Early reports of azole antifungal therapy for the treatment of coccidioidal meningitis offered some encouragement. Parenterally administered miconazole (Monistat IV, Janssen Pharmaceutica, Inc., Piscataway, New Jersey) or orally administered ketoconazole (Nizoral, Janssen Pharmaceutica, Inc.) at high doses appeared to produce therapeutic responses in some patients [13-16]. Both drugs showed significant toxicity, not all patients responded, and responses often were not sustained after therapy was discontinued. Consequently, miconazole and ketoconazole are regarded as second-line therapies compared with intrathecal amphotericin B.

Recently, triazole antifungal agents with an improved toxicology profile have become available. Both itraconazole (Sporanox, Janssen Pharmaceutica, Inc.) and fluconazole (Diflucan, Pfizer, Inc., New York, New York) are systemically absorbed after oral administration, and cerebrospinal fluid penetration of fluconazole is excellent [17-21]. Although neither drug is approved by the Food and Drug Administration for the treatment of any manifestation of coccidioidomycosis, preliminary reports of itraconazole and fluconazole treatment of coccidioidal meningitis have suggested that these agents might be useful for the treatment of this disease [22-24]. However, most of these patients were treated previously or concurrently with amphotericin B, and it is therefore difficult to determine accurately the contribution of the azole component to the observed responses. We describe the results of a collaborative study of the treatment of coccidioidal meningitis with fluconazole. Our results corroborate those of earlier investigations and offer guidance for future studies.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Patients were enrolled by collaborating investigators of the NIAID-Mycoses Study Group (Table 1). All patients provided informed consent, and the study was conducted with the approval of the Human Subjects Institutional Review Board at each site. Enrollment began on 2 August 1988, and the 50th patient began therapy on 10 September 1990.

Entry and Exclusion Criteria

The diagnosis of coccidioidal meningitis was established by 1) isolation of C. immitis from the cerebrospinal fluid; 2) complement fixing-type anticoccidioidal antibodies [25] detected in the cerebrospinal fluid in the presence of other cerebrospinal fluid abnormalities typical of coccidioidal meningitis [12]; or 3) illness plus cerebrospinal fluid abnormalities compatible with chronic meningitis and either detection of serum complement-fixing type antibodies or isolation of C. immitis from an extraneural site. For the purpose of analysis, patients with coccidioidal meningitis were categorized as having infections without previous therapy if they had received no intrathecal amphotericin B (two patients who had received only 0.16 and 2.35 mg intrathecal amphotericin were also included in this group) or if they experienced infection relapse after previous therapy. The latter category comprised patients in whom intrathecal amphotericin B could no longer be administered because of lack of access to the cerebrospinal space, neurologic or hemodynamic reactions to amphotericin B that precluded further intrathecal administrations, or coccidioidal infection that had recurred after discontinuation of antifungal therapy. Recurrent infection was documented by a twofold deterioration of at least three of the following cerebrospinal fluid abnormalities on two separate determinations made at least 1 week apart: cerebrospinal fluid white blood cells of 50 cells/mm³ or more, glucose of 30 mg/dL or less, protein of 100 mg/dL or more, and complement fixing-type antibodies detectable at a level 1:2 dilutions. Patients who had achieved adequate control with conventional therapies were ineligible for enrollment.

Treatment Regimen

Initial therapy for most patients was 400 mg of fluconazole given orally once per day. For patients weighing less than 50 kg or more than 90 kg, the initial dose was adjusted to the multiple of 100 mg that was closest to but did not exceed 6 mg per kg. In some patients, intravenous fluconazole was substituted for oral fluconazole for up to several days when oral intake was not possible. For patients whose creatinine clearance was estimated to be less than 50 mL/min, the dose was reduced by 50%. Patients who could not tolerate 400 mg/d were treated with the highest tolerable dose. Frequently therapy was initiated without hospitalization and most patients were managed on an ambulatory basis.

Conduct of Study

At the start of treatment, in addition to a general medical history and physical examination, specific baseline information was recorded (Table 2). Most patients had a computed tomographic scan of the head at the time coccidioidal meningitis was first diagnosed or at the initiation of fluconazole therapy. Cerebrospinal fluid was examined by enumerating leukocytes, measuring glucose and protein concentrations, and submitting portions for mycologic cultures. Cerebrospinal fluid and serum complement fixing-type antibodies were measured by a single reference laboratory using a quantitative double immunodiffusion technique [26]. Most cerebrospinal fluid specimens were obtained by lumbar puncture. Because the route used to obtain cerebrospinal fluid measurements systematically influences results [27], however, subsequent cerebrospinal fluid evaluations were done at the same site of cerebrospinal fluid access as at baseline in all cases. Additional laboratory studies included evaluations of complete peripheral blood count, partial thromboplastin time, prothrombin time, serum electrolytes, blood urea nitrogen, serum creatinine, serum alanine aminotransferase, serum aspartate aminotransferase, lactate dehydrogenase, alkaline phosphatase, total bilirubin, and urinalysis.

During the course of therapy, signs, symptoms, and routine laboratory studies were assessed every 2 to 4 weeks during the first 4 months and every 2 months thereafter. Examinations of the cerebrospinal fluid and measurements of serum complement fixing-type antibodies were repeated at least every 4 months; early in the course of treatment, these studies were usually more frequent. A computed tomographic scan of the head was repeated if the development of hydrocephalus was suspected.

Data Analysis

Abnormalities detected in baseline and subsequent evaluations were assigned numeric values based on the predefined point system outlined in Table 2, and an overall assessment score was computed as the point total. Specific observations that were absent from the baseline assessment were not assessed in follow-up evaluations. When information regarding specific abnormalities was unavailable on reassessment, the previous abnormality was assumed to still be present. Thus, all changes in score reflect demonstrated changes in the parameters evaluated. The baseline score was used as 100% abnormality for each patient, and subsequent evaluations were regarded as showing improvement if the scores decreased compared with baseline or as showing deterioration if they increased compared with baseline.

A response was defined as 40% or greater reduction in abnormalities without subsequent relapse during fluconazole treatment. Relapse during therapy used the same criteria as described above for recurrence with previous therapy. After 8 months, a patient not achieving this level of improvement was considered to be a nonresponder and was removed from the treatment protocol. Statistical significance was assessed using the Pearson chi-square test for categorical variables and the nonparametric Wilcoxon test for differences in rates of response among different patient subgroups. The data reported here excluded unevaluable patients; however, no differences in interpretation resulted from identical analyses that followed the intent-to-treat rule (unevaluable patients counted as nonresponders).

Results

Top Methods Results Discussion Author & Article Info References

Patient Demographics and Baseline Abnormalities

Of the 50 patients, most patients were white men who ranged in age from 8 to 77 years (see Table 1). Nine patients were infected with human immunodeficiency virus [HIV]. Of these, eight were men, and three had potentially immunosuppressing conditions (Hodgkin disease, corticosteroid therapy, and uremia). By chance, exactly one half of the enrolled patients had received no previous therapy for their meningitis. During the course of enrollment, however, the proportion of the two groups was not uniform. Patients who had relapsed from previous therapy enrolled earlier compared with patients who had not previously been treated (average time of entry after the study had begun, 11.1 and 16.3 months, respectively; Wilcoxon nonparametric test; P = 0.01). We attributed this pattern to increasing awareness within our communities of the availability of fluconazole as initial treatment for coccidioidal meningitis. Of the nine HIV-infected patients, eight received fluconazole as initial treatment (chi-square test; P = 0.01). In other respects, the two groups appeared to be similar. In 47 patients, the diagnosis of meningitis was established by isolating C. immitis or by detecting complement fixing-type antibodies from the cerebrospinal fluid. In the other three patients, the diagnosis was based on the findings of chronic meningitis and extrameningeal evidence of coccidioidomycosis; all had had repeated courses of intrathecal amphotericin B before enrollment. For all patients, meningitis was the sole manifestation of extrapulmonary coccidioidal dissemination.

The frequency of abnormalities recorded as resulting from coccidioidal meningitis and their contribution to the baseline numeric assessment are shown in Table 2. Patients with no previous therapy manifested more symptoms, had higher concentrations of cerebrospinal fluid leukocytes, and yielded C. immitis more frequently from cerebrospinal fluid cultures. Overall median scores for patients with no previous therapy and for those who had relapsed from previous therapy were 7 and 6, respectively (P = 0.52). Thus, patients without previous therapy who were enrolled in our study showed a tendency toward more serious illness than did those who had survived one or more courses of previous treatment.

Response to Treatment

Of the 50 patients in the study, 3 were considered unevaluable. Two failed to return after enrollment. One of these patients had the acquired immunodeficiency syndrome (AIDS), withdrew from all therapy (including experimental fluconazole therapy), and died at home. A third patient had a stroke 3 months after beginning fluconazole therapy, and subsequent evaluations were not deemed reliable for assessment of outcome.

Of the 47 evaluable patients, 37 (79%; CI, 61% to 90%) responded to treatment and 10 (21%) failed treatment (Table 3). Of the 37 responders, 36 were categorized by the predefined scoring system. A malfunction of his ventriculopleural shunt developed in one other patient with pre-existing hydrocephalus unrelated to coccidioidal infection. At the time of its revision, C. immitis was isolated from his cerebrospinal fluid and spherules were seen in granulomatous tissue attached to the ventricular end of the excised apparatus. With fluconazole therapy, the revision has functioned without incident for more than 3 years and the patient has had no symptoms from his infection. It was not possible to obtain cerebrospinal fluid from this patient, making the numeric assessment of his response impractical.

Ten responding patients died of unrelated causes. Six of nine patients with AIDS achieved a response and survived from 9 to 26 months. Each had other complications of AIDS (chronic diarrhea, weight loss, inanition, and bacterial sepsis) that were sufficient to account for their deaths. In two of these patients, however, it is possible that recrudescence of coccidioidal infection may have contributed to their deterioration. One HIV-infected patient developed a focal cerebral mass that on biopsy yielded C. immitis after 25 months of therapy. In the following 2 weeks, respiratory distress ensued, leading to death. At autopsy, C. immitis was also recovered from the lungs. Another HIV-infected patient showed an increase in cerebrospinal fluid white blood cells 2 weeks before death. A cerebrovascular accident, pulmonary embolism, uncontrolled Hodgkin's disease, and urosepsis accounted for four other deaths unrelated to coccidioidal meningitis in patients uninfected with HIV.

Of responders who did not die of other causes, therapy has been continued for 2 to 4 years in 25 patients (median follow-up, 38 months). One responding patient discontinued therapy after 7 months, against the investigator's recommendation. Eight months later, the patient was hospitalized with hydrocephalus and other evidence of active infection. A second patient was stable but chose to continue fluconazole off study after 479 days of treatment. Another patient, who was infected with HIV, showed evidence of hydrocephalus on computed tomographic scan. The condition was mild, however, and did not require a decompression procedure.

Ten patients failed to respond to 400 mg/d of fluconazole and were removed from the study. One patient, in whom hydrocephalus developed after 6 weeks, was given intrathecal amphotericin B, received a ventriculoperitoneal shunt, and 3 months later was returned to fluconazole therapy (not as part of this study), 400 mg/d, which he has continued. Of the other nine patients, three showed no evidence of improvement and were withdrawn after 4, 5, and 7 months, respectively, whereas six patients were withdrawn between 8 and 16 months after an initial interval of improvement. The subsequent courses of these eight patients' therapies were not controlled by the treatment protocol. Two HIV-infected patients died (one despite receiving 2 months of fluconazole therapy at 800 mg/d), a third patient was given intrathecal amphotericin B, and the remaining six patients were treated with 800 mg/d of fluconazole, which they have received from 15 to 20 months. Four of these patients have shown improvement with the higher fluconazole dose, one patient developed hydrocephalus and had intrathecal amphotericin B reinstituted after 14 months of fluconazole therapy (800 mg/d), and one patient has been lost to follow-up.

In comparing responders and nonresponders for each of the baseline characteristics shown in Table 2, meningismus and headache showed an apparent difference between the two groups. Meningismus was present in 7 of 10 (70%) nonresponders but in only 8 of 37 (22%) responders (P = 0.004). Similarly but less significantly, headache was present in all nonresponders but in only 76% of responders (P = 0.08). These observations suggest that nonresponse was more likely in patients with more severe symptoms. Although responses were more frequent in patients with previous therapy compared with those without previous therapy (88% compared with 70%), this difference did not reach statistical significance (P = 0.13). Patients with HIV were nearly as likely to respond, as were non-HIV-infected patients (75% compared with 79%, P > 0.2).

Rates of Change of Specific Abnormalities during Therapy

Kaplan-Meier calculations of the likelihood of achieving a response with fluconazole as either initial or retreatment therapy over the course of study is shown in Figure 1. Most responses were evident within the first 4 months of treatment, regardless of previous therapy. In similar analyses, neither HIV infection nor pre-existing hydrocephalus influenced the rate of response.

View larger version (10K): [in this window] [in a new window]   Figure 1. Kaplan-Meier estimates of the likelihood of achieving a response with fluconazole as either initial or retreatment therapy during the first 20 months of therapy. No difference in the rate of response was noted between patients treated for the first time and those who had relapsed from previous therapies. P > 0.2.

All components of the composite score showed reduced scores for the responding group as a whole, but the rate of improvement varied (Figure 2). The symptom component of the score was the quickest to improve, whereas other parameters such as cerebrospinal fluid white blood cell count and the glucose, protein, and antibody concentrations improved more slowly. In two responding patients, cerebrospinal fluid white blood cell counts actually increased from 150 to 734 cells/mm³ at 64 days and from 34 to 96 cells/mm³ at 33 days, respectively, before normalizing. Serologic responses were the slowest to improve. After fluconazole treatment, however, 15 of the responding patients had levels of antibodies in their cerebrospinal fluid that were undetectable by our reference laboratory.

View larger version (34K): [in this window] [in a new window]   Figure 2. Changes in the scores for different components of assessment during the course of therapy. All components decreased with treatment. Symptoms showed the earliest change, whereas cerebrospinal fluid abnormalities improved more slowly. CSF = cerebrospinal fluid.

Although most abnormalities eventually improved, low-level abnormalities have persisted in some patients. For example, in 20 responders for whom cerebrospinal fluid specimens were obtained after at least 20 months, cerebrospinal fluid leukocyte counts in 5 patients (24%) have persisted at levels of 32 to 109 cells/mm³ despite an absence of associated symptoms (Figure 3). Similarly, glucose values were between 31 and 39 mg/dL in 3 patients, and protein values were between 50 and 312 mg/dL in 9 of 16 unobstructed patients. Overall, 15 patients had abnormalities in at least one of these cerebrospinal fluid parameters. Thus, despite the generally favorable pattern, considerable heterogeneity was seen in the rate and completeness of the resolution of evaluated abnormalities.

View larger version (19K): [in this window] [in a new window]   Figure 3. Changes in cerebrospinal fluid leukocyte counts in responding patients during fluconazole therapy. Although most patients' cerebrospinal fluid leukocyte counts normalized, several did not, even after 20 months of treatment. CSF = cerebrospinal fluid.

Drug Tolerance and Toxicity

Only two patients were unable to tolerate the initial dosage, both because of mental status complaints. One patient was a 24-year-old man with recurrent Hodgkin disease who had developed reticulonodular pulmonary infiltrates due to C. immitis. His pulmonary infection was controlled with 3.0 g intravenous amphotericin B over 4 months, but 6 weeks after completing this treatment headache and recurrent fever developed; a lumbar puncture was diagnostic of coccidioidal meningitis. On the 28th day after starting 400 mg/d of fluconazole, he was hospitalized because of psychotic behavior. The dose of fluconazole was reduced to 200 mg/d, and the psychosis resolved. Several months later, fluconazole was again increased to 400 mg/d, and mental status changes were again noted within a few days. A dose of 200 mg/d was administered for the remaining 22 months of treatment. Another patient, after receiving 400 mg/d for 2 weeks, noted difficulty concentrating. This symptom was not present with a dose of 200 mg/d, which has been used for most of his therapy. This association was shown on two subsequent attempts to increase the dose to 400 mg/d, the second of which took place 2.5 years after starting treatment.

Liver function abnormalities during treatment developed in three patients. One HIV-infected patient developed isolated alkaline phosphatase measurements ranging from two to six times the upper limits of normal after 8 months of treatment. She continued therapy for 11 additional months until her death (autopsy not done). The other two patients had aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase abnormalities ranging from two to five times the upper limits of normal 5 and 10 months after the start of therapy, respectively. Both were associated with multiple drugs for the treatment of underlying Hodgkin disease and AIDS. Both continued therapy for 10 or 12 months until their deaths, and an autopsy in one patient showed no evidence of drug-induced hepatotoxicity.

Minor complaints that appeared possibly or probably related to fluconazole included hair loss (three patients), nausea or anorexia (two patients), tinnitus (one patient), and somnolence (one patient). Splitting the dose or giving fluconazole at bed time resolved all of these symptoms except hair loss. None of these symptoms necessitated the discontinuation of therapy.

Discussion

Top Methods Results Discussion Author & Article Info References

Fluconazole given orally at a dose of 400 mg/d produced improvement in most of our patients with coccidioidal meningitis. Not only was this therapy very easy to administer but adverse drug effects were few and were much easier to manage than those commonly encountered with intrathecal amphotericin B treatments. For evaluable patients who had received no previous treatment of their meningitis, 16 of 23 (70%) achieved responses, whereas 21 of 24 evaluable patients (88%) with previous treatment responded (P = 0.13). Although the difference in response rates between the two groups did not reach statistical significance, it is possible that patients who have survived previous courses of therapy before enrollment in this study formed a subset of patients with milder or more responsive infections. Consistent with this possibility are differences in the baseline characteristics of the two groups and the apparent association of severe symptoms with nonresponse. Alternatively, previous therapy with amphotericin B may in some way enhance the response to fluconazole.

Similar response rates were found regardless of the presence of either HIV infection or of hydrocephalus. This finding is particularly encouraging because both of these patient groups have added difficulties with standard intrathecal amphotericin B therapy. For HIV-infected patients, the morbidity and medical support associated with intrathecal administrations would detract substantially from their independence and quality of life. For patients with hydrocephalus, cerebrospinal fluid shunts may divert intrathecally administered amphotericin B from reaching the site of coccidioidal infection and render such treatment ineffective [12]. Thus, in those patient groups, fluconazole therapy seems to offer particular advantages over standard options.

The natural history of this disease can only be gleaned from historical information, and past reports have indicated that most patients with coccidioidal meningitis died within 2 years [1-5]. Recently, a detailed analysis of Veterans Affairs and Armed Forces records collected between 1955 and 1958 has provided detailed survival curves that lend themselves to statistical comparison with our study [28]. A significant finding in that study was that patients with coccidioidal meningitis as the sole site of extrapulmonary dissemination had a significantly better survival than did those with multiple sites of dissemination in addition to meningitis. Even so, all patients identified with meningitis during the study period had died within 31 months. In contrast, in our study of 23 patients first treated with fluconazole for their meningitis, 16 survived longer than 2 years or died of unrelated causes (see Table 3). Thus, compared with historical controls, fluconazole is of clear value in this serious disease.

Before this study, intrathecal amphotericin B was the mainstay of therapy for coccidioidal meningitis. A comparison between fluconazole and intrathecal amphotericin B by randomly assigning patients to one therapy or the other would be the most direct approach to determine which is more useful. However, the wide diversity of drug administration methods among the collaborating centers precluded agreement on a uniform regimen for the amphotericin B arm of such a study, and no individual center was capable of sufficient enrollment to complete a randomized trial. Labadie and Hamilton [11] reviewed the survival of 505 patients reported in seven separate studies published between 1964 and 1986. The overall survival rate was 56.6%, ranging from 51% to 100% among the different reports, depending on the specific technique of drug administration, drug dosing, patient factors, and length of observation [2, 7-11, 29]. Compared with these figures, the response rate of 79% (95% CI, 61% to 90%) for fluconazole reported here seems at least as satisfactory. On the other hand, intrathecal amphotericin B therapy is difficult to administer. Nearly always it results in major toxicity, and each dose requires a great deal of physician time and ancillary support, in striking contrast to the ease of administration and lack of untoward effects seen with fluconazole. Even without a formal comparison, however, fluconazole therapy is a much simpler and less toxic mode of therapy. Currently, the cost of a year's treatment with 400 mg/d of fluconazole and routine laboratory monitoring is approximately $10 000. The costs of administering intrathecal amphotericin B vary widely among centers and are more difficult to estimate; however, we believe that the costs exceed those of fluconazole therapy.

Another triazole antifungal agent, itraconazole, has been reported to also be effective in patients relapsing from previous intrathecal therapy for coccidioidal meningitis [22]. In that report, seven of eight evaluable patients improved, a result similar to those of our study. However, three of the patients received concurrent amphotericin B during part of their treatment, and three patients manifested minimal evidence of activity of their infection at the time itraconazole therapy was begun. These differences make it difficult to rank the two drugs' efficacy without additional studies.

Only one responding patient discontinued fluconazole therapy, and he subsequently relapsed. In a recent evaluation of 14 patients who have discontinued therapy for coccidioidal meningitis with any of several azoles (ketoconazole, itraconazole, or fluconazole), 9 patients relapsed [30]. These observations, in concert with the lack of adverse drug-related events, indicate the advisability of continuing treatment indefinitely, although the precise risk for relapse after stopping fluconazole is unknown.

The response rate of coccidioidal meningitis to 400 mg/d of fluconazole is encouraging, and we believe that this dose is reasonable for the initiation of therapy for most patients. However, at least two important unresolved issues remain. First, 10 (21%) of our patients did not respond to this dose. Also, many patients who did respond did so slowly. Because some of the nonresponders appeared to improve on higher doses of fluconazole, we believe that it is important for future studies to determine the frequency and rate of response to doses of 800 mg or more per day. Second, despite the relative absence of symptoms, 24% of responding patients exhibited a persistent cerebrospinal fluid pleocytosis as well as other cerebrospinal fluid abnormalities. Whether these findings have long-term significance is unknown but conceivably may be associated with arachnoiditis or presage chronic neurologic complications. Because of this uncertainty, extended supervision of our patients will be needed to determine if such complications occur. If complications are a risk, higher doses of fluconazole may be useful to reduce the frequency of their occurrence. Alternatively, abbreviated courses of intrathecal amphotericin B before or during some portion of fluconazole therapy may prove necessary to prevent late complications.

In light of the present study, fluconazole has emerged as an effective therapy for a disease that is difficult to treat. Fluconazole may also be effective treatment for other manifestations of coccidioidal infection, although this possibility cannot be assessed in our patients because meningitis was their sole manifestation of disseminated infection. The ease of administration and lack of untoward effects of fluconazole make this treatment even more attractive. Not all patients have responded to 400 mg/d, however, and increasing the dose of fluconazole or instituting intrathecal amphotericin B may be required in some patients. Furthermore, because fluconazole may not be curative, we recommend that patients continue treatment indefinitely. Finally, experience is limited to treatment durations of 4 years or less, and the possibility exists that as yet unforeseen chronic neurologic complications of persistent low-level infection may arise. Further studies are needed to optimize the use of fluconazole for patients with coccidioidal meningitis.

Note Added in Proof

Between acceptance and publication, two patients receiving fluconazole as initial therapy have worsened while still receiving treatment. A pediatric patient, weighing less than 50 kg, developed hydrocephalus and after 778 days of therapy had her fluconazole dose increased from 200 mg to 400 mg per day. Another patient had an increase in CSF leukocyte count and a decrease in glucose, and after 1172 days of therapy, her dose was doubled to 800 mg per day.