ARTICLES
Fluconazole Compared with Amphotericin B plus Flucytosine for Cryptococcal Meningitis in AIDS A Randomized Trial Robert A. Larsen, MD; Mary Ann E. Leal, MD; and Linda S. Chan, PhD
Objective: To compare the efficacy offluconazolewith amphotericin B plusflucytosinein the treatment of cryptococcal meningitis. Design: Patients were randomly assigned to oralfluconazole,400 mg/d, for 10 weeks or to amphotericin B, 0.7 mg/kg body weight daily for 1 week, then three times weekly for 9 weeks combined with flucytosine, ISO mg/kg' d, in four divided doses. Setting: Los Angeles County-University of Southern California Medical Center. Patients: Between 15 February and 7 December 1988, 42 patients had evidence of their first episode of cryptococcal meningitis, of whom 21 participated in the trial. All patients enrolled were men with the acquired immunodeficiency syndrome (AIDS) except one woman who was receiving prednisone therapy and was excluded from the final analysis. Results: Of 14 patients with AIDS assigned to fluconazole, 8 (57%; 95% CI, 29% to 82%) failed; none of the 6 patients with AIDS failed who were assigned to amphotericin B plus flucytosine therapy (0%; CI, 0% to 46%) (Fisher exact test, P = 0.04). The mean duration of positive cerebrospinal fluid cultures was 40.6 ± 5.4 days in patients receivingfluconazoleand 15.6 ± 6.6 days in patients receiving amphotericin B plusflucytosine(Mann-Whitney test, P = 0.02). Overall, 4 patients assigned tofluconazoletherapy died whereas no patient assigned to amphotericin B plus flucytosine therapy died (Fisher exact test, P = 0.27). Conclusion: Amphotericin B used in combination with flucytosine has superior mycologic and clinical efficacy compared with fluconazole for the treatment of cryptococcal meningitis in patients with AIDS.
Annals of Internal Medicine. 1990;113:183-187. From Los Angeles County Hospital and the University of Southern California School of Medicine, Los Angeles, California. For current author addresses, see end of text.
(cryptococcal meningitis is one of the most devastating opportunistic infections suffered by patients with the acquired immunodeficiency syndrome (AIDS). Approximately 10% of patients with AIDS develop cryptococcal meningitis, and nearly 60% of these patients die from this infection (1-3). Treatment with amphotericin B alone or in combination with flucytosine, although preventing death and serious neurologic impairment in some patients, is associated with significant morbidity, particularly renal and hematologic injury (4-7). In addition, many patients develop fevers, chills, rigors, nausea, and vomiting with each dose of amphotericin B. Alternatives to treatment with amphotericin B and flucytosine are needed. Fluconazole, a triazole antifungal agent, shows considerable promise as an effective antifungal for infections caused by Cryptococcus neoformans, both in animal trials and in uncontrolled human evaluations (8-18). We report our experience with fluconazole in a randomized, controlled, prospective clinical trial comparing fluconazole with amphotericin B plus flucytosine in the treatment of cryptococcal meningitis.
Patients and Methods Patients were enrolled if they had clinical evidence of cryptococcal meningitis defined as headache and positive serum or cerebrospinal fluid cryptococcal antigen; were older than 18 years; were able to give informed consent; had an anticipated survival of longer than 2 weeks; were not receiving immune stimulants such as interferon or lymphocyte transfusion; were not taking rifampin, phenytoin, or phenobarbital; did not have significant impairment of liver or renal function (serum bilirubin > 50 /Ltmol/L or creatinine > 310 /imol/L); and were not thought to have another concurrent central nervous system infection (for example, toxoplasmosis or syphilis). Patients with and without AIDS were eligible to participate in the trial. After giving informed consent, patients were randomly assigned (stratified for the presence or absence of AIDS) to either fluconazole or amphotericin B treatment in a 2:1 ratio, respectively. Fluconazole was given orally, 400 mg/d, for 10 weeks. Patients with AIDS were then given 200 mg/d orally thereafter as maintenance therapy. Fluconazole doses during maintenance therapy were increased to 400 mg/d if urine cultures became or were positive at or after the conclusion of the initial 10-week treatment period. Amphotericin B was given intravenously at 0.7 mg/kg body weight daily for the first 7 days, then three times weekly for the next 9 weeks. Patients assigned to amphotericin B treatment were also given oral flucytosine, 150 mg/kg • d, in four divided doses during the initial 10-week treatment period. All patients were followed for 62 weeks after diagnosis unless they died, failed study therapy, or declined further study participation. Renal, hepatic, and hematologic function, and, where appropriate, flucytosine levels were monitored weekly. Blood and cerebrospinal fluid culture and C. neoformans antigen titers © 1990 American College of Physicians
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19709/ by a University of California San Diego User on 02/28/2017
183
Table 1. Severity
of
Crypto coccal
Index of Severity
Meningitis
Fluconazole
,in Patients
with
Amphotericin B plus Flucytosine
AllDS on Entry
into the Clinical
Tri al
P Value*
Relative Riskt
95% CI*
1.00 0.22 0.55 1.00
1.00 0.72 0.78 1.17
0.40 to 1.30 0.42 to 1.42 0.94 to 1.44
Positive/Num ber Tested (%) Cerebrospinal fluid culture Blood culture Urine culture India ink preparation
14/14 12/13 12/14 12/14
(100) (92) (86) (86)
6/6 4/6 4/6 6/6
(100) (67) (67) (100)
* P values were based on the Fisher exact test. izole group. t Relative risks were the ratio ol" the incidence in the arnphotericin B plus flucytiosine group to the inc:idence in the flucons $ 95% confidence intervals were based on the Taylor seiries from the EPI-INFO :statistical program.
were obtained every 2 weeks during the 10-week initial treatment period. Blood cultures consisted of a single Dupont Isolator 10 (DuPont, Wilmington, Delaware) inoculated onto blood, chocolate, and Lowenstein-Jensen agar supplemented with pyruvate and brain-heart infusion agar plates. The brainheart infusion agar plates were held 14 days before being discarded as showing no fungal growth. A minimum of 3.0 mL of cerebrospinal fluid was centrifuged for 30 minutes at 2500 rpm and the sediment used to prepare an India ink mount and inoculate Sauboraud agar with and without chloramphenicol. Urine cultures were inoculated on Sauboraud agar with and without chloramphenicol using the sediment obtained from 50 mL of urine after 30 minutes of centrifugation at 2500 rpm. Cerebrospinal fluid and urine cultures were held for 42 days before being discarded as sterile. All fungal isolates were identified using standard mycologic techniques. Cryptococcus neoformans was identified by its typical colonial morphology, narrow-based budding yeast form, lack of germ tube production when grown in serum broth, and production of urease. All isolates had characteristic sugar fermentation reactions (API, Plainview, New York). Serial twofold dilutions of serum and cerebrospinal fluid were made to obtain a semiquantitative cryptococcal antigen titer (Meridian Diagnostics, Cincinnati, Ohio). An endpoint titer was defined as the last dilution to produce 2+ agglutination of the latex particles. We prospectively assigned a clinical assessment of outcome to all patients. Patients were categorized as a treatment success if blood and cerebrospinal fluid cultures were negative at the end of 10 weeks. Even if urine cultures remained positive, the patient's meningitis was considered successfully treated. Patients were considered to have failed treatment if cultures of cerebrospinal fluid remained positive for 10 weeks with or without neurologic deterioration, or they showed neurologic deterioration requiring a change in study therapy before 10 weeks. Patients who died were categorized as successfully treated or failed based on the last available cerebrospinal fluid and blood culture. For analyzing characteristics recorded as dichotomous variables, the differences between the two treatment groups were tested using the two-tailed Fisher exact test. Relative risk ratios of these characteristics and their 95% confidence intervals (CIs) based on the Taylor series estimation were obtained from the EPI-INFO statistical software package (19). For analyzing characteristics recorded as continuous variables, differences of their mean values between the two treatment groups were tested using the Mann-Whitney two-sample test based on the BMDP03D statistical program (20). The 95% CIs for the mean differences were based on the Student t-statistic. For analyzing the cryptococcal antigen titers, the mean number of twofold dilutions was used. Results Between 15 February and 7 December 1988, 42 patients with evidence of their first episode of cryptococcal meningitis were seen at the Los Angeles CountyUniversity of Southern California Medical Center. Twenty-six patients gave their consent to participate in 184
the trial and were randomized. Of the 16 patients who did not participate, 6 refused, 4 were confused and unable to give informed consent, 2 had significant liver disease, 2 were expected to survive less than 2 weeks, 1 was a prisoner, and 1 was enrolled in a zidovudine trial that excluded the use of fluconazole. Of the 26 patients who entered the trial, 6 were excluded from the final study population. Of the 6 exclusions, 3 patients did not have meningitis (although 2 had other sites of cryptococcal infection), 1 patient was thought to have a second concurrent central nervous system infection and was withdrawn after 6 days of treatment, 1 was the only patient without AIDS (assigned to amphotericin B plus flucytosine treatment and successfully treated), and 1 patient was lost to follow-up after 4 weeks of treatment. This last patient was assigned to fluconazole therapy and had positive cerebrospinal fluid and blood cultures at 4 weeks after which he failed to return for follow-up evaluations. Enrollment was stopped by the principal investigator after 26 patients were entered because all failures were in the group treated with fluconazole. Randomization appeared to produce similar treatment groups. All final study patients were men with AIDS. There were no statistical differences between the groups for the following characteristics (mean ± SE): age, 38 ± 2 years; weight, 60 ± 2 kg; hemoglobin, 107 ± 4 g/L; serum albumin, 34.6 ± 1.5 g/L; and serum sodium, 133 ± 1 mmol/L. The total CD4 cell counts were markedly abnormal in all patients with the mean total CD4 lymphocyte count of 44 ± 13 per mm 3 for those assigned to fluconazole, and 97 ± 36 per mm 3 for those assigned to amphotericin B plus flucytosine (Mann-Whitney, P = 0.055). The presence of previous and concurrent AIDSrelated illnesses were also similar in both groups, with three patients assigned fluconazole and one patient assigned amphotericin B plus flucytosine having no previous AIDS-related infections. The severity of cryptococcal meningitis as measured in the two groups by percent with positive blood and urine cultures, and the proportion of patients with visible cryptococci on India ink mount of cerebrospinal fluid was comparable (Table 1). Cerebrospinal fluid characteristics, such as mean cerebrospinal fluid glucose and protein, serum and cerebrospinal fluid geometric mean cryptococcal antigen titers, were also comparable (Table 2). The geometric mean cryptococcal antigen titers in cerebrospinal fluid were 1:465 and 1:2304 for the fluconazole and amphotericin B plus flucytosine groups, respectively; and the
1 August 1990 • Annals of Internal Medicine • Volume 113 • Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19709/ by a University of California San Diego User on 02/28/2017
geometric mean titers in serum were 1:1771 and 1:2574, respectively. The cerebrospinal fluid opening pressure on lumbar puncture was higher among patients randomized to amphotericin B therapy; however, not all patients had measurements recorded on their initial lumbar puncture. Three of ten patients in the fluconazole treatment group but all three in the amphotericin B treatment group had abnormal opening pressures on their initial lumbar puncture (Fisher exact test, P = 0.07). Of the 14 patients receiving fluconazole, 4 died within 10 weeks of enrollment while none of those receiving amphotericin B plus flucytosine died (Fisher exact test, P = 0.27). Three of those who died had positive cerebrospinal fluid cultures immediately before death, whereas 1 had negative blood and cerebrospinal fluid cultures on two occasions before he died. Death was directly related to cryptococcal meningitis in 2 patients (death on days 6 and 42), one death was due to Pseudomonas aeruginosa pneumonia (the last cerebrospinal fluid culture was positive on day 21, death occurred on day 22), and one death was due to respiratory failure thought to be unrelated to cryptococcal infection (the last cerebrospinal fluid culture was positive on day 28, death occurred on day 69). Three additional patients taking fluconazole were withdrawn and declared treatment failures because of progressive mental deterioration. The cerebrospinal fluid cultures obtained on these three patients were positive at the time they were discontinued from study therapy (on days 42, 28, and 42, respectively). Two patients completed 10 weeks of fluconazole therapy and were declared treatment failures on the basis of a positive 10-week cerebrospinal fluid culture. Both patients continued fluconazole therapy at 400 mg/d and subsequently failed treatment. Thus, at 10 weeks treatment was considered a failure in 8 of 14 patients who received fluconazole. One patient taking fluconazole had a positive urine culture at 10 weeks of treatment, but blood and cerebrospinal fluid cultures were negative. All 6 patients assigned to amphotericin B plus flucytosine therapy survived 10 weeks and had negative blood and cerebrospinal fluid cultures. Urine cultures were positive in 2 patients at 10 weeks. Overall, at 10 weeks of treatment, of 14 patients assigned to fluconazole 8 (57%; CI, 29% to 82%) failed whereas none failed among the 6 patients (0% failures; CI, 0% to
Table 2. Severity
of
Cryptococcal Meningitis
Index of Severity Cryptococcal antigen titerU Cerebrospinal fluid Blood Cerebrospinal fluid profile Pressure, mm of cerebrospinal fluid Protein, gIL Glucose, mmollL
46%) assigned to amphotericin B plus flucytosine therapy (Fisher exact test, P = 0.04). One patient randomized to receive fluconazole whose cultures of blood, cerebrospinal fluid, and urine were negative at 10 weeks developed a positive urine culture at 16 weeks. He subsequently developed cryptococcemia at 32 weeks while taking 400 mg of fluconazole. A second patient randomized to fluconazole whose cultures of blood, cerebrospinal fluid, and urine were negative at 10 weeks had a positive cerebrospinal fluid culture at 14 weeks and recurrence of meningeal symptoms while taking 200 mg of fluconazole. Thus, two of the four surviving patients taking fluconazole and shown to be culture negative in blood, cerebrospinal fluid, and urine at 10 weeks had a subsequent relapse during follow-up evaluations. One patient who received amphotericin B plus flucytosine and had negative blood, cerebrospinal fluid, and urine cultures at 10 weeks was found to have C. neoformans in a urine culture after prostatic massage 1 year after completing study therapy. This patient's cerebrospinal fluid remained culture negative. Mycologic success, measured by conversion of positive to negative cultures, was slower in those receiving fluconazole compared with those receiving amphotericin B. Cerebrospinal fluid cultures remained positive for a mean of 40.6 ± 5 . 4 days (SE) for those receiving fluconazole compared with 15.6 ± 6.6 days for those receiving amphotericin B plus flucytosine (Mann-Whitney test, P = 0.02). Blood cultures had become negative within 7 days (4.2 ± 1.4 days) for all patients receiving amphotericin B plus flucytosine, whereas the mean duration of cryptococcemia was 11.4 ± 3.0 days for those receiving fluconazole (P = 0.19, Mann-Whitney test) and was as long as 4 weeks in some patients. Patients with severe cryptococcal disease, defined as those with a positive blood culture for C. neoformans, with a positive cerebrospinal fluid India ink preparation, who had fewer than 20 leukocytes x 106/L cerebrospinal fluid and a cerebrospinal fluid cryptococcal antigen titer greater than 1:32, did poorly. Cerebrospinal fluid cultures failed to convert to negative in seven of nine patients treated with fluconazole compared with none of three patients treated with amphotericin B plus flucytosine. Those patients with less severe disease fared better: Four of five patients treated with fluconazole
in Pa tients with AIDS
Fluconazole Compared with Amphotericin B plus Flucytosine for Cryptococcal Meningitis in AIDS A Randomized Trial Robert A. Larsen, MD; Mary Ann E. Leal, MD; and Linda S. Chan, PhD
Objective: To compare the efficacy offluconazolewith amphotericin B plusflucytosinein the treatment of cryptococcal meningitis. Design: Patients were randomly assigned to oralfluconazole,400 mg/d, for 10 weeks or to amphotericin B, 0.7 mg/kg body weight daily for 1 week, then three times weekly for 9 weeks combined with flucytosine, ISO mg/kg' d, in four divided doses. Setting: Los Angeles County-University of Southern California Medical Center. Patients: Between 15 February and 7 December 1988, 42 patients had evidence of their first episode of cryptococcal meningitis, of whom 21 participated in the trial. All patients enrolled were men with the acquired immunodeficiency syndrome (AIDS) except one woman who was receiving prednisone therapy and was excluded from the final analysis. Results: Of 14 patients with AIDS assigned to fluconazole, 8 (57%; 95% CI, 29% to 82%) failed; none of the 6 patients with AIDS failed who were assigned to amphotericin B plus flucytosine therapy (0%; CI, 0% to 46%) (Fisher exact test, P = 0.04). The mean duration of positive cerebrospinal fluid cultures was 40.6 ± 5.4 days in patients receivingfluconazoleand 15.6 ± 6.6 days in patients receiving amphotericin B plusflucytosine(Mann-Whitney test, P = 0.02). Overall, 4 patients assigned tofluconazoletherapy died whereas no patient assigned to amphotericin B plus flucytosine therapy died (Fisher exact test, P = 0.27). Conclusion: Amphotericin B used in combination with flucytosine has superior mycologic and clinical efficacy compared with fluconazole for the treatment of cryptococcal meningitis in patients with AIDS.
Annals of Internal Medicine. 1990;113:183-187. From Los Angeles County Hospital and the University of Southern California School of Medicine, Los Angeles, California. For current author addresses, see end of text.
(cryptococcal meningitis is one of the most devastating opportunistic infections suffered by patients with the acquired immunodeficiency syndrome (AIDS). Approximately 10% of patients with AIDS develop cryptococcal meningitis, and nearly 60% of these patients die from this infection (1-3). Treatment with amphotericin B alone or in combination with flucytosine, although preventing death and serious neurologic impairment in some patients, is associated with significant morbidity, particularly renal and hematologic injury (4-7). In addition, many patients develop fevers, chills, rigors, nausea, and vomiting with each dose of amphotericin B. Alternatives to treatment with amphotericin B and flucytosine are needed. Fluconazole, a triazole antifungal agent, shows considerable promise as an effective antifungal for infections caused by Cryptococcus neoformans, both in animal trials and in uncontrolled human evaluations (8-18). We report our experience with fluconazole in a randomized, controlled, prospective clinical trial comparing fluconazole with amphotericin B plus flucytosine in the treatment of cryptococcal meningitis.
Patients and Methods Patients were enrolled if they had clinical evidence of cryptococcal meningitis defined as headache and positive serum or cerebrospinal fluid cryptococcal antigen; were older than 18 years; were able to give informed consent; had an anticipated survival of longer than 2 weeks; were not receiving immune stimulants such as interferon or lymphocyte transfusion; were not taking rifampin, phenytoin, or phenobarbital; did not have significant impairment of liver or renal function (serum bilirubin > 50 /Ltmol/L or creatinine > 310 /imol/L); and were not thought to have another concurrent central nervous system infection (for example, toxoplasmosis or syphilis). Patients with and without AIDS were eligible to participate in the trial. After giving informed consent, patients were randomly assigned (stratified for the presence or absence of AIDS) to either fluconazole or amphotericin B treatment in a 2:1 ratio, respectively. Fluconazole was given orally, 400 mg/d, for 10 weeks. Patients with AIDS were then given 200 mg/d orally thereafter as maintenance therapy. Fluconazole doses during maintenance therapy were increased to 400 mg/d if urine cultures became or were positive at or after the conclusion of the initial 10-week treatment period. Amphotericin B was given intravenously at 0.7 mg/kg body weight daily for the first 7 days, then three times weekly for the next 9 weeks. Patients assigned to amphotericin B treatment were also given oral flucytosine, 150 mg/kg • d, in four divided doses during the initial 10-week treatment period. All patients were followed for 62 weeks after diagnosis unless they died, failed study therapy, or declined further study participation. Renal, hepatic, and hematologic function, and, where appropriate, flucytosine levels were monitored weekly. Blood and cerebrospinal fluid culture and C. neoformans antigen titers © 1990 American College of Physicians
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19709/ by a University of California San Diego User on 02/28/2017
183
Table 1. Severity
of
Crypto coccal
Index of Severity
Meningitis
Fluconazole
,in Patients
with
Amphotericin B plus Flucytosine
AllDS on Entry
into the Clinical
Tri al
P Value*
Relative Riskt
95% CI*
1.00 0.22 0.55 1.00
1.00 0.72 0.78 1.17
0.40 to 1.30 0.42 to 1.42 0.94 to 1.44
Positive/Num ber Tested (%) Cerebrospinal fluid culture Blood culture Urine culture India ink preparation
14/14 12/13 12/14 12/14
(100) (92) (86) (86)
6/6 4/6 4/6 6/6
(100) (67) (67) (100)
* P values were based on the Fisher exact test. izole group. t Relative risks were the ratio ol" the incidence in the arnphotericin B plus flucytiosine group to the inc:idence in the flucons $ 95% confidence intervals were based on the Taylor seiries from the EPI-INFO :statistical program.
were obtained every 2 weeks during the 10-week initial treatment period. Blood cultures consisted of a single Dupont Isolator 10 (DuPont, Wilmington, Delaware) inoculated onto blood, chocolate, and Lowenstein-Jensen agar supplemented with pyruvate and brain-heart infusion agar plates. The brainheart infusion agar plates were held 14 days before being discarded as showing no fungal growth. A minimum of 3.0 mL of cerebrospinal fluid was centrifuged for 30 minutes at 2500 rpm and the sediment used to prepare an India ink mount and inoculate Sauboraud agar with and without chloramphenicol. Urine cultures were inoculated on Sauboraud agar with and without chloramphenicol using the sediment obtained from 50 mL of urine after 30 minutes of centrifugation at 2500 rpm. Cerebrospinal fluid and urine cultures were held for 42 days before being discarded as sterile. All fungal isolates were identified using standard mycologic techniques. Cryptococcus neoformans was identified by its typical colonial morphology, narrow-based budding yeast form, lack of germ tube production when grown in serum broth, and production of urease. All isolates had characteristic sugar fermentation reactions (API, Plainview, New York). Serial twofold dilutions of serum and cerebrospinal fluid were made to obtain a semiquantitative cryptococcal antigen titer (Meridian Diagnostics, Cincinnati, Ohio). An endpoint titer was defined as the last dilution to produce 2+ agglutination of the latex particles. We prospectively assigned a clinical assessment of outcome to all patients. Patients were categorized as a treatment success if blood and cerebrospinal fluid cultures were negative at the end of 10 weeks. Even if urine cultures remained positive, the patient's meningitis was considered successfully treated. Patients were considered to have failed treatment if cultures of cerebrospinal fluid remained positive for 10 weeks with or without neurologic deterioration, or they showed neurologic deterioration requiring a change in study therapy before 10 weeks. Patients who died were categorized as successfully treated or failed based on the last available cerebrospinal fluid and blood culture. For analyzing characteristics recorded as dichotomous variables, the differences between the two treatment groups were tested using the two-tailed Fisher exact test. Relative risk ratios of these characteristics and their 95% confidence intervals (CIs) based on the Taylor series estimation were obtained from the EPI-INFO statistical software package (19). For analyzing characteristics recorded as continuous variables, differences of their mean values between the two treatment groups were tested using the Mann-Whitney two-sample test based on the BMDP03D statistical program (20). The 95% CIs for the mean differences were based on the Student t-statistic. For analyzing the cryptococcal antigen titers, the mean number of twofold dilutions was used. Results Between 15 February and 7 December 1988, 42 patients with evidence of their first episode of cryptococcal meningitis were seen at the Los Angeles CountyUniversity of Southern California Medical Center. Twenty-six patients gave their consent to participate in 184
the trial and were randomized. Of the 16 patients who did not participate, 6 refused, 4 were confused and unable to give informed consent, 2 had significant liver disease, 2 were expected to survive less than 2 weeks, 1 was a prisoner, and 1 was enrolled in a zidovudine trial that excluded the use of fluconazole. Of the 26 patients who entered the trial, 6 were excluded from the final study population. Of the 6 exclusions, 3 patients did not have meningitis (although 2 had other sites of cryptococcal infection), 1 patient was thought to have a second concurrent central nervous system infection and was withdrawn after 6 days of treatment, 1 was the only patient without AIDS (assigned to amphotericin B plus flucytosine treatment and successfully treated), and 1 patient was lost to follow-up after 4 weeks of treatment. This last patient was assigned to fluconazole therapy and had positive cerebrospinal fluid and blood cultures at 4 weeks after which he failed to return for follow-up evaluations. Enrollment was stopped by the principal investigator after 26 patients were entered because all failures were in the group treated with fluconazole. Randomization appeared to produce similar treatment groups. All final study patients were men with AIDS. There were no statistical differences between the groups for the following characteristics (mean ± SE): age, 38 ± 2 years; weight, 60 ± 2 kg; hemoglobin, 107 ± 4 g/L; serum albumin, 34.6 ± 1.5 g/L; and serum sodium, 133 ± 1 mmol/L. The total CD4 cell counts were markedly abnormal in all patients with the mean total CD4 lymphocyte count of 44 ± 13 per mm 3 for those assigned to fluconazole, and 97 ± 36 per mm 3 for those assigned to amphotericin B plus flucytosine (Mann-Whitney, P = 0.055). The presence of previous and concurrent AIDSrelated illnesses were also similar in both groups, with three patients assigned fluconazole and one patient assigned amphotericin B plus flucytosine having no previous AIDS-related infections. The severity of cryptococcal meningitis as measured in the two groups by percent with positive blood and urine cultures, and the proportion of patients with visible cryptococci on India ink mount of cerebrospinal fluid was comparable (Table 1). Cerebrospinal fluid characteristics, such as mean cerebrospinal fluid glucose and protein, serum and cerebrospinal fluid geometric mean cryptococcal antigen titers, were also comparable (Table 2). The geometric mean cryptococcal antigen titers in cerebrospinal fluid were 1:465 and 1:2304 for the fluconazole and amphotericin B plus flucytosine groups, respectively; and the
1 August 1990 • Annals of Internal Medicine • Volume 113 • Number 3
Downloaded From: http://annals.org/pdfaccess.ashx?url=/data/journals/aim/19709/ by a University of California San Diego User on 02/28/2017
geometric mean titers in serum were 1:1771 and 1:2574, respectively. The cerebrospinal fluid opening pressure on lumbar puncture was higher among patients randomized to amphotericin B therapy; however, not all patients had measurements recorded on their initial lumbar puncture. Three of ten patients in the fluconazole treatment group but all three in the amphotericin B treatment group had abnormal opening pressures on their initial lumbar puncture (Fisher exact test, P = 0.07). Of the 14 patients receiving fluconazole, 4 died within 10 weeks of enrollment while none of those receiving amphotericin B plus flucytosine died (Fisher exact test, P = 0.27). Three of those who died had positive cerebrospinal fluid cultures immediately before death, whereas 1 had negative blood and cerebrospinal fluid cultures on two occasions before he died. Death was directly related to cryptococcal meningitis in 2 patients (death on days 6 and 42), one death was due to Pseudomonas aeruginosa pneumonia (the last cerebrospinal fluid culture was positive on day 21, death occurred on day 22), and one death was due to respiratory failure thought to be unrelated to cryptococcal infection (the last cerebrospinal fluid culture was positive on day 28, death occurred on day 69). Three additional patients taking fluconazole were withdrawn and declared treatment failures because of progressive mental deterioration. The cerebrospinal fluid cultures obtained on these three patients were positive at the time they were discontinued from study therapy (on days 42, 28, and 42, respectively). Two patients completed 10 weeks of fluconazole therapy and were declared treatment failures on the basis of a positive 10-week cerebrospinal fluid culture. Both patients continued fluconazole therapy at 400 mg/d and subsequently failed treatment. Thus, at 10 weeks treatment was considered a failure in 8 of 14 patients who received fluconazole. One patient taking fluconazole had a positive urine culture at 10 weeks of treatment, but blood and cerebrospinal fluid cultures were negative. All 6 patients assigned to amphotericin B plus flucytosine therapy survived 10 weeks and had negative blood and cerebrospinal fluid cultures. Urine cultures were positive in 2 patients at 10 weeks. Overall, at 10 weeks of treatment, of 14 patients assigned to fluconazole 8 (57%; CI, 29% to 82%) failed whereas none failed among the 6 patients (0% failures; CI, 0% to
Table 2. Severity
of
Cryptococcal Meningitis
Index of Severity Cryptococcal antigen titerU Cerebrospinal fluid Blood Cerebrospinal fluid profile Pressure, mm of cerebrospinal fluid Protein, gIL Glucose, mmollL
46%) assigned to amphotericin B plus flucytosine therapy (Fisher exact test, P = 0.04). One patient randomized to receive fluconazole whose cultures of blood, cerebrospinal fluid, and urine were negative at 10 weeks developed a positive urine culture at 16 weeks. He subsequently developed cryptococcemia at 32 weeks while taking 400 mg of fluconazole. A second patient randomized to fluconazole whose cultures of blood, cerebrospinal fluid, and urine were negative at 10 weeks had a positive cerebrospinal fluid culture at 14 weeks and recurrence of meningeal symptoms while taking 200 mg of fluconazole. Thus, two of the four surviving patients taking fluconazole and shown to be culture negative in blood, cerebrospinal fluid, and urine at 10 weeks had a subsequent relapse during follow-up evaluations. One patient who received amphotericin B plus flucytosine and had negative blood, cerebrospinal fluid, and urine cultures at 10 weeks was found to have C. neoformans in a urine culture after prostatic massage 1 year after completing study therapy. This patient's cerebrospinal fluid remained culture negative. Mycologic success, measured by conversion of positive to negative cultures, was slower in those receiving fluconazole compared with those receiving amphotericin B. Cerebrospinal fluid cultures remained positive for a mean of 40.6 ± 5 . 4 days (SE) for those receiving fluconazole compared with 15.6 ± 6.6 days for those receiving amphotericin B plus flucytosine (Mann-Whitney test, P = 0.02). Blood cultures had become negative within 7 days (4.2 ± 1.4 days) for all patients receiving amphotericin B plus flucytosine, whereas the mean duration of cryptococcemia was 11.4 ± 3.0 days for those receiving fluconazole (P = 0.19, Mann-Whitney test) and was as long as 4 weeks in some patients. Patients with severe cryptococcal disease, defined as those with a positive blood culture for C. neoformans, with a positive cerebrospinal fluid India ink preparation, who had fewer than 20 leukocytes x 106/L cerebrospinal fluid and a cerebrospinal fluid cryptococcal antigen titer greater than 1:32, did poorly. Cerebrospinal fluid cultures failed to convert to negative in seven of nine patients treated with fluconazole compared with none of three patients treated with amphotericin B plus flucytosine. Those patients with less severe disease fared better: Four of five patients treated with fluconazole
in Pa tients with AIDS
