patients with AIDS [16] and has been attributed to adrenal insufficiency, hyporeninemic hypoaldosteronism, or advanced renal failure [1-4]. An increased plasma potassium level may also result from pharmacologic agents used to treat these patients. Although hyperkalemia is not generally recognized as a complication of trimethoprim therapy, disorders of potassium metabolism have been described in patients receiving high-dose trimethoprim (20 mg/kg per day). Four patients with AIDS described by Kalin and coworkers [4] developed mild-to-moderate hyperkalemia that was ascribed to the hyporeninemic hypoaldosterone state. However, these patients were also receiving trimethoprim at the time they developed an increased plasma potassium level. It is possible that trimethoprim could have caused or at least contributed to the hyperkalemia. In a prospective trial of therapeutic efficacy, 20% of patients treated with trimethoprim-sulfamethoxazole and 53% of patients treated with trimethoprim-dapsone developed mild hyperkalemia [10].

Our results show that patients with AIDS frequently have an increase in plasma potassium concentration when treated with trimethoprim (combined with either dapsone or sulfamethoxazole), at the dose recommended for P. carinii pneumonia (20 mg/kg per day). In some patients (10% of 30 consecutive patients treated with trimethoprim), hyperkalemia was severe (potassium > 6 mmol/L). A similar incidence was reported by Greenberg and colleagues [17], who found that serum potassium concentration was more than 6 mmol/L in 3 of 25 (12%) consecutive patients treated with trimethoprim-sulfamethoxazole, despite normal adrenal function and normal glomerular filtration rate. Patients with AIDS may be at a greater risk for developing hyperkalemia during trimethoprim treatment because they have a coexisting but clinically asymptomatic defect in renal or adrenal function. Merenich and colleagues [18] recently studied 40 HIV-positive asymptomatic patients who had decreased baseline cortisol and aldosterone levels as well as impaired cortisol response to adrenocorticotropic hormone infusion when compared with normal controls. The authors suggested there might be a subtle defect in adrenal function underlying the abnormal response. It is conceivable that the superimposition of trimethoprim therapy on these patients with AIDS could precipitate hyperkalemia that would not have occurred otherwise.

Although many factors may contribute to hyperkalemia in patients with AIDS, our results indicate that trimethoprim (a dihydrofolate reductase inhibitor), by itself, inhibits renal potassium excretion and may participate importantly in this disturbance. Trimethoprim and amiloride are both organic cations, and the similarity of their actions may relate to their common structure [19]. Fonseca and coworkers [20] reported that trimethoprim inhibits sodium transport across frog skin and competes with amiloride for blockade of conductive sodium channels. More recently, trimethoprim has also been shown to inhibit amiloride-sensitive short circuit current in the epithelial cell line A6 [19]. Our results indicate that trimethoprim blocks sodium channels in the distal nephron of the mammalian kidney. When trimethoprim was perfused directly into the lumen of distal tubules of rats, potassium secretion decreased by 59% (see Table 1, Figure 3, and the lumen negative transepithelial voltage decreased by 66% (Figure 4). These results are similar to results obtained when the diuretic amiloride is perfused into the distal tubule [21, 22]. In the distal tubule, amiloride inhibits potassium secretion indirectly by blocking luminal sodium channels. Because the lumen negative voltage depends on electrogenic sodium transport in this nephron segment, blockade of this pathway depolarizes the transepithelial voltage. In contrast, when potassium channels in the luminal membrane are blocked directly, as with barium (a divalent cation), potassium secretion decreases, but the magnitude of the transepithelial voltage increases [23]. Thus, the presence of both inhibited potassium secretion and depolarization of the transepithelial voltage during luminal application of trimethoprim indicates that the drug blocks sodium channels in the luminal membrane of the mammalian distal tubule.

If trimethoprim acts as a sodium channel blocker to reduce potassium secretion by decreasing the electrical driving force favoring potassium secretion, a decrease in net sodium transport would be predicted. Trimethoprim statistically increased renal sodium excretion compared with control Figure 2, a result consistent with this mechanism of action. In perfused distal tubules, however, net sodium absorption was not statistically decreased by trimethoprim. It may be difficult to show inhibition of sodium transport by trimethoprim in perfused distal tubules because most of the sodium absorption by this nephron segment traverses pathways that are insensitive to amiloride [22]. A change in net sodium flux through an amiloride-sensitive pathway can be masked by variations in salt absorption through other transport pathways. In contrast, sodium absorption downstream along the cortical collecting duct is mediated predominantly by an amiloride-sensitive pathway. Effects of trimethoprim on renal sodium transport along the collecting duct may lead to natriuresis and salt wasting. Treatment with trimethoprim-sulfamethoxazole has been associated with hyponatremia in previous reports [5], but this effect was attributed to the free water infused during drug administration rather than to a direct effect of the drug. Hyponatremia may occur in more than 50% of patients with AIDS and may be associated with salt wasting [3]. Although many factors including inappropriate release of antidiuretic hormone, adrenocortical insufficiency, and renal tubular dysfunction contribute to hyponatremia and salt wasting, our results suggest that trimethoprim may also cause salt wasting and hyponatremia. Even when trimethoprim itself does not lead to disorders of sodium or fluid balance, the administration of this dihydrofolate reductase inhibitor may exacerbate or unmask preexisting abnormalities.

Previous results indicating a higher incidence of hyperkalemia in patients treated with trimethoprim-dapsone were interpreted [5] as indicating that dapsone alters potassium metabolism. Other results, however, suggest an alternative explanation. Plasma trimethoprim concentrations are 48% higher in patients receiving trimethoprim-dapsone compared with those receiving trimethoprim-sulfamethoxazole and are thought to be related to dapsone-induced decreases of trimethoprim metabolism by the liver [10, 24]. The higher plasma trimethoprim levels would increase trimethoprim concentration in tubule fluid and could lead to a more complete inhibition of renal potassium excretion.

Our data indicate that high-dose trimethoprim (20 mg/kg per day) impairs renal potassium excretion and leads to an increase in serum potassium concentration. The effect of lower doses on potassium metabolism was not studied. However, even when the serum potassium concentration is not more than 5 mmol/L, it appears from Figure 1 that in nearly all patients the serum potassium concentration was greater during trimethoprim treatment compared with the pretreatment value. After a single oral dose of 200 mg trimethoprim, the urinary concentration of trimethoprim may reach 320 µg/mL (1.1 mmol/L) [25]. Figure 4 shows, however, that a 10-fold lower dose of trimethoprim (0.1 mmol/L) in the lumen of the distal tubule is sufficient to inhibit the transepithelial voltage by about one half of the maximal level. Although trimethoprim concentrations in tubule fluid during treatment with standard doses (160 mg trimethoprim twice daily) may not cause frank hyperkalemia in otherwise normal adults, such treatment may predispose to hyperkalemia when trimethoprim is administered to patients with underlying disorders of potassium metabolism, such as hyporeninemic hypoaldosteronism.

Portions of this work were presented at the 25th meeting of the American Society of Nephrology and were published in abstract form (J Am Soc Nephrology. 1992; 3:322).