When the oral azole antifungal drugs were introduced [5-10], they provided a safe and effective alternative to intravenous therapy in patients with disseminated coccidioidomycosis. Initially with ketoconazole [11], and later with the triazoles, fluconazole and itraconazole [12, 13], it was clear that some patients with meningitis could achieve both remission of symptoms and changes in the cerebrospinal fluid toward normal with oral therapy. Initially, azoles were used to taper and discontinue amphotericin therapy in patients who were receiving amphotericin into the cerebrospinal fluid and thus to avoid the complications of this form of amphotericin therapy. It was later shown that some patients with meningitis could achieve remission with oral therapy alone [11-15]. Recrudescences of meningeal diseases in patients who achieve remission and continue to receive therapy appear to be rare (6 of 87 evaluable patients in previous studies [11-15]).

In the 25 years during which amphotericin administered into the cerebrospinal fluid was the only form of therapy for coccidioidal meningitis, cures were obtained in some patients [3, 4]. Clinicians developed treatment schedules that involved initial intensive amphotericin therapy, and, if symptomatic and cerebrospinal fluid remissions were achieved, therapy was gradually withdrawn. Cure could be defined in those patients who tolerated withdrawal of therapy without exacerbation of disease and then remained in remission without receiving any therapy for a prolonged time (for example, 2 years) [3, 4]. Because 1) experience with the new oral therapy was limited, 2) there was concern that azoles might be fungistatic in vivo compared with amphotericin, and 3) it was less urgent to discontinue therapy with azoles than with amphotericin, it was unclear whether patients successfully treated with azoles were cured or whether their disease was only suppressed. Criteria for discontinuing azole therapy were needed.

In this study, we were concerned with the subset of patients who responded to therapy with oral azoles and maintained their response. We report the results of long-term follow-up of all patients known to us who responded to oral therapy and in whom anti-fungal therapy was subsequently discontinued.

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Our patients were drawn from previous studies in which we have participated [11-15], from our practices, and from referrals by colleagues. In some instances, follow-up information was obtained through correspondence, circulated questionnaires, or presentations at medical meetings. Because most of the patients included came from studies or from our practices, and thus received follow-up attention from us, and because requests for information specified that all patients in whom therapy had been discontinued should be identified, we have no reason to suspect bias in accrual for or against patients who have had relapse. Our patients include those in Mycoses Study Group study 3C (ketoconazole) [11], studies of fluconazole [12] and itraconazole [13], the Pan-American Study Group study of fluconazole [14], and the NIAID-Mycoses Study Group study 11 (fluconazole) [15]. We began our study in 1990 and have continued it to the present, as we attempt to obtain information on all patients with coccidioidal meningitis in whom azole therapy is discontinued for any reason. No patient who initially received intracerebrospinal fluid therapy was included in our study if a sample of the cerebrospinal fluid indicated disease inactivity at the time that intracerebrospinal fluid therapy was tapered concurrent with the initiation of azole therapy.

Table 1 lists the characteristics of the 18 patients who have completed treatment for coccidioidal meningitis. Two patients (patients 3 and 13), who have each had two relapses, are included in this series. To avoid skewing the analyses of factors influencing relapse, each patient was counted once for statistical purposes and is presented only once in the line listing in tabular form (Table 1). For patients 3 and 13, this listing shows their first and second relapses, respectively. The data for these two patients were selected on the basis of the patients' status at the start of our study; patient 3 had not yet had his second relapse, and patient 13 had already had both relapses. Both relapses in both patients are discussed below.

Relapse was defined in patients receiving no antifungal therapy as central nervous system symptoms or signs plus worsening cerebrospinal fluid indices (protein or glucose concentrations, leukocyte counts, or coccidioidal antibody titers) relative to the last cerebrospinal fluid sample available before the end of therapy, if the combination of these factors had caused the patient's physician to reinstitute therapy.

Laboratory Analyses

Antibody assay was done using the complement fixation or immunodiffusion methods [16] for IgG at reference laboratories in California, Arizona, and Texas. Serial measurements for a given patient were all done in the same laboratory.

Statistical Analysis

Bivariate analyses were done on categorical measurements using the Fisher exact test; the Kruskal-Wallis test was used for ordinal measurements. Significance was set at P < 0.05.

Results

Top Methods Results Discussion Author & Article Info References

Results of Discontinuing Therapy

At the time of initial re-contact in our follow-up study, 12 patients had had a total of 13 relapses, and 6 patients had not had relapse (Table 1). During the course of this follow-up study, an additional 3 relapses occurred in 2 patients. Therefore, of the 18 patients in whom azole therapy for coccidioidal meningitis was discontinued, 14 had relapse of disseminated coccidioidomycosis after discontinuation. One relapse (patient 8) was of nonmeningeal (cutaneous) disseminated coccidioidomycosis, for which fluconazole therapy was restarted; the period for which patient 8 was at risk for meningeal relapse while not receiving therapy was only 2 months, and we cannot be certain whether meningeal relapse would also have occurred if therapy had not been resumed.

Two of the 14 patients who had relapse each had two relapses of meningitis after azole therapy was discontinued. Patient 13 was treated unsuccessfully with intraventricular and intravenous amphotericin for 1 year and then with 1200 mg of ketoconazole daily for 4 years, during which time the cerebrospinal fluid culture results converted to negative, the cerebrospinal fluid antibody titers decreased, and symptoms disappeared. This therapy was discontinued, but in the following 4 months, the patient developed recurrent headaches and cerebrospinal fluid pleocytosis. Intraventricular amphotericin therapy was reinitiated but was later discontinued because of therapy-related bizarre behavior and depression. Fluconazole therapy was initiated at that time; the results are shown in Table 1. After the course summarized in Table 1, patient 3 again began receiving ketoconazole, 1200 mg/d. One year later, this patient's dose was decreased to 1200 mg every other day. Five years after the patient started his second course of ketoconazole and during the course of this study, therapy was stopped, and the patient again had relapse of meningitis 3 to 4 months later. Therefore, a total of 16 relapses have occurred in the 18 patients.

Relapse could occur early or late after therapy (range, 0.5 to 30 months). Six patients had relapse within 6 months after azole therapy was discontinued, but 5 had relapse as late as 12 months or more after therapy. Of the 4 patients receiving ketoconazole who had relapse, 3 had relapse more than 1 year after stopping therapy. Of the 6 patients receiving fluconazole who had relapse, all had relapse 7 months or less after stopping therapy. Relapses occurred in 1 of 3 women compared with 13 of 15 men (P = 0.1) and in 2 of 4 Hispanic patients compared with 12 of 14 white patients (P = 0.2).

Correlations with Relapse

Relapse occurred in all 3 patients who received oral azoles as their sole form of therapy for meningitis, but meningeal relapse also occurred in 11 of 15 patients who had received amphotericin B into the cerebrospinal fluid before receiving therapy with azoles alone (P = 0.3). The duration of amphotericin B therapy in these patients was not insubstantial; it ranged from only 3 weeks in 1 patient (who could then be maintained on azole therapy alone) to 18 years of intrathecal therapy (without cure) in another patient, before the initiation of azole therapy, which allowed discontinuation of amphotericin therapy. The patient who had received 18 years of therapy had relapse of meningitis, despite almost 4 years of itraconazole therapy.

Relapse was common with any azole. It occurred in 6 of 8 patients who received fluconazole as their only form of azole therapy, in 4 of 6 patients who received ketoconazole alone, in 2 of 2 patients who received itraconazole alone, and in 2 of 2 patients who received ketoconazole and fluconazole successively (P = 0.6).

The duration of azole therapy in these patients, all of whom were considered to have responded to azole therapy, was prolonged (median, 39 months). All 5 patients who received azoles for fewer than 24 months had relapse. The duration of azole therapy was 25, 25, 57, and 101 months in those four patients who have not had relapse, but 9 of 14 patients who had relapse also received azole therapy for more than 25 months, and 2 of the 9 received it for more than 57 months. The median duration of azole therapy was 39 months in patients who had relapse and 41 months in patients who did not.

Relapse was common regardless of the reason for stopping therapy. In 13 instances, azole therapy was discontinued because the physician, the patient, or (often) both felt that the duration of treatment was adequate. In all of these 13 cases, the physician at least assented to discontinuation. In some instances, physicians felt that the patients were cured; in others, they were influenced by previous experience with durations of azole therapy that produced relatively low relapse rates in patients with nonmeningeal disseminated coccidioidomycosis [6, 9, 17, 18]. Ten of these 13 patients had relapse after a median of 28 months (range, 8 to 101 months) of azole therapy. In this group, it is difficult to separate the influence of the patient in the decision to discontinue therapy. It is clear that, in many instances, physicians could not justify prolonging therapy to a questioning patient who had been receiving therapy for a prolonged period (usually years), was asymptomatic, and had improved cerebrospinal fluid indices or cerebrospinal fluid that was normal according to one or all criteria. Even if we disregard three instances in which the patient appeared to have been the dominant force in discontinuing therapy, the relapse rate among the remaining patients was 80% (8 of 10 patients).

Azole therapy was discontinued in four other instances (median, 43 months; range, 9 to 57 months) because patients were clearly noncompliant with continuing therapy. One of these four patients had had gynecomastia, azoospermia, and loss of libido while receiving ketoconazole therapy. Three of these four patients had relapse. In the remaining patient, the physician discontinued ketoconazole therapy at 44 months because of side effects (hypertriglyceridemia, elevated liver function test results, gynecomastia, and loss of libido). This patient also had relapse of meningitis.

We attempted to define whether other patient characteristics were associated with relapse or non-relapse. Although in a patient with active coccidioidal meningitis, the cerebrospinal fluid indices may vary according to the site of cerebrospinal fluid sampling, and although some indices may not, even in a patient whose disease has become inactive, revert completely to normal, the cerebrospinal fluid coccidioidal antibody titer, the leukocyte count, and the glucose concentration are the three indices most often used to assess response to therapy [3, 4]. As shown in Table 2, the presence or absence of normality in any of these indices or in any combination of them before discontinuation of azole therapy was not associated with success. Even when antibody titer, leukocyte count, or glucose concentration was normal at discontinuation of therapy, relapse occurred in 12 of 13, 9 of 11, and 10 of 14 patients, respectively (Table 2). Relapse occurred in 7 of the 8 patients in whom all three of these indices were normal. Only one patient (patient 6) had an underlying disease associated with predisposition to relapse (systemic lupus erythematosus treated with prednisone).

Bivariate statistical analysis did not show any clinical, laboratory, or treatment variables to be associated with therapeutic success as compared with the group of patients having relapse. The exception to this, paradoxically, was the correlation of the presence of cerebrospinal fluid antibody at the end of azole therapy with lack of relapse; this correlation just reached significance (P = 0.04). Because antibody is a marker for disease activity, this phenomenon requires further validation.

Outcomes of Relapses

Relapses were marked by abnormal cerebrospinal fluid (for example, mean leukocyte count, 372 cells/mm³; compare with leukocyte count in patients receiving therapy Table 1), plus headache (11 patients), fever (3 patients), hydrocephalus (2 patients), vomiting (1 patient), confusion (1 patient), or vertical diplopia (1 patient). The relapses summarized in Table 1 were successfully treated in 7 patients with fluconazole alone and in 1 patient with itraconazole plus intracisternal amphotericin; these patients are currently clinically stable while receiving therapy. Another patient (patient 3), whose previously described relapse was treated successfully with ketoconazole, had a second relapse when ketoconazole therapy was stopped but then became clinically stable when given fluconazole. A tenth patient who had relapse is currently clinically stable while receiving fluconazole but required ventriculoperitoneal shunting to relieve hydrocephalus exacerbated by relapse. The 9 patients mentioned who are currently receiving fluconazole include the patient who had only a nonmeningeal relapse and the 2 patients who each had two meningeal relapses after each of two apparently successful previous courses of azole therapy. The final patient with relapse in this subgroup was successfully re-treated with intrathecal and intravenous amphotericin B but was then lost to follow-up.

Patient 6 responded initially within 10 days to fluconazole re-treatment but died 17 days after therapy was restarted, in a clinical setting in which lupus coronary arteritis was considered to be the cause of death (autopsy was not done). The other two patients died as a result of complications more directly attributable to the relapse despite re-treatment with fluconazole: One had increased intracranial pressure, and the other had renal failure and electrolyte imbalance (both findings were probably due to dehydration because of unresponsiveness).

Discussion

Top Methods Results Discussion Author & Article Info References

These clinical data provide a sobering counterpoint to the promise of safe and convenient therapy for a lethal disease. Our original goal was to define predictors of relapse [19] that could guide clinicians as to when it might be safe to discontinue azole therapy. As the data accumulated, the gravity of the situation became clearer. At this time, we believe that the frequency of relapse, despite treatment for very long periods, is so high that cessation of therapy is not justified. This conclusion reprises an earlier experience with miconazole, which was the first azole shown to be effective for meningitis. Although remissions in patients with meningitis were produced with intravenous or intracerebrospinal fluid therapy or both, relapse later occurred in at least three fourths of responding patients [20]. The lack of an oral preparation meant that only short-term courses were practical, and it was hoped that the advent of oral agents and the possibility of long-term therapy might produce cure.

Our analyses do not indicate that any set of clinical conditions or combination of conditions unequivocally suggests that therapy can be discontinued. However, our power to detect differences is limited by our sample size. Some patients had their therapy discontinued by their physician, and some discontinued it on their own, with a presumption of continued good health. Our data indicate that this presumption was usually wrong.

A nineteenth patient was not included in the Results section (see above) because, strictly defined, he had a recrudescence rather than a relapse. This patient's coccidioidal meningitis was initially treated with intraventricular amphotericin plus ketoconazole, 800 mg/d, for 1 year, and was then maintained in remission with ketoconazole alone, 800 mg/d, for 9 years. At that time, because of an increase in drug price, the patient halved his dose. One year later, he developed blurred vision, headaches, and back pain. Papilledema and hydrocephalus were found, and the patient was re-treated. We have little doubt that had his treatment been discontinued after 10 years, he would have been included in this study as a case of relapse.

The data indicate that relapse can have serious consequences and that it can occur long after the discontinuation of therapy, after an interval of apparent cure. Thus, the four patients who have not had relapse cannot yet be considered cured. Two of the four have been followed while not receiving therapy for longer than the longest interval to relapse in the others, but further follow-up is needed to show whether these patients are indeed cured. Our data suggest that they are at great risk for relapse. However, after we circulated our conclusions, these patients and their physicians elected to continue observation while the patients received no therapy. If these patients are indeed cured, the cure rate with azoles may be similar to that with amphotericin B received into the cerebrospinal fluid [3, 4], although randomized comparisons have not been done.

If cure is not possible, or even if it is possible in an unpredictable few persons, the alternative to discontinuation of therapy and risk for relapse is life-long therapy with oral azoles, recognizing that this may only suppress quiescent disease. This does not seem to us an unpromising conclusion to the experience with oral therapy for coccidioidal meningitis, given the documented lethality of the disease and the problems of other therapy. In our study and in others [11-15], long-term therapy, particularly with fluconazole or itraconazole, appears to be well tolerated. Coccidioidal meningitis is one of the few infectious diseases for which lifelong therapy can be recommended. The toxicity of such prolonged therapy is, however, unknown at this time.

Although relapse after oral azoles with other forms of disseminated coccidioidomycosis has not yet appeared to be as problematic [14, 17, 18], our experience may also be relevant to other diseases for which azole therapy has shown promise, particularly cryptococcal meningitis in persons with the acquired immunodeficiency syndrome [21, 22]. Follow-up after therapy with fluconazole or itraconazole is limited, but the incidence of recrudescence during long-term follow-up in patients with this condition who continue to receive therapy [21, 23] suggests that these patients constitute another group in which cure may be rare.

Dr. Galgiani: Department of Medicine, Division of Infectious Disease, Veterans Affairs Medical Center, Tucson, AZ 85723.

Dr. Graybill: University of Texas, Health Science Center, Department of Infectious Disease, Audie Murphy Veterans Affairs Hospital, 7400 Merton Minter (111F), San Antonio, TX 78284.

Drs. Diaz and Rendon: Centro De Diagnostico, Dr. Cantu 300 Col Los Doctores, 4d. PISO, Monterrey, N.L., Mexico C.P. 64710.

Ms. Cloud: Tumor Institute, Room 153, University of Alabama Medical Center, University Station, Birmingham, AL 35294.

Dr. Stevens: Santa Clara Valley Medical Center, Department of Medicine, Division of Infectious Disease, 751 South Bascom Avenue, San Jose, CA 95128-2699.