|
|
| |||||||||||||||||||||||||||||||
Background In patients with the acquired immunodeficiency syndrome (AIDS), toxoplasmic encephalitis is usually a presumptive diagnosis based on the clinical manifestations, a positive antitoxoplasma-antibody titer, and characteristic neuroradiologic abnormalities. A response to specific therapy helps to confirm the diagnosis, but it is unclear how rapid the response should be. We studied the course of patients treated for acute toxoplasmic encephalitis and evaluated objective clinical criteria for this empirical diagnosis.
Methods A quantifiable neurologic assessment was used prospectively to evaluate the clinical outcome of patients with AIDS and toxoplasmic encephalitis who were treated with oral clindamycin (600 mg four times a day) and pyrimethamine (75 mg every day) for six weeks.
Results Thirty-five of 49 patients (71 percent) responded to therapy, and 30 of these (86 percent) had improvement by day 7. Thirty-two of those with a response (91 percent) improved with respect to at least half of their base-line abnormalities by day 14. Improvement in neurologic abnormalities within 7 to 14 days after the start of therapy was strongly associated with the neurologic response at 6 weeks. The four patients in whom treatment failed and the two patients with lymphoma had progressing neurologic abnormalities or new abnormalities during the first 12 days of therapy. Nonlocalizing abnormalities (headache and seizure) improved regardless of the clinical outcome.
Conclusions Oral clindamycin and pyrimethamine are an effective treatment for toxoplasmic encephalitis. Patients who have early neurologic deterioration despite treatment or who do not improve neurologically after 10 to 14 days of appropriate antitoxoplasma therapy should be considered candidates for brain biopsy.
Treatment is usually begun on the basis of a presumptive diagnosis, and the standard therapy is the combination of pyrimethamine and sulfadiazine1. New treatment regimens are needed because of the frequency of adverse reactions (more than 40 percent) to sulfonamides in patients with AIDS2,3. Recently, a prospective study demonstrated levels of efficacy and toxicity for clindamycin (1200 mg four times a day, given first intravenously, then orally) and oral pyrimethamine (75 mg per day) comparable to those for standard therapy with pyrimethamine and sulfadiazine4.
After the start of specific antitoxoplasma chemotherapy, a clinical and radiologic response to therapy is sought to confirm the diagnosis. An important difficulty with this approach is the possibility of an inaccurate diagnosis and thus of undue delay in pursuing an alternative diagnosis. In this pilot study, we examined the effectiveness and safety of oral clindamycin (600 mg four times a day) and pyrimethamine (75 mg per day) as therapy for acute toxoplasmic encephalitis. In addition, we evaluated a quantifiable neurologic assessment using objective criteria to assess the time to a response to empirical chemotherapy for toxoplasmic encephalitis.
Methods
Study Design
This international, multicenter, open-label phase 2 study of the combination of oral clindamycin plus pyrimethamine used a quantifiable-neurologic-assessment tool to determine the natural history of treated toxoplasmic encephalitis. Data on clinical responses were correlated with base-line measurements and radiographic examinations. The duration of the study treatment was six weeks.
Recruitment and Eligibility of Patients
Patients who were known to be infected with the human immunodeficiency virus (HIV) or who belonged to a group at high risk for HIV infection and who also had a neuroradiographic abnormality compatible with the diagnosis of toxoplasmic encephalitis5 were considered for enrollment at 1 of 14 sites in the United States, 2 in France, and 1 in Switzerland. Patients with biopsy-proved or relapsed disease were also eligible. The presence of antitoxoplasma antibody in the serum of all patients was determined by the Sabin-Feldman dye test at the Palo Alto Medical Foundation, Palo Alto, California4. Those found to be negative were excluded from the analysis of treatment efficacy. The criteria for exclusion also included a score of 20 or less on the Karnofsky performance scale,6 a history of hypersensitivity to the study drugs, known concomitant infection of the central nervous system or central nervous system lymphoma, a positive test for cryptococcal antigen in serum, and the inability to take oral therapy.
Study Therapy
All the patients received a 200-mg loading dose of pyrimethamine followed by 75 mg per day by mouth, with 10 mg per day of leucovorin. The dose of clindamycin was 600 mg orally every six hours. Therapy was administered for six weeks in the absence of adverse events or treatment failure requiring discontinuation of the study drugs.
Care of Patients
The patients were hospitalized for the first week of the study and evaluated daily by selective history-taking and physical examination. Thereafter, they were evaluated weekly or more often, as medically appropriate. The patients were assessed neurologically at entry into the study, on days 3 and 7, and then weekly for six weeks. If physical examination showed a patient's condition to be deteriorating on any day during the first week, the neurologic assessment was repeated more frequently. Blood was obtained at base line for laboratory tests, including a chemistry profile, hematologic studies, CD4 and CD8 T-lymphocyte counts, serologic tests for toxoplasma and syphilis, and tests for cryptococcal antigen in serum. Chemistry profiles and hematologic studies were repeated weekly or as medically indicated. Neuroradiographic scans using computed tomography (CT) or magnetic resonance imaging (MRI) were performed at base line and repeated at weeks 3 and 6. If patients had clinical progression of disease after 7 days of therapy or unchanged clinical status after 14 days, repeat scans were performed.
The use of dexamethasone was permitted only for the management of intracranial hypertension or mass effect. Zidovudine use was discontinued during the study. Adverse reactions were scored with the standardized toxicity table of the National Institute of Allergy and Infectious Diseases (NIAID). The management of less severe toxic effects included increasing the dose of leucovorin from 10 mg to 20 mg per day in the case of hematologic abnormalities and giving transfusions to maintain hemoglobin levels above 7.0 g per deciliter.
Neurologic Assessment
The neurologic assessment involved 26 signs and symptoms (Table 1), each of which was listed with objective criteria that allowed the investigator to assess the patient's neurologic status at each evaluation point. For each sign or symptom, normal was assigned a value of 0, and numbers from 1 to as high as 6 were used to quantify the level of severity. Responses in three categories (language fluency, language comprehension, and repetition) were assessed as either normal or abnormal.
|
Neurologic responses and final clinical outcomes were determined at the completion of the study therapy, either at six weeks in the case of patients completing the study or earlier in the case of those who discontinued study therapy because of voluntary withdrawal, progression of disease, or toxicity (Table 2). A patient with a neurologic response was defined as one who had improvement by one or more levels with regard to at least half the neurologic abnormalities that had been identified in that patient at base line (ignoring signs and symptoms that were not recorded or could not be evaluated), provided that there was no worsening of any of the other base-line abnormalities and no development of new abnormalities that persisted for at least two consecutive examinations. A patient who did not meet this definition was considered not to have had a neurologic response. By following the course of each patient's abnormalities, we were able to determine when the patient had a neurologic response. This information was used in the analysis of the time required to respond to therapy and in the evaluation of prognostic variables. The patient's initial improvement was also quantified and was defined as a decrease in severity by at least one level from the base-line value with regard to at least one abnormality. Alternatively, patients were considered to have progressing disease when the neurologic assessment showed that at least one abnormality had increased in severity by at least one level and when the increased severity persisted for two consecutive examinations.
|
|
The association between different types of response was evaluated with odds ratios and 95 percent confidence intervals. Differences in response rates were evaluated by logistic regression and Fisher's exact test. Time-to-response curves and medians were estimated according to the Kaplan-Meier method. Differences in the distribution of the time to a response were evaluated by the log-rank test.
Results
Patients
Between December 1989 and December 1990, 56 patients (49 from the United States, 5 from France, and 2 from Switzerland) were enrolled. Seven patients were excluded from the analysis of efficacy; six had negative serologic tests for toxoplasma, and one never started treatment. The majority of the patients were men (88 percent) and the median age was 35 years. Twenty-nine of the 49 patients included in the study (59 percent) had a previously diagnosed AIDS-related illness, and 3 of these 29 had had toxoplasmic encephalitis, with a median interval of 369 days (range, 6 to 2902) since the first diagnosis. Sixteen patients (33 percent) were receiving antipneumocystis prophylaxis -- aerosolized pentamidine in 14 and trimethoprim-sulfamethoxazole in 2. The Karnofsky scores of 29 patients (59 percent) were 60 or less. Among the 41 patients for whom CD4 and CD8 counts were reported, the median counts were 34 (range, 2 to 269) and 470 (range, 67 to 1504) cells per cubic millimeter, respectively; 24 of these patients (59 percent) had CD4 counts below 50 per cubic millimeter.
Clinical Response to Study Therapy
There were 35 patients with a response (71 percent), 12 patients who had no response (24 percent), and 2 patients whose response was unknown because no neurologic assessment was made after the base-line assessment (4 percent) (Table 2). Eighteen (51 percent) of those with a neurologic response had initial improvement by day 3, 30 (86 percent) by day 7, and 32 (91 percent) by day 14. Furthermore, 26 patients (74 percent) had improvement with regard to half of their base-line abnormalities by day 7 of therapy, and 32 (91 percent) had such improvement by day 14 (Figure 1). Among those with a neurologic response, 9 patients had a complete recovery and 20 had improvement with respect to base-line neurologic abnormalities, but without a complete resolution at six weeks (Table 2). The remaining six patients had neurologic responses but discontinued therapy early because of toxicity or voluntary withdrawal from the study.
|
The neurologic assessments of the four patients who did not respond to treatment revealed worsening of base-line abnormalities by day 5 in two and by day 12 in the remaining two. Both patients whose disease progressed during the first week of therapy were receiving corticosteroids. Only one underwent biopsy and was found to have toxoplasmic encephalitis on histologic examination. Although there was waxing and waning of base-line abnormalities during the first week of therapy, each of the patients who did not respond to treatment and each of those with lymphoma had progression of some base-line abnormalities and the development of new abnormalities after 5 to 12 days of therapy. One patient, in whom treatment failed after day 12, died of documented toxoplasmosis five days after therapy was discontinued. Two patients were switched to pyrimethamine and sulfadiazine, and they responded to therapy. The final patient who did not respond to treatment improved clinically after receiving a higher dose of intravenous clindamycin and oral pyrimethamine.
Radiologic Response
Base-line brain scans were available for review by independent neuroradiologists for 45 patients (CT in 35 and MRI in 10), and all revealed focal lesions. The remaining four scans were unavailable for independent review, although all four patients were reported by the local institutions to have had focal lesions. Of these 45 scans, 31 (69 percent) showed multiple lesions (median, 3; range, 2 to 11), and 14 (31 percent) showed single lesions. Of the 14 single lesions, 10 (71 percent) were detected by CT and 4 (29 percent) by MRI. The overall radiologic response rate was 76 percent (37 of 49 patients who had complete, good, or partial responses).
To determine whether the neurologic assessment could be used as a tool to evaluate the response to therapy reliably, we compared the clinical and radiologic responses at week 3 and at the end of therapy. At week 3, 28 patients had radiologic responses (1 complete, 19 good, and 8 partial), whereas at the end of therapy 37 patients had such responses (8 complete, 13 good, and 16 partial). The association between the neurologic and radiologic findings was stronger at the end of therapy (odds ratio, 14.9; 95 percent confidence interval, 2.9 to 77.6) than at week 3 (odds ratio, 0.88; 95 percent confidence interval, 0.15 to 5.3). Of the patients with neurologic responses, 32 (91 percent) had improvement that was detectable radiologically (Table 3). Odds ratios were computed by collapsing the radiologic responses into the categories of complete, good, and partial and comparing them with the others. At the end of therapy, the radiologic responses in the four patients who did not respond to treatment were as follows: mixed, no response, progression of disease, and unknown response.
Time to Response and Prognostic Indicators
The neurologic assessment was divided into five neuroanatomical categories (Table 1) to examine patients' times to a response. The median time needed to achieve 50 percent improvement in base-line neurologic abnormalities was three days for all neuroanatomical categories, except those abnormalities attributable to focal cortical causes; for these the median time to response was six days. The patients who responded to therapy had responses in all neuroanatomical categories, whereas those who did not respond either did not improve or had deterioration in all categories except that including nonlocalizing abnormalities (i.e., headache and seizure); these improved regardless of the ultimate response. There was no significant difference in the response rate (P = 0.56 by Fisher's exact test) or in the time to a response (P = 0.36 by the log-rank test) among the 14 patients who received corticosteroids. CD4 count, severity of symptoms at base line, and Karnofsky score were not significantly associated with the time to response. The only factors approaching a statistically significant association with neurologic response were the number of focal lesions (P = 0.05) and the number of days since the AIDS-related diagnosis (P = 0.07).
Natural History of Treated Toxoplasmic Encephalitis
The abnormalities noted in at least half the patients at base line are shown in Table 4. Ambulation was abnormal in 73 percent of patients at base line. Nineteen patients (39 percent) presented with abnormal mental status, but it was considered severe (with the patient lethargic, requiring frequent arousal, or stuporous) in only three. Some residual neurologic problems persisted at the end of the six-week treatment period in patients who had a partial clinical response, but they were mild (level 1 or 2), except in two patients who had level 3 abnormalities in ambulation and gait.
|
Adverse Events
Among the eight patients with dose-limiting toxic effects, six (75 percent) had clinical improvement within one week after discontinuing therapy. The dose-limiting toxic effect in seven patients was rash; the eighth patient had neutropenia. One patient, excluded from the study because of negative serologic tests for toxoplasma, had grade 4 hepatitis during therapy.
Discussion
In this prospective study, the combination of oral clindamycin and pyrimethamine for the treatment of toxoplasmic encephalitis was effective. We generated a comprehensive, ordered, quantifiable neurologic assessment that had the sensitivity to assess not only the ultimate response to treatment but also the actual time required for the response to occur. In so doing, we established that the response to this combination of drugs can be both rapid and unequivocal and that pyrimethamine and clindamycin are a satisfactory alternative to pyrimethamine and sulfadiazine for patients who cannot tolerate sulfonamides. These findings contribute to those of others who found that regimens of pyrimethamine combined with a higher dose of intravenous clindamycin or with oral clindamycin have efficacy comparable to that of pyrimethamine and sulfadiazine4,7.
We examined a variety of potential prognostic factors to determine whether they affected either the neurologic response or the time to a response to therapy. The base-line severity of the neurologic abnormalities did not significantly affect the time to response, although patients with severe focal cortical abnormalities (e.g., in motor strength and language and speech) were slower to have resolution and were more often left with residual deficits. Among the 20 patients with a response who did not have complete resolution of base-line signs and symptoms, 6 had a waxing and waning course in which normal neurologic assessments were achieved after the start of therapy, only to be interrupted intermittently by abnormalities of memory, recall, and language fluency. Whether these abnormalities were due to toxoplasmic encephalitis or to concomitant HIV encephalitis is unknown. Headache and seizure tended to be nonspecific indicators of a response to therapy, since these abnormalities resolved both in those with a response and in those without a response. This was probably due to the use of palliative medications. The patients receiving corticosteroids did not differ significantly in their time to a response from those not receiving corticosteroids. This may be attributable to the fact that steroids were reserved only for the treatment of patients with evidence of elevated intracranial pressure.
Standard clinical practice calls for treatment of toxoplasmic encephalitis to be initiated on presumptive diagnosis when a characteristic neuroradiologic abnormality is noted. The clinical diagnosis is confirmed by the occurrence of a clinical and radiologic response to therapy1. This clinical practice has evolved with the advent of AIDS because of the morbidity associated with brain biopsy, the inability of the clinician to perform biopsies of all the numerous and often inaccessible lesions, and the reluctance of many neurosurgeons to perform brain biopsy. Since toxoplasma is the most common opportunistic pathogen of the brain in patients with AIDS, the practice of presumptive therapy in patients who have a characteristic finding on CT or MRI and a positive serologic test for toxoplasma is widely accepted. With these criteria, the predictive value in one study was estimated to be as high as 80 percent8. However, for populations of patients such as intravenous drug users in whom other central nervous system processes are more prevalent, the predictive value of a positive serologic test for toxoplasma may be diminished9. Furthermore, with the widespread use of trimethoprim-sulfamethoxazole for antipneumocystis prophylaxis,10 the predictive value of these indicators may be further reduced.
Guidelines for establishing a diagnosis on the basis of empirical therapy require that data be available on the time to response. It was suggested, on the basis of anecdotal and retrospectively collected information, that patients who were treated empirically should have a clinical response within 14 to 21 days and a radiologic response in all lesions within 3 weeks11. In this study, we demonstrated that more than half of patients who ultimately responded to therapy had initial neurologic improvement by day 3 of therapy, and 86 percent had such improvement by day 7. Also, for 74 and 91 percent of the patients with a response, more than half of their base-line neurologic abnormalities clearly improved by days 7 and 14 of therapy, respectively. In contrast, the base-line signs and symptoms of all the patients who ultimately did not respond or had lymphoma worsened, and new abnormalities appeared between day 5 and day 12 of therapy. The correlation between the neurologic assessment and radiographic results was particularly good at the end of therapy. Therefore, for patients whose condition deteriorates early in the course of therapy or whose base-line abnormalities (except headache and seizure) do not improve within the first 10 to 14 days of appropriate treatment, brain biopsy should be seriously considered, with or without a change in therapy. Particularly close monitoring may be important for patients with a solitary lesion on MRIs who are more likely to have lymphoma than toxoplasmic encephalitis12 or who are receiving trimethoprim-sulfamethoxazole for pneumocystis prophylaxis. A more rapid time to empirical diagnosis will allow earlier initiation of therapy in patients who do not have toxoplasmic encephalitis.
Supported in part by grants (UO1AI-131808, to Dr. Luft and Ms. Bosler; and AI-04717, to Dr. Remington) from the U.S. Public Health Service.
Source Information
From the State University of New York at Stony Brook (B.J.L., E.M.B., J.F.); the Division of AIDS, National Institute of Allergy and Infectious Diseases, Bethesda, Md. (R.H.); Harvard School of Public Health, Boston (A.H.K., D.D.B.); Groupe Hospitalier Bichat-Claude Bernard, Paris (C.L., J.-L.V.); Los Angeles County-University of Southern California Medical Center, Los Angeles (D.A.); University of Miami, Miami (R.U.); Mount Sinai Medical Center, New York (J.J.); Hopital Pellegrin, Bordeaux, France (P.M.); and the Palo Alto Medical Foundation, Palo Alto, Calif. (J.S.R.) The contributing members of the Agence Nationale de Recherches sur le SIDA (ANRS) of the Institut National de la Sante et de la Recherche Medicale, the AIDS Clinical Trials Group (ACTG) of the National Institute of Allergy and Infectious Diseases, and the independent investigators are listed in the Appendix.
Address reprint requests to Dr. Luft at the Division of Infectious Diseases, State University of New York at Stony Brook, Stony Brook, NY 11794.
References
The following persons and institutions participated in the development, implementation, or analysis of the ACTG 077p/ANRS 009 trial.
Frontier Science and Technology Research Foundation, Buffalo, N.Y. -- C. Phillips; Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland -- M. Glauser and J. Chave; Washington University School of Medicine, St. Louis -- D. Clifford, W. Powderly, and M. Klebert; Stanford University, Stanford, Calif. -- R. DeLaPaz and D. Israelski; Memorial Sloan-Kettering Cancer Center, New York -- D. Armstrong and M. Binion; University of California, San Diego -- S. Spector and C. McIvor; University of Miami, Miami -- J. Reese; University of Pittsburgh, Pittsburgh -- M. Ho and P. Horn; University of Southern California, Los Angeles -- J. Leedom, D. Diamond, and Y. Barrandy; Tulane University, New Orleans -- R. Clark and D. Greenspan; Bronx Veterans Affairs Hospital, New York -- N. Ostrow and D. Tomesch; Bronx Municipal Hospital Center, New York -- R. Soeiro and P. Edwards; University of Massachusetts, Worcester -- S. Cheeseman and C. Andrews; State University of New York at Stony Brook -- E. Cantos, R. Steigbigel, and R. Burk; St. Luke's-Roosevelt Hospital, New York -- M. Grieco and B. Kolatch; Duke University, Durham, N.C. -- H. Waskin and K. Shipp; and Cornell Medical Center, New York -- M. Giordano and B. Greenhill.
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
