BackgroundThe most appropriate alternative to induction therapy for HIV-associated cryptococcal meningitis (CM) remains unclear when standard treatment is unavailable, inaccessible, intolerable, or ineffective.MethodsA prospective, multi-centre cohort study was conducted to analyze the data of 156 HIV-infected patients with CM who were treated with amphotericin B deoxycholate (AmB-D) + flucytosine (5FC), voriconazole (VCZ) + 5FC, or AmB-D + Fluconazole (Flu) as induction regimens. Clinical efficacy, cumulative mortality, and adverse effects were compared among the three treatment groups.ResultsFewer deaths occurred by week 4 and week 10 among patients receiving AmB-D + 5FC than among those receiving AmB-D + Flu [4 (5.1%) vs. 8 (16.0%) deaths by week 4; hazard ratio, 1.8; 95% confidence interval [CI], 1.0 to 3.3; p = 0.039; and 8 (10.3%) vs. 14 (28.0%) deaths by week 10; hazard ratio, 1.8; 95% CI, 1.1 to 2.7; p = 0.008, respectively]. AmB-D plus 5FC was found to result in significantly higher rates of cerebrospinal fluid (CSF) culture sterility (57.6% vs. 34% by week 2; 87.9% vs. 70% by week 10; p < 0.05 for both comparisons). However, the differences in CSF culture sterility and mortality between the VCZ + 5FC group and the AmB-D + 5FC group were not statistically significant. VCZ plus 5FC had a significantly advantageous effect on the incidence of new AIDS-defining illness and length of hospital stay, compared with AmB-D plus 5FC. Laboratory adverse events (grade 3 or 4), such as severe anemia, were less frequent with VCZ + 5FC use than with AmB-D combined with 5FC or Flu use.ConclusionOur results suggest that AmB-D combined with 5FC remains the more efficacious induction regimen compared to AmB-D plus Flu, and that VCZ + 5FC might be a potential alternative when the standard regimen is not readily available, accessible, tolerated, or effective.Clinical Trials: Registration number, ChiCTR1900021195. Registered 1 February 2019, http://www.chictr.org.cn/showproj.aspx?proj=35362.