Safety and tolerability of intrathecal liposomal amphotericin B (AmBisome) for cryptococcal meningitis: a retrospective study in HIV-infected patients

Abstract

Background:

While mortality of HIV-related cryptococcal meningitis (CM) in developed countries is relatively low, in developing countries over half of patients die within 10 weeks. Current recommended therapies are often not suitable for resource-poor settings, and new shorter regimens are urgently needed. Intrathecal administration of liposomal amphotericin B (lAmB) has shown promising results in animal models. However, the safety and tolerability of intrathecal lAmB in humans are not well known.

Methods:

In this retrospective observational study, we report the tolerability and safety of intrathecal lAmB in patients with CM from an HIV cohort study in India.

Results:

In all, 18 patients were included in the analysis. Six were female and the median age was 40 years [interquartile range (IQR): 35–45]. The median CD4 count was 42 cells/µl (IQR: 19–127). Compared with a historical control group, the hazard ratio for mortality was 0.59 (95% confidence interval: 0.26–1.29). Two patients complained of transient lumbar pain in single occasion. One patient had a skin reaction to chlorhexidine, which was used as skin disinfectant. After initial improvement, one patient requested to stop lumbar punctures for the last 2 days of treatment.

Conclusion:

Intrathecal lAmB was safe and well tolerated in HIV-infected patients with CM.

Introduction

Despite the dramatic expansion of antiretroviral treatment in developing countries, cryptococcal meningitis (CM) is still a frequent cause of death among HIV-infected patients. It is estimated that 223,100 (95% confidence interval (CI): 150,600–282,400) HIV- infected patients had CM in 2014 and 181,100 (95% CI: 119,400–234,300) died.¹ While over half of patients with CM from low- and middle-income countries die within 10 weeks of diagnosis, the mortality in developed countries is nearly 10%.²

Amphotericin B (AmB) has a strong fungicidal activity against Cryptococcus, but it has poor penetration into the cerebrospinal fluid (CSF).^3,4 To achieve higher concentration in CSF, intrathecal administration of AmB deoxycholate (AmBd) has been used for the treatment of CM, and observational studies from Asia suggest that it could be associated with improved survival.⁵ However, AmB deoxycholate has a direct irritant effect and intrathecal administration is poorly tolerated.^6,7 Lipid formulations of AmB are less irritant^8–10 and have shown promising results in animal models of CM.^11,12 In a previous study from our cohort, we reported our experience with a CM regimen using intrathecal AmB lipid emulsion (Amphomul^®).¹³ However, this lipid form of AmB is not available in many parts of the world. In this retrospective study, we aimed to describe the safety and tolerability of the intrathecal administration of liposomal AmB (lAmB) for the treatment of CM in HIV-infected patients.¹⁴

Methods

The Vicente Ferrer HIV Cohort Study (VFHCS) is a prospective cohort study of HIV-infected patients who have attended the Rural Development Trust Hospital in Bathalapalli, Anantapur District, AP, India. The VFHCS is registered at clinicaltrials.gov (No. NCT02454569). The hospital belongs to a nongovernmental organization and provides medical care to HIV-infected people free of charge. In our setting, 72% of the population live in rural areas,¹⁵ and the HIV epidemic is largely driven by heterosexual transmission and it is characterized by low CD4 cell counts at presentation, poor socioeconomic conditions and high levels of illiteracy.^16–18

The diagnosis of CM was based on the presence of Cryptococcus antigens in CSF by latex agglutination.¹⁹ Before 15 June 2013, patients were treated with the standard 2-week induction treatment with intravenous AmBd 0.7–1 mg/kg and oral fluconazole 1200 mg once daily followed by 600 mg once daily for 8 weeks (consolidation phase) and 200 mg once daily thereafter (maintenance phase).²⁰ From 15 June 2013 to 31 October 2016, the induction phase was changed to AmBd 0.7 mg/kg once daily for 7 days, intrathecal AmB lipid emulsion (Amphomul^®; Bharat Serums and Vaccines, India) once daily for 7 days, and oral fluconazole 600 mg twice daily for 14 days. Although the use of the short regimen with intrathecal Amphomul^® was associated with a significant reduction in mortality,¹³ the supply of Amphomul^® to the hospital stopped on 31 October 2016, and it was replaced by lAmB (AmBisome^®). In this study, we describe our experience with patients who received intrathecal lAmB.

For the preparation of the intrathecal administration of lAmB, a vial containing 50 mg of AmBisome^® powder was diluted with 10 ml of sterile water to achieve a final concentration of 5 mg/ml in the vial. Lumbar punctures (LPs) were performed daily during the first week of treatment. After removing at least 20 ml of CSF (typically 30–40 ml), 2 ml of the lAmB solution (10 mg) was extracted and mixed with 3 ml of 5% dextrose in a 5-ml syringe and administered intrathecally slowly over 2 min. After the LP, patients were asked to lay flat on the bed for 1 h. The rest of the AmBisome^® vial was kept in a fridge at 2–8°C and used as required for a maximum of 7 days, according to the stability information given by the manufacturer. Intravenous AmBd was diluted in 500 ml of 5% dextrose and administered slowly over 4 h.

Medical records of all HIV-infected patients with CM who received lAmB until 21 May 2017 were reviewed searching for side effects secondary to the intrathecal administration of lAmB. Patients’ attitudes towards repeated LPs were also collected. To study the mortality of the treatment compared with the standard of care, patients who received the standard 2-week induction treatment with AmBd (but no intrathecal administration of any form of AmB) were included in the control group.

Statistical analysis was performed using Stata Statistical Software (Stata Corporation. Release 14.1; College Station, TX, USA). The log-rank test was used to compare mortality during the first 10 weeks after the initiation of treatment. The study was approved by the Ethics Committee of the Rural Development Trust Hospital.

Results

We identified 18 patients who received intrathecal lAmB. Six were female and the median age was 40 years [interquartile range (IQR) = 35–45]. The median CD4 count 42 cells/µl (IQR = 19–127). Two patients complained of transient lumbar pain after the intrathecal administration of lAmB in single occasions. One patient had a skin reaction to chlorhexidine, which was used as skin disinfectant before LPs. After initial improvement, one patient requested to stop LPs for the last 2 days of treatment.

Compared with a control group of 73 patients who received the standard 2-week induction phase regimen with AmBd, there was a statistically nonsignificant trend towards improved survival during the first 10 weeks after initiation of CM treatment (hazard ratio 0.59; 95% CI: 0.26–1.29) in the intrathecal lAmB group (Figure 1). In the control group, the median CD4 count 62.5 cells/µl (IQR = 28.5–142.5). Two patients, one in each group, did not complete the 10-week follow-up period.

Figure 1.

Kaplan–Meier survival estimates of patients receiving 1-week induction treatment with intrathecal liposomal amphotericin B (lAmB) or the standard 2-week regimen with AmB deoxycholate.

Discussion

CM is still responsible for 15% of deaths in patients with AIDS.¹ The recommended therapy for CM comprises 2-week induction treatment with AmBd accompanied by flucytosine or fluconazole.²⁰ However, flucytosine is rarely available in developing countries, and AmBd is underused in resource-poor settings because long hospital admissions and close renal and electrolyte monitoring are required.²¹ Shortening treatment to 1 week could improve substantially the feasibility of the treatment for resource-poor settings because the nephrotoxicity of AmBd occurs mainly during the second week of therapy, and severe renal toxicity is rare when the treatment is shortened to 7 days or less.^22–24 Our results suggest that intrathecal administration of lAmB could help shorten AmBd treatment to 1 week without observing an increase in mortality.

Previous studies have reported that the use of intrathecal AmBd is poorly tolerated and adverse events such as leg pain, vomiting, prostration or altered mental status were common.^6,7 However, in our study, adverse events in patients who received intrathecal lAmB were mild and transient.

The study has important limitations. We observed a clinically important reduction of mortality in the intrathecal lAmB group similar to the one observed in a previous study from our cohort using another lipid form of AmB.¹³ However, differences were nonstatistically significant and patients were not randomized. Historical confounders could have influenced in the results. Moreover, the number of patients treated with intrathecal lipid forms of AmB was relatively small and from a single centre. Clinical trials are needed to confirm the survival benefits of the intrathecal administration of lAmB in CM.

Conclusion

Current pharmaceutical forms of AmB have poor penetration into CSF. The results from our cohort indicate that intrathecal administration of lAmB could help reduce the duration of the intravenous AmBd to 7 days, making the treatment more feasible for resource-poor settings. Intrathecal lAmB was safe and well tolerated by patients, and it was not associated with an increased mortality compared with the standard of care.