A descriptive study of vancomycin use at Red Cross War Memorial Children's Hospital, Cape Town.

Background: Antimicrobial stewardship principles guide the clinical use of antimicrobials, including vancomycin, but paediatric vancomycin prescribing practices have not been evaluated in South Africa. Objectives: To document the use, prescribing practices and monitoring of intravenous vancomycin and the spectrum of bacteria isolated on microbiological culture in children treated with intravenous vancomycin during a 12-month period at Red Cross War Memorial Children's Hospital (RCWMCH). Method: A retrospective audit of intravenous vancomycin use in children admitted to RCWMCH during 2019 was performed. Results: All 158 vancomycin prescription episodes for 143 children were included. Overall usage of intravenous vancomycin was 63 days of therapy per 1000 patient days (interquartile range [IQR]: 38-72). The median starting dose was 15 mg/kg per dose (IQR: 14-15) and median daily dose was 45 mg/kg per day (IQR: 43-60). Vancomycin was prescribed as empiric (127/158, 80%) and directed (31/158, 20%) treatment. The median duration of treatment for the directed group (7 days) was longer than the empiric group (4 days) (p = 0.001). Vancomycin serum trough concentrations were performed in 65/98 (66%) episodes where vancomycin treatment exceeded 3 days, with only 16/65 (25%) of these samples obtained before the fourth dose. Prolonged antibiotic treatment of 14 days or more was not associated with Gram-positive bacteria on culture (odds ratio [OR]: 1.02, 95% confidence interval [CI]: 0.17-4.2). Conclusion: Dosing errors, prolonged empiric treatment and inappropriate vancomycin monitoring were problems associated with vancomycin prescriptions. Contribution: The study identified multiple opportunities for improved vancomycin prescribing and monitoring. Further research and implementation of improved prescribing practices could contribute to the preservation of vancomycin as an effective antibiotic.

Candida bloodstream infection among children hospitalised in three public-sector hospitals in the Metro West region of Cape Town, South Africa.

BACKGROUND: Candida bloodstream infection (BSI) causes appreciable mortality in neonates and children. There are few studies describing the epidemiology of Candida BSI in children living in sub-Saharan Africa. METHODS: A retrospective descriptive study was conducted at three public sector hospitals in Cape Town, South Africa. Demographic and clinical details, antifungal management and patient outcome data were obtained by medical record review. Candida species distribution and antifungal susceptibility testing results were obtained from the National Health Laboratory Service database. RESULTS: Of the 97 Candida BSI episodes identified during a five-year period, 48/97 (49%) were Candida albicans (C. albicans), and 49/97 (51%) were non-C. albicans species. The overall incidence risk was 0.8 Candida BSI episodes per 1000 admissions at Red Cross War Memorial Children's Hospital. Of the 77/97 (79%) Candida BSI episodes with available clinical information, the median age (interquartile range) at the time of BSI was 7 (1-25) months, 36/77 (47%) were associated with moderate or severe underweight-for-age and vasopressor therapy was administered to 22/77 (29%) study participants. Most of the Candida BSI episodes were healthcare-associated infections, 63/77 (82%). Fluconazole resistance was documented among 17%, 0% and 0% of C. parapsilosis, C. tropicalis and C. albicans isolates, respectively. All Candida isolates tested were susceptible to amphotericin B and the echinocandins. The mortality rate within 30 days of Candida BSI diagnosis was 13/75 (17%). On multivariable analysis, factors associated with mortality within 30 days of Candida BSI diagnosis included vasopressor therapy requirement during Candida BSI, adjusted Odds ratio (aOR) 53 (95% confidence interval 2-1029); hepatic dysfunction, aOR 13 (95% CI 1-146); and concomitant bacterial BSI, aOR 10 (95% CI 2-60). CONCLUSION: The study adds to the limited number of studies describing paediatric Candida BSI in sub-Saharan Africa. Non-C. Albicans BSI episodes occurred more frequently than C. albicans episodes, and vasopressor therapy requirement, hepatic dysfunction and concomitant bacterial BSI were associated with an increase in 30-day mortality.

Hospitalization among infants who initiate antiretroviral therapy before 3 months of age.

INTRODUCTION: Studies examining hospitalization among infants with HIV in resource-limited settings, in the context of early infant diagnosis and early antiretroviral therapy (ART) initiation, are limited. METHODS: We used routinely collected data on infants who initiated ART aged <3 months (Western Cape province, South Africa; 2013-2017) to describe hospitalization from birth until 12 months post-ART initiation. Record reviews were additionally performed at three tertiary-level facilities. We used mixed-effects Poisson regression to examine factors associated with hospitalization. RESULTS: Among 840 infants, 579 (69%) were hospitalized; 36% had >1 hospitalization. Median age at ART initiation decreased from 57 days (interquartile range [IQR] 22-74; 2013-2015) to 19 days (IQR 5-54; 2016-2017). Early neonatal hospitalization (age <7 days) occurred in 271 infants (32%) and represented 24% of hospitalizations (272/1131). Overall, 443 infants (53%) were hospitalized at age ≥7 days, including 13% with hospitalizations pre-ART initiation, 15% pre and post-ART initiation and 25% post-ART initiation. Excluding early neonatal hospitalizations, initiating ART at older age vs. age <1 week was associated with higher hospitalization rates: adjusted incidence rate ratios (95% confidence interval) were 1.86 (1.31-2.64); 2.31 (1.62-3.29) and 2.47 (1.76-3.46) if ART initiation age was 1-4 weeks; 5-8 weeks and 9-12 weeks respectively. Among infants whose hospital records were reviewed, reasons for early neonatal hospitalizations mostly related to prematurity or low birthweight (n = 46/60; 77%) whereas hospitalizations at age ≥7 days were mostly due to infections (n = 206/243; 85%). CONCLUSIONS: Earlier ART initiation is associated with lower hospitalization rates. High hospitalization rates, despite initiation age <3 months, is concerning.

Balamuthia mandrillaris Granulomatous Amoebic Encephalitis: The First African Experience.

We report the first case of Balamuthia mandrillaris granulomatous amoebic encephalitis definitively acquired in Africa. Our case emphasizes initial nonspecific dermatological features, delays in confirmation of the diagnosis, difficulties accessing recommended medication, and uncertainty about optimal treatment of a disease with a frequently fatal outcome.

Xpert MTB/RIF Ultra and mycobacterial culture in routine clinical care at a paediatric hospital.

Background: Microbiological confirmation of pulmonary tuberculosis (PTB) in children is a well-documented challenge. This study evaluated Xpert Mycobacterium Tuberculosis (MTB)/Rifampicin (RIF) Ultra (Ultra) and mycobacterial cultures in routine clinical care at a tertiary paediatric hospital. Methods: Children treated for PTB and who had at least one respiratory specimen investigated by Ultra and mycobacterial culture before tuberculosis (TB) treatment was commenced were included. The findings of this retrospective study were summarised using descriptive and inferential statistics. Results: A total of 174 children were included. The median age was 2.5 years. Microcytic anaemia, airway compression, cavitary disease and miliary TB were significantly observed in children with microbiologically confirmed TB (cTB). Tuberculosis was microbiologically confirmed in 93 (53.4%) children. The positive yield from testing the first respiratory specimens was 68/174 (39.1%) on Ultra and 82/174 (47.1%) on combined Ultra and mycobacterial culture. In the subset of children (n = 70) tested with Ultra on two sequential respiratory specimens, the incremental yield from the second specimen was 30.3%. In the subset of children (n = 16) tested with Ultra on three sequential respiratory specimens, the incremental yield from the second and third specimens was 16.7% and 0.0%, respectively. When Ultra and mycobacterial culture results were combined, the incremental yield in children who had two sequential respiratory specimens tested was 24.4% and 3.1% on Ultra and mycobacterial culture, respectively. Conclusion: Ultra and mycobacterial culture on a single respiratory specimen resulted in a high microbiological yield. Ultra-testing on a second respiratory specimen increased the yield of microbiologically cTB. Additional diagnostic testing may require further study.

Increased infectious-cause hospitalization among infants who are HIV-exposed uninfected compared with HIV-unexposed.

OBJECTIVES: Increased risk of morbidity and hospitalization has been observed in children who are HIV-exposed uninfected (HEU) compared with HIV-unexposed uninfected (HUU). Studies in the era of universal maternal antiretroviral treatment (ART) are limited. DESIGN: Prospective cohort. METHODS: We investigated hospitalization between 29 days and 12 months of life in a South African cohort of infants born between February 2017 and January 2019 (HEU = 455; HUU = 458). All mothers known with HIV during pregnancy received ART. We reviewed hospital records and classified and graded infectious diagnoses using a standardized tool. We examined factors associated with infectious-cause hospitalization using mixed-effects Poisson regression. RESULTS: Infants HEU vs. HUU had higher all-cause and infectious-cause hospitalization (13 vs. 7%, P = 0.004 and 10 vs. 6%, P = 0.014, respectively). Infectious causes accounted for most hospitalizations (77%). More infants HEU were hospitalized with severe or very severe infections than those HUU (9 vs. 6%; P = 0.031). Mortality (<1%) did not differ between groups. HIV exposure was a significant risk factor for infectious-cause hospitalization [adjusted incidence rate ratios (aIRRs) = 2.8; 95% confidence interval (CI) 1.5-5.4]. Although increased incidence of preterm birth (14 vs. 10%; P < 0.05) and shorter duration of breastfeeding (44 vs. 68% breastfed for ≥3 months, P < 0.001) among infants HEU vs. HUU contributed to increased hospitalization, they did not account for all the increased risk. CONCLUSION: Infectious-cause hospitalization incidence was higher among infants HEU vs. HUU, likely partly because of higher incidence of preterm birth and lower breastfeeding rates among infants HEU. The increased infectious disease burden in HEU infants has important implications for health services in sub-Saharan Africa.

Preterm birth and severe morbidity in hospitalized neonates who are HIV exposed and uninfected compared with HIV unexposed.

OBJECTIVES: Infants who are HIV exposed but uninfected (HEU) compared with HIV unexposed uninfected (HUU) have an increased risk of adverse birth outcomes, morbidity and hospitalization. In the era of universal maternal antiretroviral treatment, there are few insights into patterns of neonatal morbidity specifically. DESIGN: A prospective cohort study. METHODS: We compared neonatal hospitalizations among infants who were HEU (n = 463) vs. HUU (n = 466) born between 2017 and 2019 to a cohort of pregnant women from a large antenatal clinic in South Africa. We examined maternal and infant factors associated with hospitalization using logistic regression. RESULTS: Hospitalization rates were similar between neonates who were HEU and HUU (13 vs. 16%; P = 0.25). Overall, most hospitalizations occurred directly after birth (87%); infection-related causes were identified in 34%. The most common reason for hospitalization unrelated to infection was respiratory distress (25%). Very preterm birth (<32 weeks) (29 vs. 11%; P = 0.01) as well as very low birthweight (<1500 g) (34 vs. 16%; P = 0.02) occurred more frequently among hospitalized neonates who were HEU. Of those hospitalized, risk of intensive care unit (ICU) admission was higher in neonates who were HEU (53%) than HUU (27%) [risk ratio = 2.1; 95% confidence interval (95% CI) 1.3-3.3]. Adjusted for very preterm birth, the risk of ICU admission remained higher among neonates who were HEU (aRR = 1.8; 95% CI 1.1-2.9). CONCLUSION: Neonates who were HEU (vs. HUU) did not have increased all-cause or infection-related hospitalization. However, very preterm birth, very low birthweight and ICU admission were more likely in hospitalized neonates who were HEU, indicating increased severity of neonatal morbidity.

South African measles outbreak 2009 - 2010 as experienced by a paediatric hospital.

INTRODUCTION: Between 2009 and 2010, South Africa experienced a major measles outbreak, with more than 18 000 confirmed cases reported to the National Institute of Communicable Diseases. METHODS: We studied measles admissions during the outbreak to Red Cross War Memorial Children's Hospital, Cape Town, between 1 November 2009 and 31 July 2010. Factors associated with mortality were retrospectively identified from notification records and hospital admissions data. Multivariate logistic regression was used to investigate potential risk factors for death. RESULTS: In total, 1 861 children were diagnosed with measles; 552 (30%) were admitted to hospital. The most common reason for admission was pneumonia (379 (68%)) and/or diarrhoea (262 (48%)). The median age at admission was 7.36 months (interquartile range (IQR) 5.0 - 10.7). The median duration of admission was 4 days (IQR 2 - 6); total hospital admission time was 3 746 days (10.3 child-years). HIV status was known in 404 (73%) children: 39/400 (14%) were HIV-infected. Eighteen children died (3% of all admissions); 15 (83%) of them were less than 1 year old. In the regression model, HIV-infection (adjusted odds ratio (aOR) 7.55, 95% confidence interval (CI) 2.27 - 25.12) and female sex (aOR 3.86, 95% CI 1.26 - 11.84) were associated with higher odds of death. CONCLUSIONS: There was a large paediatric admission burden during the 2009 - 2010 measles outbreak in Cape Town; young children were predominantly affected. HIV-infected children had a significantly higher case fatality.

Systemic amyloidosis complicating multidrug-resistant tuberculosis in childhood.

Multidrug-resistant tuberculosis is increasingly common and is associated with long diagnostic delay and high morbidity. We present a 7-year-old child who developed steroid-resistant nephrotic syndrome while receiving treatment for tuberculosis. Renal biopsy results showed systemic amyloidosis; culture of peritoneal tissue confirmed disseminated multidrug-resistant tuberculosis.

Antiretroviral therapy for children in resource-limited settings: current regimens and the role of newer agents.

WHO antiretroviral treatment guidelines for HIV-infected children have influenced the design of treatment programmes in resource-limited settings. This review analyses the latest WHO first- and second-line regimen recommendations. The recommendation to use lopinavir/ritonavir-containing first-line regimens in young children with prior non-nucleoside reverse transcriptase inhibitor (NNRTI) exposure is based on good quality evidence. Recent research suggests that lopinavir/ritonavir-containing first-line regimens should be extended to all young children, irrespective of prior NNRTI exposure. Strategies for overcoming the adverse metabolic effects of rifampicin-containing anti-tuberculosis therapy on antiretroviral therapy regimens have been under-researched in HIV-infected children, creating uncertainty about global recommendations. Preferred second-line recommendations are largely predictable. The exception is that NNRTI-containing second-line regimens are recommended for children previously exposed to NNRTIs and who subsequently did not respond to lopinavir/ritonavir-containing first-line therapy. In these patients, second-line regimens containing newer protease inhibitors (PIs) such as darunavir and tipranavir, or integrase inhibitors such as raltegravir, should be evaluated. Newer antiretroviral agents including second-generation NNRTIs and PIs, C-C chemokine receptor type 5 inhibitors, and integrase inhibitors may assist in further refinement of existing regimen options.

Primary immunodeficiencies: a 27-year review at a tertiary paediatric hospital in Cape Town, South Africa.

INTRODUCTION: The epidemiology of primary immunodeficiencies (PID) is not well documented in Africa. The objective of this study was to describe the spectrum of PID at a tertiary paediatric centre in South Africa. METHODS: A retrospective study was conducted on 168 patients diagnosed with PID from 1983 to 2009. RESULTS: Over the study period, antibody deficiencies predominated (51%) followed by well-defined syndromes (24%). Common variable immunodeficiency was the commonest antibody deficiency. The mean age of diagnosis was 51 months overall but decreased significantly to 35 months over the last 9 years. Recurrent infections were the most common presenting complaint (74%). The overall mortality rate was 25% while combined immunodeficiencies accounted for 40% of the deaths. CONCLUSIONS: The spectrum of PID in South Africa was similar to international trends. The declining mean age of diagnosis indicated improved recognition of PID. Future research should focus on identifying children with PID more effectively.

BCG Vaccination in HIV-Infected Children.

Despite the use of Bacillus Calmette-Guérin (BCG) vaccination for many years, infants and young children exposed to adults with infectious forms of tuberculosis (TB) are at high risk of developing complicated TB disease. This risk is much higher among HIV-infected children, and data on BCG protective efficacy in HIV-infected children is lacking. Recent research on BCG safety in HIV-infected infants has resulted in policy shifts, but implementation is challenging. New approaches to preventing TB among infants and children, particularly HIV-infected infants, are needed. This paper briefly reviews BCG safety and efficacy considerations in HIV-infected infants and discusses other approaches to preventing TB, including new TB vaccines and vaccination strategies.

Effect of rifampicin on efavirenz pharmacokinetics in HIV-infected children with tuberculosis.

SETTING: Rifampicin may reduce plasma efavirenz concentrations by inducing the expression of the cytochrome P450 2B6, which metabolizes efavirenz. However, there is no data in pediatric patient populations. METHODS: We measured plasma efavirenz concentrations in 15 children during and after rifampicin-based antitubercular treatment. They were receiving standard doses of efavirenz as part of antiretroviral treatment. Trough concentration (Cmin) was estimated by extrapolation of the log-linear concentration-time line to 24 hours after the previous dose. RESULTS: Wide interpatient variation and marked bimodality of efavirenz concentrations were observed. Efavirenz Cmin was not significantly different during vs. after antitubercular treatment (median 0.83 mg/L interquartile range 0.59-6.57 vs. median 0.86 mg/L interquartile range 0.61-3.56; P = 0.125). Nine (60%) and 8 (53%) children had subtherapeutic Cmin (<1 mg/L) during and after antitubercular treatment, respectively. CONCLUSIONS: Concomitant rifampicin-based antitubercular treatment was not an important determinant of efavirenz concentrations. The substantial proportion of participants with estimated Cmin <1 mg/L could result in the rapid emergence of efavirenz-resistant mutations and treatment failure.

Antiretroviral therapy during infancy: essential intervention for resource-limited settings.

EVALUATION OF: Prendergast A, Mphatswe W, Tudor-Williams G et al. Early virological suppression with three-class antiretroviral therapy in HIV-infected African infants. AIDS 22, 1333-1343 (2008). There is limited information regarding the outcome of infants treated with highly active antiretroviral therapy (HAART) in middle- and low-income countries. The mean/median age of children enrolled on HAART in these countries is generally high; however, it is acknowledged that untreated HIV-infected children under 2 years of age, including infected infants, are an extremely vulnerable group. This article assessed the findings of a recently published paper describing the outcome of 63 infants randomized to commence immediate or deferred HAART in a resource-poor setting and documented favorable short-term outcome in both groups, with correspondingly high adherence rates. The recent change in global treatment guidelines, recommending that all HIV-infected infants should be commenced on HAART soon after diagnosis, irrespective of their clinical status and/or immunological severity, is discussed in relation to the key findings of the article.

Bacillus Calmette-Guérin (BCG) vaccine-induced complications in children treated with highly active antiretroviral therapy.

OBJECTIVE: To describe the frequency, risk factors, and clinical features of bacillus Calmette-Guérin (BCG) complications in HIV-infected children treated with highly active antiretroviral therapy (HAART). METHODS: A retrospective study of children started on HAART between August 2002 and November 2004 was completed. RESULTS: Six percent (21/352; 95% CI 3.7-8.0%) developed BCG complications. All developed ipsilateral axillary lymphadenitis; one child had suspected disseminated BCG infection. There were 14 females; median age at start of HAART was 5 months. BCG disease developed a median of 34 days after starting HAART. At baseline and 6 months into HAART, the median CD4 percentage and log(10) viral load were 12.3/6.1 and 23.9/4.5, respectively. Seventeen (81%) of the patients were treated with either zidovudine or stavudine combined with lamivudine and ritonavir. Young age and high baseline viral load were independent risk factors for development of BCG complications. Mycobacterium bovis BCG was isolated in 70% of patients who underwent incision and drainage of abscesses at the vaccination site or regional lymph nodes. CONCLUSIONS: This study identified a high prevalence of BCG complications in children on HAART. A clinical case definition of BCG immune reconstitution syndrome independent of laboratory parameters for use in resource-limited settings should be developed.

Effect of rifampicin on lopinavir pharmacokinetics in HIV-infected children with tuberculosis.

OBJECTIVE: Rifampicin dramatically reduces plasma lopinavir concentrations (coformulated with ritonavir in a 4:1 ratio). A study in healthy adult volunteers showed that this reduction could be ameliorated if additional ritonavir is given. We evaluated the effect of additional ritonavir on plasma lopinavir concentrations in HIV-infected children receiving rifampicin-based treatment for tuberculosis. METHODS: We measured plasma lopinavir concentrations in 2 parallel groups receiving combination antiretroviral therapy that included lopinavir-ritonavir, with and without rifampicin-based antitubercular treatment. Additional ritonavir was given (lopinavir/ritonavir ratio of 1:1) during antitubercular treatment. Lopinavir concentrations were determined using liquid chromatography-tandem mass spectrometry. RESULTS: There were 15 children (aged 7 months to 3.9 years) in each group. Lopinavir pharmacokinetic measures (median [interquartile range]) for children with and without rifampicin, respectively, were maximum concentration (Cmax) of 10.5 [7.1 to 14.3] versus 14.2 [11.9 to 23.5] mg/L (P = 0.018), area under the curve from 0 to 12 hours (AUC0-12) of 80.9 [50.9 to 121.7] versus 117.8 [80.4 to 176.1] mg/h/L (P = 0.036), and trough concentration (Cmin) of 3.94 [2.26 to 7.66] versus 4.64 [2.32 to 10.40] mg/L (P = 0.468). Thirteen of 15 children receiving antitubercular treatment (87%) had a lopinavir Cmin greater than the recommended minimum therapeutic concentration (1 mg/L). CONCLUSIONS: The effect of rifampicin-based antitubercular treatment on lopinavir concentrations was attenuated by adding ritonavir to rifampicin. Although the median Cmax and AUC0-12 were lowered by 26% and 31%. respectively, the Cmin was greater than the minimum recommended concentration in most children.

High prevalence of subtherapeutic plasma concentrations of efavirenz in children.

OBJECTIVE: Efavirenz-containing regimens using concentration-controlled dosing have been shown to provide potent antiretroviral activity in children. In many settings, concentration-controlled dosing is not available. In this study, efavirenz plasma concentrations were evaluated in South African HIV-infected children receiving efavirenz-based antiretroviral treatment. METHODS: Three consecutive blood samples were drawn between 12 and 24 hours after dosing in 15 HIV-infected children receiving the recommended daily doses of efavirenz. Validated liquid chromatography tandem mass spectrometry methods were used to determine plasma levels of efavirenz. The trough concentration (Cmin) of efavirenz was estimated by extrapolation of the log-linear regression line of the 3 concentration versus time points to 24 hours. RESULTS: The estimated Cmin was <1 mg/L in 6 (40%) of the children. Three of the 5 children with detectable viral loads had low efavirenz concentrations. Marked bimodality in efavirenz concentrations was observed. CONCLUSIONS: Our findings, together with those of previous studies, indicate that many children dosed according to the current guidelines do not achieve adequate efavirenz exposure. Because low efavirenz concentrations are associated with the rapid emergence of efavirenz-resistant mutations and treatment failure, the current recommended efavirenz doses should be re-evaluated, especially in developing countries, where therapeutic drug monitoring is seldom available.

Comparison of diagnostic criteria of tuberculous meningitis in human immunodeficiency virus-infected and uninfected children.

INTRODUCTION: Tuberculous (TB) meningitis is sometimes difficult to diagnose in young children. The decision to start anti-TB treatment of TB meningitis is usually made on clinical grounds and results of special investigations, such as cerebrospinal fluid examination and cranial computerized tomography (CT), because bacteriologic yield is low and the results delayed. AIM: To determine whether the clinical, laboratory, and radiologic criteria used in the diagnosis of TB meningitis in human immunodeficiency virus (HIV)-uninfected children apply to HIV-infected children. METHODS: Retrospective, case-control study. Clinical, laboratory, and radiologic features of TB meningitis were compared in 34 HIV-infected and 56 HIV-uninfected patients matched for age and stage of TB meningitis. RESULTS: All clinical differences found between the 2 groups at admission were related to the underlying HIV disease. Neurologic presentation and cerebrospinal fluid findings at admission did not differ significantly between the 2 groups. Significantly more HIV-infected than HIV-uninfected children had evidence of TB on chest radiography. The classic CT signs of TB meningitis (obstructive hydrocephalus and basal enhancement) were significantly less prominent in the HIV-infected group (P < 005). CONCLUSION: The diagnostic criteria for clinical diagnosis of TB meningitis apply to HIV-infected children. However, cranial CT findings in this group may be misleading and delay the diagnosis of TB meningitis.

Progressive multifocal leukoencephalopathy after initiation of highly active antiretroviral therapy in a child with advanced human immunodeficiency virus infection: a case of immune reconstitution inflammatory syndrome.

Five weeks after commencing highly active antiretroviral therapy, a 12-year-old boy with advanced human immunodeficiency virus infection presented with acute cerebellar dysfunction and hemiparesis. Progressive multifocal leukoencephalopathy was diagnosed by cerebrospinal fluid polymerase chain reaction for JC virus and magnetic resonance imaging of the brain. Rapid and sustained improvement followed a prolonged course of glucocorticosteroid therapy while continuing antiretrovirals.