Trimethoprim-sulfamethoxazole is considered the drug of choice for the prevention of P. carinii pneumonia [1, 4, 5], and it seems to be effective in the primary prophylaxis of toxoplasmosis [6-10]. However, trimethoprim-sulfamethoxazole is not well tolerated; rates of intolerance as high as 64% have been seen [11, 12]. Although a daily regimen has been recommended by the Centers for Disease Control and Prevention [4], several investigators prefer to administer intermittent regimens to minimize adverse effects and to improve compliance and quality of life [6, 7, 13-15]. Therefore, the best trimethoprim-sulfamethoxazole regimen has not yet been defined.

Dapsone and pyrimethamine are active against both P. carinii and Toxoplasma gondii [16-18]. Dapsone alone or combined with pyrimethamine appeared to be a good alternative for the primary prophylaxis of P. carinii pneumonia and toxoplasmosis when assessed in small studies [19, 20] and when different regimens (between 100 and 350 mg weekly of dapsone alone or combined with pyrimethamine, 50 to 75 mg weekly) were compared with inhaled pentamidine [21-23]. However, daily dapsone doses of 50 to 100 mg have a high toxicity rate [12, 21]. Several intermittent regimens (100 mg of dapsone plus 25 mg of pyrimethamine given once, twice, or three times weekly) either failed to prevent P. carinii pneumonia in controlled studies [6, 24, 25] or proved effective but were associated with considerable toxicity (200 mg of dapsone plus 75 mg of pyrimethamine once weekly) [23]. Nevertheless, the prolonged half-life of dapsone and the results of in vitro and in vivo studies suggest a potential efficacy for intermittent dapsone-pyrimethamine regimens in the prophylaxis of P. carinii pneumonia and toxoplasmosis [18, 26].

After failure of a low-dose, once-weekly regimen of dapsone-pyrimethamine in preventing P. carinii pneumonia [6], we compared thrice-weekly trimethoprim-sulfamethoxazole with twice-weekly dapsone-pyrimethamine for the simultaneous primary prophylaxis of P. carinii pneumonia and toxoplasmosis in patients infected with HIV who were at risk for these other infections. The trimethoprim-sulfamethoxazole regimen was identical to that used in our previous trial [6]. The doses of dapsone-pyrimethamine were chosen on the basis of data from a noncontrolled study that suggested the efficacy of a regimen of 100 mg of dapsone combined with 25 mg of pyrimethamine administered twice a week [20].

Methods

Top Methods Results Discussion Author & Article Info References

Between April 1992 and July 1994, we enrolled 230 patients at the HIV outpatient clinic of our infectious disease service in a 1000-bed university teaching hospital in the Barcelona area. Patients were selected for inclusion in our study if they had HIV infection and if their CD4 cell counts were less than 200 x 10⁶/L. Selected patients gave written informed consent.

Patients were excluded from the study if they had a history of P. carinii pneumonia or toxoplasmosis; had a history of allergy to sulfonamides or sulfones; were currently receiving other drugs with known anti-P. carinii or antitoxoplasma activity (such as clindamycin or clarithromycin); or had hemoglobin levels less than 90 g/L, neutrophil counts of less than 1.0 x 10 (⁹/L), alanine aminotransferase levels more than five times normal, and creatinine levels more than twice normal.

Treatment and Follow-up

The study was designed as a nonblinded randomized trial. Patients were randomly assigned [using a random-number table] to receive either 1) 160 mg of trimethoprim and 800 mg of sulfamethoxazole [cotrimoxazole] twice a day on Mondays, Wednesdays, and Fridays or 2) dapsone (100 mg) plus pyrimethamine (50 mg) twice weekly (on Tuesdays and Fridays): Both regimens were given orally. Patients who had adverse effects could switch to the other study drug or to pentamidine (300 mg per month), with or without pyrimethamine (50 mg thrice weekly), according to T. gondii serologic results. Baseline evaluation included a complete physical examination; a chest radiograph; and a laboratory evaluation (complete blood counts; CD4 cell counts; toxoplasma serologic results; and levels of alkaline phosphatase, alanine aminotransferase, and creatinine). Antibodies (IgG) to T. gondii were detected by ELFA (Enzyme Linked Fluorescent Assay, bioMerieux SA, RCS LYON B 673 620 399, 69280 Marcy-l'Etoile, France). According to the manufacturer's recommendation, toxoplasma serologic results were considered positive when the titer was 10 IU/mL or greater.

Patients were examined at least every 30 to 60 days. Compliance and symptoms of toxicity were evaluated, physical examinations were done, and blood samples for blood cell counts and biochemical enzyme levels (alkaline phosphatase, alanine aminotransferase, and creatinine) were obtained at every examination. CD4 counts were repeated every 3 to 4 months.

End points were P. carinii pneumonia, toxoplasmosis, and death. To limit potential biases from not blinding, we used strict criteria to diagnose both infections. Pneumocystis carinii pneumonia was diagnosed if suggestive clinical and radiographic findings appeared and if P. carinii was detected in samples of bronchoalveolar lavage fluid or induced sputum. Toxoplasmic encephalitis was diagnosed by clinical findings (fever, neurologic symptoms, or both) and by the appearance of one or more contrast-enhanced focal brain images on computed tomographic scans. A positive response of the patient with toxoplasmosis to sulfadiazine-pyrimethamine or clindamycin-pyrimethamine therapy was defined as improvement in clinical signs and substantial reduction or disappearance of brain images on computed tomographic scan. In addition, cases of P. carinii pneumonia and toxoplasmosis were reviewed by a physician who was not directly involved in the study.

If patients had symptoms suggesting bacterial infection, cultures were done to confirm the diagnosis. Bacteremia was diagnosed by fever and positive blood cultures. Pneumonia was diagnosed by fever, respiratory symptoms, and lobar consolidation in chest radiographs, with or without isolation of bacteria in respiratory or blood samples, or both. Tracheobronchitis was diagnosed by fever, cough, purulent sputum, the absence of new infiltrates on chest radiographs, and response to antibacterial therapy, with or without isolation of bacteria in respiratory samples. Sinusitis was diagnosed by nasal purulent secretion, paranasal sinus opacification on radiographs or computed tomographic scans (or both), and response to antibacterial therapy, with or without isolation of bacteria from sinus secretion obtained by sinus puncture (if it could be done).

Antiretroviral therapy was given concomitantly during the study period. Briefly, zidovudine was offered to patients with CD4 counts of less than 500 x 10⁶/L. Didanosine was given to patients who were intolerant to or whose disease progressed during therapy with zidovudine. Zalcitabine was used in patients intolerant of zidovudine and didanosine. In selected patients, didanosine or zalcitabine was added to zidovudine.

Toxicity Criteria

Adverse reactions thought to be related to study drugs included 1) hematologic toxicity, defined as a reduction of values to below normal limits [normal levels: hemoglobin, 120 g/L; leukocytes, 4.0 x 10⁹/L; neutrophils, 2.5 x 10⁹/L; and platelets, 150 x 10⁹/L] or a greater than 25% reduction of baseline values if they were initially below normal limits, 2) hepatotoxicity, defined as an elevation of aminotransferase levels to more than twice the initial values, and 3) nephrotoxicity, defined as an elevation of creatinine levels to more than twice the initial values.

The following clinical signs were considered to be adverse effects when they were not attributable to other causes and when they disappeared with discontinuation of drug therapy: temperature greater than 38.5 °C, skin rash (generalized erythematous macules or papules with pruritus), and gastric intolerance (epigastric pain, vomiting, or both).

Statistical Analysis

Baseline characteristics of enrolled patients were compared using the Student t-test for quantitative variables or the chi-square or Fisher exact tests for qualitative variables (depending on the expected cell frequencies). The efficacy of both regimens was evaluated by intention-to-treat analysis. We assumed that 95% of patients receiving trimethoprim-sulfamethoxazole would remain free of P. carinii pneumonia and toxoplasmosis at 1 year of follow-up, and we estimated that 100 evaluable patients at risk would be needed in each arm to detect a difference of 15% or greater if it existed, with 90% certainty and a 5% significance level. When 230 patients (149 of whom were seropositive for T. gondii) had been assigned to treatment, we did an interim analysis; because significant differences were found for P. carinii pneumonia, we decided to interrupt the trial.

The significance of the difference in terms of efficacy (preventing P. carinii pneumonia, toxoplasmosis, and death) between treatment groups was calculated by the Mantel-Cox log-rank test of the Kaplan-Meier product-limit estimates, using the 1L program of the BMDP statistical package. All P values were two-sided. The outcome measures were also compared by chi-square or Fisher exact tests.

Patients were censored at the final visit. The eight patients who presented with P. carinii pneumonia or toxoplasmosis (one patient presented with both simultaneously) were censored for these end points at the time of the diagnosis, but they continued to be followed until the last visit for evaluation of survival. Patients who discontinued therapy with the initial allocated drug (because of adverse reactions or poor compliance), with or without switching to other drugs, were censored at the end of the study unless they presented with an end point before the end of the study. In the on-treatment analysis, patients were censored when they discontinued therapy with the initial allocated drug.

Results

Top Methods Results Discussion Author & Article Info References

Patients

Thirty of the 230 initially randomized patients (19 in the dapsone-pyrimethamine arm and 11 in trimethoprim-sulfamethoxazole arm) were not included in the analysis: P. carinii pneumonia was diagnosed in 3 patients within 2 weeks after initiation of the study; 1 patient was receiving prophylaxis with inhaled pentamidine when assigned to treatment; and 26 patients had been lost to follow-up since the day of inclusion in the study, and data could not be collected from them. The only differences in baseline characteristics between these 30 ineligible patients and the 200 eligible patients were that the former were younger (28 compared with 32 years; P = 0.02) and that almost all of them were intravenous drug users (27 [90%] compared with 143 [72%]; P = 0.03). Thus, 200 patients were included in the analysis: 104 in the trimethoprim-sulfamethoxazole arm and 96 in the dapsone-pyrimethamine arm.

The two groups were similar regarding such baseline characteristics as age, sex, risk behavior, CD4 cell counts (in percentages and absolute numbers), CD8 cell counts, CD4/CD8 ratio, blood cell counts, T. gondii serologic results, previous diagnosis of the acquired immunodeficiency syndrome (AIDS), and antiretroviral therapy (Table 1). All patients received antiretroviral monotherapy or combined therapy at some time during the follow-up.

Of the 175 patients receiving zidovudine at the beginning of the study, 111 (63%) continued to receive zidovudine throughout the follow-up: Ninety-seven patients received zidovudine monotherapy; 10 received zidovudine with zalcitabine; and 4 received zidovudine with didanosine. Eight (4%) patients discontinued zidovudine therapy because of poor compliance. Twenty-three (13%) patients (10 in the trimethoprim-sulfamethoxazole arm and 13 in the dapsone-pyrimethamine arm) discontinued zidovudine therapy because of adverse reactions (21 because of hematologic toxicity, 1 because of hepatotoxicity, and 1 because of allergy; of these 23 patients, 14 switched to didanosine, 1 switched to zalcitabine, and 8 took no other antiretroviral drugs). Thirty-three (19%) patients switched to didanosine monotherapy because of HIV disease progression.

Of the 16 patients receiving didanosine at the beginning of the study, 3 had to discontinue therapy because of adverse reactions (1 patient had diarrhea, 1 had peripheral neuropathy, and 1 died of pancreatitis). Finally, of the 9 patients who initiated antiretroviral therapy after the study began, 8 received zidovudine and 1 received zidovudine plus zalcitabine. Overall, no differences were observed between treatment arms in the type of antiretroviral agents received, the duration and the frequency of treatment, or the reasons for discontinuation of treatment.

Efficacy

After a median follow-up of 430 days (range, 30 to 810 days), six patients in the dapsone-pyrimethamine arm (6.3%) and none in the trimethoprim-sulfamethoxazole arm developed P. carinii pneumonia (P < 0.0001). The cumulative rates of P. carinii pneumonia at 12 and 24 months were 0% and 0% for patients receiving trimethoprim-sulfamethoxazole and 4% and 11% for patients receiving dapsone-pyrimethamine (Mantel-Cox, P = 0.014) (Figure 1). The rate of P. carinii pneumonia in the dapsone-pyrimethamine arm was 0.43 per 100 patient-months. Five of the six patients who developed P. carinii pneumonia (one of them also developed toxoplasmosis) had voluntarily stopped taking the drug at least 2 months before the episode of P. carinii pneumonia and had received no other prophylaxis during that period.

View larger version (17K): [in this window] [in a new window]   Figure 1. Cumulative rates of Pneumocystis carinii pneumonia according to prophylaxis regimen (Kaplan-Meier product-limit estimates). DP = dapsone-pyrimethamine; PCP = Pneumocystis carinii pneumonia; TMP-SMX = trimethoprim-sulfamethoxazole (cotrimoxazole).

In the analysis of patients receiving treatment, the cumulative rates of P. carinii pneumonia at 12 and 24 months were 0% and 0% for patients receiving trimethoprim-sulfamethoxazole and 0% and 6% for patients receiving dapsone-pyrimethamine (Mantel-Cox, P = 0.26). The rate of P. carinii pneumonia in patients receiving dapsone-pyrimethamine was 0.08 per 100 patient-months.

One hundred thirty-one patients, 65 in the trimethoprim-sulfamethoxazole arm and 66 in the dapsone-pyrimethamine arm, were seropositive for T. gondii at the beginning of the study. Of these 131 patients, only 1 patient receiving trimethoprim-sulfamethoxazole and 2 patients receiving dapsone-pyrimethamine developed toxoplasmosis during follow-up (for rates of 0.07 and 0.13 per 100 patient-months, respectively; not statistically significant). The cumulative rates of toxoplasmosis at 12 and 24 months were 0% and 4% for patients receiving trimethoprim-sulfamethoxazole and 2% and 7% for patients receiving dapsone-pyrimethamine (Mantel-Cox, P = 0.65) (Figure 2).

View larger version (18K): [in this window] [in a new window]   Figure 2. Cumulative rates of toxoplasmosis in patients seropositive for Toxoplasma gondii according to prophylaxis regimen. (Kaplan-Meier product-limit estimates). DP = dapsone-pyrimethamine; TMP-SMX = trimethoprim-sulfamethoxazole (co-trimoxazole).

In the on-treatment analysis, the cumulative rates of toxoplasmosis at 12 and 24 months were 0% and 4% for patients receiving trimethoprim-sulfamethoxazole and 0% and 8% for patients receiving dapsone-pyrimethamine (Mantel-Cox, P = 0.86). The toxoplasmosis rate was 0.11 per 100 patient-months in patients receiving trimethoprim-sulfamethoxazole and 0.12 per 100 patient-months in patients receiving dapsone-pyrimethamine. Of the 96 patients receiving dapsone-pyrimethamine, 33 received didanosine or rifampin or both (the latter was included in a combination regimen for tuberculosis therapy) during the study period: Twenty received didanosine, 10 received rifampin, and 3 received both drugs. Patients were advised to take dapsone at least 2 hours before or after didanosine. Two of these 33 patients (6%) developed P. carinii pneumonia or toxoplasmosis compared with 5 of 63 patients (7.9%) who did not receive didanosine or rifampin (1 of these 5 patients developed P. carinii pneumonia and toxoplasmosis simultaneously).

When the trimethoprim-sulfamethoxazole group was compared with the dapsone-pyrimethamine group, no differences were found in mortality rates (15 patients [14.4%] and 14 patients [14.6%]; Mantel-Cox, P = 0.85) or in bacterial infection rates (12 episodes and 17 episodes): bacteremia (4 episodes and 6 episodes); pneumonia (4 and 7 episodes); tracheobronchitis (3 episodes and 1 episode); and sinusitis (1 episode and 3 episodes) (P = 0.21). Patient outcome at the end of follow-up is shown in Table 2. No differences were found in the incidence of infections caused by Mycobacterium avium complex (4 [3.8%] in the trimethoprim-sulfamethoxazole arm compared with 2 [2.1%] in the dapsone-pyrimethamine arm; P = 0.46).

Although no differences were observed in the proportion of patients who discontinued drug therapy because of adverse effects or who were lost to follow-up, a statistically significant difference was seen in the number of noncompliant patients (defined as patients who failed to take 50% or more of the drugs for at least 2 consecutive months); 5 (4.8%) patients in the trimethoprim-sulfamethoxazole arm were noncompliant compared with 15 (15.6%) in the dapsone-pyrimethamine arm (P = 0.01).

Toxicity

Nineteen patients (9.5%) had to switch the study drug or switch to another drug because of adverse reactions: 10 (9.6%) patients receiving trimethoprim-sulfamethoxazole (8 switched to dapsone-pyrimethamine and 2 switched to inhaled pentamidine) and 9 (9.3%) patients receiving dapsone-pyrimethamine (6 switched to trimethoprim-sulfamethoxazole and 3 switched to inhaled pentamidine). None of these patients developed P. carinii pneumonia or toxoplasmosis. Of the 14 patients who switched from one study drug to another, 10 (71.4%) could tolerate the second drug even though they had previously presented with fever plus rash (4 patients), rash alone (3 patients), or gastric intolerance (3 patients).

Some differences were observed in the type of adverse reactions that occurred during therapy. More patients receiving trimethoprim-sulfamethoxazole had fever (13.5% compared with 4.2%; P = 0.04) and gastric intolerance (16.3% compared with 8.4%; P = 0.09), whereas more patients receiving dapsone-pyrimethamine had anemia (33.3% compared with 20.2%; P = 0.03) (Table 3).

Discussion

Top Methods Results Discussion Author & Article Info References

Our data suggest that thrice-weekly trimethoprim-sulfamethoxazole is effective in the simultaneous primary prophylaxis of P. carinii pneumonia and toxoplasmosis in patients with HIV infection who are at risk for these other infections. Although dapsone-pyrimethamine was significantly less effective as prophylaxis for P. carinii pneumonia in the intention-to-treat analysis, the regimen used appears to be a good alternative to trimethoprim-sulfamethoxazole. This was especially true in compliant patients because no differences were observed in analyses of patients who were receiving treatment: Only one patient in the dapsone-pyrimethamine arm and no patients in the trimethoprim-sulfamethoxazole arm presented with P. carinii pneumonia while receiving treatment. Dapsone-pyrimethamine was also useful for toxoplasmosis prevention.

Even though patients receiving dapsone-pyrimethamine did not appear to be as compliant as those receiving trimethoprim-sulfamethoxazole, it is difficult to know whether these data are reliable because, as in other published studies [21, 22], compliance was only evaluated by periodic patient questioning. Social problems or problems related to drug abuse and not adverse reactions (including moderate or mild reactions) were identified as the cause of poor compliance in these patients.

Although daily and thrice-weekly trimethoprim-sulfamethoxazole regimens have not been compared, our results reinforce previous data that suggest that the thrice-weekly regimen is as effective as and probably better tolerated than daily trimethoprim-sulfamethoxazole regimens for the simultaneous prophylaxis of P. carinii pneumonia and toxoplasmosis in patients infected with HIV who are at risk for these other infections [11-15].

Retrospective studies have suggested that a trimethoprim-sulfamethoxazole regimen of one double-strength tablet 3 days a week is effective prophylaxis for P. carinii pneumonia [13, 14]. Because few data have been reported on the efficacy of this low-dose trimethoprim-sulfamethoxazole regimen in toxoplasmosis prophylaxis, we decided to use two double-strength trimethoprim-sulfamethoxazole tablets 3 days a week for this purpose.

Although an untreated control group was not included in our study, rates of toxoplasmosis in the trimethoprim-sulfamethoxazole arm (0% at 12 months and 4% at 24 months) were lower than those observed in prospective, comparative studies of patients receiving inhaled pentamidine (10% to 24% after 11 to 24 months) [10, 21, 23] or in observational studies of patients seropositive for T. gondii (26% at 24 months) [27]. However, caution must be exercised when comparing data from different trials, because differences in such baseline characteristics as CD4 counts may influence the results. At the end of our study, approximately 60% of patients in both arms had CD4 counts of less than 100 x 10⁶/L; although these patients are at less than the highest risk for toxoplasmosis, patients with CD4 cell counts between 100 and 200 x 10⁶/L are also at a substantial risk [10, 28, 29]. Of 236 patients with toxoplasmosis who were retrospectively evaluated, 147 had CD4 determinations within 3 months before the diagnosis and 60 (40%) had CD4 counts greater than 100 x 10⁶/L [30].

The only dapsone-pyrimethamine regimen that has shown efficacy in toxoplasmosis prophylaxis in a randomized study was that used by Girard and colleagues [21]. Our dapsone-pyrimethamine regimen proved to be as effective as the schedule used by these investigators in the prevention of both P. carinii pneumonia and toxoplasmosis, and our regimen was better tolerated (9.3% of patients withdrew because of adverse reactions compared with 24%) [21].

Torres and colleagues [22] compared twice-weekly dapsone with inhaled pentamidine as prophylaxis for primary and secondary P. carinii pneumonia in a study not designed to evaluate prophylaxis of toxoplasmosis. Their study showed that dapsone alone may be effective in preventing P. carinii pneumonia; however, pyrimethamine has been shown to have synergistic effects against T. gondii both in vitro and in a murine model [31]. Moreover, when comparing our data with those of Torres and colleagues [22], the addition of pyrimethamine to a twice-weekly regimen of dapsone does not seem to enhance toxicity.

A recent three-arm study [32] by the AIDS Clinical Trials Group (ACTG) compared oral daily dapsone (50 mg twice daily for 7 days/wk) with both oral daily trimethoprim-sulfamethoxazole (160 mg of trimethoprim and 800 mg of sulfamethoxazole twice daily for 7 days/wk) and monthly inhaled pentamidine (300 mg monthly) (ACTG 081) for the primary prophylaxis of P. carinii pneumonia and toxoplasmosis. Although the final results of the ACTG study have not yet been published, to our knowledge, an intention-to-treat analysis showed no differences in the prevention of P. carinii pneumonia and toxoplasmosis among the three arms. In contrast, a higher toxicity was found in the two oral treatment arms, and therefore trimethoprim-sulfamethoxazole and dapsone were withdrawn in 50% and 41% of patients, respectively. These data reinforce the convenience of using safe and effective intermittent regimens in these patients.

Both regimens used in our study were well tolerated; the overall rate of discontinuation of therapy due to drug toxicity was 9.5%. As described in other studies [33-35], many patients who had adverse reactions with the initially allocated study drug (71.4% in our study) could tolerate the other drug, thereby reducing the number of patients who needed another less effective alternative drug for prophylaxis against P. carinii pneumonia or toxoplasmosis.

It has been suggested that trimethoprim-sulfamethoxazole may prevent bacterial infections [33]. We did not find a statistically significant difference between the treatment arms, possibly because of low statistical power due to the small number of events. Both sensitive and resistant microorganisms appeared during follow-up despite trimethoprim-sulfamethoxazole therapy (data not shown). Although dapsone has been previously found to protect against M. avium complex infection [36, 37], we did not find differences between regimens in the appearance of this AIDS-associated complication.

Although some controversial data about the levels of dapsone in patients also receiving didanosine have been reported [38-40], our results suggest that, in patients taking these drugs at least 2 hours apart, the efficacy of dapsone is not reduced.

Our study had some limitations. That it was designed as a nonblinded study could account for potential biases. However, efficacy was measured by objective end points, such as the development of P. carinii pneumonia or toxoplasmosis and death, that were diagnosed on the basis of strict, previously defined criteria. Another limitation of our study was that the number of patients at risk for toxoplasmosis was lower than initially planned. Nevertheless, the power of the study remained great enough to detect a significant biological difference (15% with nearly 80% certainty and 5% significance level), if it existed. In fact, the observed incidence of toxoplasmosis in both arms was small (one patient in the trimethoprim-sulfamethoxazole arm and two in the dapsone-pyrimethamine arm, by intention-to-treat analysis).

Thrice-weekly trimethoprim-sulfamethoxazole with the schedule used in our study appears to be an effective and well-tolerated regimen for the simultaneous primary prophylaxis of P. carinii pneumonia and toxoplasmosis in patients infected with HIV who are at risk for these infections. Twice-weekly dapsone-pyrimethamine offers a safe and effective alternative to trimethoprim-sulfamethoxazole.