Safety of different amphotericin B formulations among AIDS patients with invasive fungal disease: a retrospective observational study

Abstract

We conducted a retrospective, observational study among acquired immune deficiency syndrome (AIDS) patients with cryptococcal meningitis or talaromycosis to assess AmB formulations-related adverse events (AEs). Total 205 eligible patients were enrolled. Of them, 139 received AmB therapy, 51 received liposomal AmB (L-AmB) therapy, and 15 received AmB cholesteryl sulfate complex (ABCD) therapy. The incidences of total AEs between the AmB, L-AmB and ABCD group had no significant differences. The ABCD group had significantly higher incidences of hepatotoxicity and hematological toxicity than the AmB and L-AmB groups. The incidence of grade 3–4 hematological toxicity in the ABCD group was significantly higher than that in the AmB and L-AmB groups. Multinomial logistic regression models showed that compared with AmB, ABCD had a higher risk for the occurrence of grade 3–4 hematological toxicity (aOR = 43.924, 95%CI 6.296-306.418; p < 0.001). We demonstrated that ABCD was more prone to hepatotoxicity and hematological toxicity than AmB and L-AmB among AIDS patients, which is worth noting.

Introduction

Acquired immune deficiency syndrome (AIDS) patients are one of the most susceptible groups to invasive fungal disease (IFD), with high mortality and morbidity. It is reported that about 50% of opportunistic infections among HIV-positive people are caused by fungi [1]. Cryptococcal meningitis and talaromycosis are common IFD in AIDS patients. Amphotericin B (AmB) is preferred for cryptococcal meningitis and talaromycosis by guidelines [2]. However, drug-induced toxicity may limit its clinical application. Liposomal AmB (L-AmB) has been reported to have lower incidences of adverse events (AEs) than AmB due to its chemical composition and rigorous manufacturing standards [3, 4]. In addition, AmB cholesteryl sulfate complex (ABCD) has also been used in treating IFD in China, which was described to have good efficacy and safety in the treatment of IFD in 30 patients with hematological malignancies [5]. Here we conducted a retrospective observational study to compare the drug-related AEs between AmB, L-AmB and ABCD therapy in AIDS patients with cryptococcal meningitis or talaromycosis, and the potential risks associated with the drug-related hematological toxicity were analyzed.

Methods

Between January 1, 2018, and April 30, 2023, AIDS patients who had confirmed diagnosis with cryptococcal meningitis or talaromycosis were retrospectively recruited from Zhongnan Hospital of Wuhan University, Hubei and Hunan university of medicine general hospital, Hunan. Patients whose age ≥ 18 years old and who received treatment with AmB or L-AmB or ABCD were included. The exclusion criteria were as follows: (1) discharged or died on the day of AmB/L-AmB/ABCD initiation; (2) AmB, L-AmB and ABCD were exchanged during the treatment. The regimens of AmB, L-AmB and ABCD in all patients were based on our guideline [2]. For patients with cryptococcal meningitis, AmB (0.5 ~ 0.7 mg/kg/d, North China pharmaceutical co.,ltd, China) or L-AmB (3 ~ 5 mg/kg/d, Ben Venue Laboratories Inc, USA) or ABCD (4 ~ 6 mg/kg/d, Unacon ouyi drug co., ltd, China) were injected in the induction period for at least 4 weeks, combining with flucytosine. For patients with talaromycosis, AmB or L-AmB or ABCD were injected in the induction period for 2 weeks.

AmB/L-AmB/ABCD related AEs were evaluated within 7 days after drug withdrawal. Drug-related AEs were divided to 4 grades [6, 7]: Grade 1, mild; Grade 2, moderate; Grade 3, severe; Grade 4, life-threatening; Grade 5, fatal. Hematological toxicity were graded according to the internationalwork shop on chronic lymphocytic leukemia (IWCLL) Working Group grading scale [7], and the other AEs including hepatotoxicity and nephrotoxicity were graded according to DAIDS Adverse Event Grading Tables (Version 2.1) [6]. In cases of pre-existent laboratory examination abnormalities, an increase in DAIDS or IWCLL grade was considered an drug-related AE. The judgement of AmB/L-AmB/ABCD related AEs was independently decided by two clinical experts, and when it was inconsistent, senior experts would issue a unified opinion.

SPSS 21.0 was used for data statistics. Variables were denoted as medians (IQR) or n (%). Group t-test or non-parametric rank sum test was used for the analyses of continuous variables according to Kolmogorov-Smirnov test, and chi-square test or Kruskal-Wallis rank sum test was used for the analyses of counts data. The potential risks associated with the drug-related AEs (no AEs; grade 1–2 AEs; grade 3–4 AEs) were performed by multinomial logistic regression model after the hypothesis of equal coefficients of independent variables tested by parallel lines. p < 0.05 was considered statistically significant.

Results

205 eligible patients who met the inclusion and exclusion criteria were enrolled in the study. 139 received AmB, 51 received L-AmB and 15 received ABCD therapy. Majority of patients (≥ 80%) in both three groups were males, and most patients (≥ 80%) didn’t initiate ART. The median duration on ART in the AmB, L-AmB and ABCD groups were 1.2, 2.0 and 3.5 months, respectively. The characteristics among these patients were shown in Additional file 1.

The comparisons of AmB, L-AmB and ABCD-related AEs were shown in Table 1. The incidence of infusion-related reactions was significantly lower in the L-AmB group than in the AmB (p = 0.04) and ABCD groups (p = 0.02). The ABCD group had significantly higher incidences of hypokalemia (p = 0.002; p = 0.05), hepatotoxicity (p = 0.01; p = 0.001) and hematological toxicity (p = 0.001; p = 0.002) than the AmB and L-AmB groups.

Table 1 Comparison of total AEs between the three groups

Table 2 Comparisons of grade 3–4 AEs between the three groups

The incidences of total grade 3–4 AEs (p = 0.02; p = 0.001) and grade 4 AEs (p < 0.001; p < 0.001) in the ABCD group were significantly higher than that in the AmB and L-AmB groups. The incidences of grade 3–4 hepatotoxicity had no significant differences between the three groups. The incidences of grade 3–4 hematological toxicity (p < 0.001; p < 0.001) and grade 4 hematological toxicity (p = 0.001; p < 0.001), mainly neutropenia and thrombocytopenia, in the ABCD group were significantly higher than that in the AmB and L-AmB groups. No grade 5 AEs were observed in all three groups. Multinomial logistic regression models showed that compared with AmB, ABCD had a higher risk for the occurrence of grade 3–4 hematological toxicity (aOR = 43.924, 95%CI 6.296-306.418; p < 0.001), see in Additional file 2).

Discussion

The AmB lipid-associated formulations, including L-AmB and ABCD, has been developed to solve the issue on the increased toxicity related to AmB. However, the safety of these drugs in AIDS patients haven’t been fully studied. This retrospective, observational study aimed at comparing the safety of AmB, L-AmB, and ABCD among AIDS patients with cryptococcal meningitis and talaromycosis in order to provide reference for clinical practice.

Our study found that the incidences of infusion-related reactions in the ABCD group were higher than that in the AmB and L-AmB group, which was consistent with the results of other studies [8, 9]. Up-regulation of IL-1β protein synthesis and decreasing IL-1ra levels may be responsible for the increased infusion-related toxic adverse effects among patients with ABCD therapy [10].

ABCD could be quickly absorbed by organs of reticuloendothelial system (such as liver, spleen and lung) after entering the blood, thus avoiding damage to renal tubules; therefore, nephrotoxicity was low by comparison with AmB [8, 11]. Studies conducted in patients with aspergillosis showed that ABCD may have superior renal safety compared with AmB [9, 12]. Our study observed that the incidence of ABCD-related nephrotoxicity was lower than AmB, however, there were no significant differences of the incidence of nephrotoxicity in the ABCD group by comparison with the AmB and L-AmB groups. Larger sample size studies are needed to investigate the renal safety of ABCD.

We also observed higher rates of hepatotoxicity, mainly grade 1–2 hepatotoxicity in the ABCD group compared with the AmB and L-AmB groups. A randomized, double-blind, multicenter trial comparing L-AmB with AmB for empirical antifungal therapy found no significant difference in the frequency of hepatotoxicity between L-AmB and AmB [13]. A meta-analysis on AmB formulations therapy for invasive fungal infection showed that patients receiving empirical therapy with ABCD, AmB lipid complex (ABLC), and L-AmB had higher pooled risk for developing mildly elevated liver enzymes in comparison to AmB [14]. Combined with our results, ABCD-induced hepatotoxicity should be concerned in the treatment of IFD among AIDS patients.

It’s noteworthy that the ABCD-related hematological toxicity, especially the grade 3–4 hematological toxicity including neutropenia and thrombocytopenia, in our study were significantly higher than AmB and L-AmB. Our multinomial logistic regression analysis also supported the data. The incidence of AmB related anemia was as high as 75%, which may be partly due to the suppression of erythropoietin production [15, 16]. Previous studies found that the incidences of neutropenia, hypohemoglobin and thrombocytopenia were lower in patients with L-AmB treatment compared with patients with AmB treatment, and ABLC had more obvious neutropenia and thrombocytopenia side effects [17, 18]. Although ABCD has been evaluated in limited studies and found to be safe in immunocompromised patients with IFD [9, 11, 19, 20], there is no data on ABCD-related hematological toxicity. Our data demonstrated that ABCD-related hematological toxicity including neutropenia, hypohemoglobin and thrombocytopenia should be of great attention. The mechanism of ABCD-induced hematological toxicity remains obscure, which needs more large sample sizes studies to further investigate.

In conclusion, L-AmB should be preferred for treatment in AIDS patients with IFD for its fewer AEs compared with AmB and ABCD. ABCD-related hematological toxicity, including neutropenia, hypohemoglobin and thrombocytopenia, should be particularly noteworthy.