Early initiation of antiretroviral therapy (ART) has been found to decrease mortality from a number of opportunistic infections (OIs) associated with HIV infection, such as Pneumocystis pneumonia and tuberculosis (TB) (1). However, whether early ART initiation provides a similar clinical benefit for HIV-positive patients with cryptococcal meningitis (CM) remains unclear. For HIV patients with CM, the challenges of early ART initiation include adverse reactions, overlapping toxicities of ART and antifungal therapy, and the development of immune reconstitution inflammatory syndrome (IRIS) (2, 3). Deferred initiation of ART may increase the risk of disease progression, which could lead to death. Clinicians must therefore tread a very fine line when choosing the precise timing of ART initiation for these patients.

The Infectious Diseases Society of America's 2010 guidelines recommend that ART in patients with CM should be started between 2 and 10 weeks after antifungal therapy has been initiated (4, 5). However, it is not clear precisely when to start ART during the prescribed period of between 2 and 10 weeks. In affluent countries, some experts may delay ART by up to 10 weeks (i.e., after completion of consolidation fluconazole therapy) in patients who have close follow-up and access to treatment that may prevent further OIs. A prospective randomized controlled trial in Zimbabwe observed that patients who received ART within 72 h of diagnosis of CM had a mortality rate of 88%, compared to 54% in those who delayed ART initiation to 10 weeks after initiation of antifungals, with all participants receiving 800 mg of fluconazole daily for CM induction therapy (6). The deferred-ART strategy minimizes the risk of development of paradoxical IRIS and drug-drug interactions between ART drugs and high-dose antifungal therapy. However, this approach is not universally applicable since the benefits of a more rapid immune recovery in resource-limited areas usually outweigh the risk of development of IRIS (7).

The classic Cryptococcal Optimal ART Timing (COAT) prospective study showed that deferring ART initiation until 5 weeks after the diagnosis of CM, was associated with significantly improved survival as compared with initiating ART at 1–2 weeks (8). One early study found that compared with deferred ART (median time to ART initiation was 32 days), early ART (median time to ART initiation was 7 days) was not associated with improved cerebrospinal fluid (CSF) fungal clearance, and resulted in a higher risk of development of IRIS (3). Additionally, observations from meta-analytic data sets have also indicated a higher probability of survival among people who initiate ART until 4 weeks after CM diagnosis (9, 10). Thus, for individuals in resource-limited settings, ART should be initiated between 4 and 6 weeks after the initial diagnosis of CM, because the benefits of a faster immune recovery usually outweigh the risks associated with development of IRIS. This approach is consistent with guidelines from the World Health Organization (7). However, the 2018 guidelines from the International Antiviral Society-USA panel suggest an earlier initiation of ART in these patients under certain circumstances (11).

The optimal timing of ART initiation thus remains somewhat controversial for HIV patients with CM. We have conducted this multicenter prospective study in order to determine the optimal timing for ART initiation in ART-naive HIV-infected patients newly diagnosed with CM in China.

Currently, there remains uncertainty regarding the optimum timing of ART initiation in HIV-infected patients presenting with CM. The uncertainty relates to determining precisely when to initiate ART after a diagnosis of CM, and involves balancing the survival benefits conferred by ART against the risk of developing IRIS. Previous studies in mixed patient cohorts or non-randomized studies have suggested that early ART treatment could decrease mortality and possibly improve clinical outcomes (1, 3). However, a systematic review which included subjects from the United States, South Africa, Zimbabwe, and Puerto Rico found insufficient evidence to support either early or deferred initiation of ART (14). In another large Latin-American cohort study, there also appeared to be no statistically significant difference in the probability of survival between patients who started ART within 2 weeks of CM vs. those initiating ART between 2 and 8 weeks after CM diagnosis (15).

In contrast to these findings, a recent review of four randomized controlled trials suggested a higher risk of mortality among people who initiate ART within four weeks of CM diagnosis (10). Of these four trials, a trial from Zimbabwe showed decreased survival in individuals with HIV/CM who were started on ART within 72 h vs. after 10 weeks from initial CM treatment (6). A classic trial conducted in Uganda and South Africa found that patients randomized to start ART within 2 weeks after diagnosis of CM had a higher risk of mortality compared to those assigned to deferred ART until 5 weeks after diagnosis of CM (8). However, both of the preceding studies were terminated early after observing increased mortality in the early ART arm, and may thus have overestimated the size of the effect of early ART on mortality (16). Another clinical trial, showing the opposite effect, found fewer deaths and AIDS progression events in those treated with early vs. deferred-ART (1). Conversely, a study conducted in Botswana failed to demonstrate a reduction in mortality with early ART.

In the present study, data from the intention-to-treat analysis found that initiating ART between 2 and 5 weeks of CM treatment showed a trend toward decreased survival in patients with HIV-associated CM, as compared with deferring ART until 5 weeks after the start of CM therapy. Moreover, the outcomes comparing early initiation of ART (<4 weeks after starting antifungal treatment) with deferred initiation of ART (6 weeks after starting antifungal treatment) suggested that earlier ART was indeed detrimental to the outcomes of patients with concurrent CM and HIV infection; this association reached statistical significance in the subgroup analysis for the primary outcome. Possible reasons for the detrimental effects of early ART in these participants include progression of cryptococcal disease, inadequate CSF pressure management, combined drug interactions, and additive toxic drug effects (17). Also, since neurological deterioration often occurred unexpectedly, it was rarely possible to perform appropriate investigations in order to establish the cause for the deterioration. We thus postulated that the higher short-term mortality observed with early ART initiation may relate to adverse consequences in the CNS. Potential causes are likely to involve either occult paradoxical immunological mechanisms, overt CNS IRIS, or the unmasking of a different comorbid CNS opportunistic infection, each occurring in the context of early initiation of ART.

More than half of the patients in research settings had succumbed by 10 weeks after starting antifungal therapy, even among patients primarily treated with amphotericin-based therapy. The results emphasize the high acute mortality in patients with CM; these acute deaths were attributed mainly to CM or complications related to CM. The findings observed in the present study are consistent with what has been observed in other studies and cohorts (6, 8). Importantly, in our subgroup analysis, between-group differences were found in the early-ART group compared to the deferred-ART group with respect to the rate of mortality during the 5-10 weeks after enrolment. The results of analyses performed in the per-protocol population were also consistent with this result. The risk factors for mortality, in addition to high fungal burden, were slow clearance of infection, altered mental status, older age, low CD4+ T-cell count, low weight, and anemia, which have all been previously identified as predictors of mortality in HIV cohorts in general (18). Cell-mediated immunity has been observed to play an important role in the control of Cryptococcus infection; low CSF white cell counts and low levels of IFN-γ, TNF-α, and IL-6 were shown to be inversely associated with survival (19). The higher short-term mortality trend observed with early ART initiation in our trial may thus be related to a poor cellular inflammatory responses at the site of infection in the CNS.

The early-ART group appeared to be associated with a higher frequency of severe (grade 4) adverse events, suggesting that early ART initiation may actually be disadvantageous in terms of adverse events. This increase in adverse events could not be accounted for by differences in ART regimens used, as the differences in adverse events at baseline were similar between the two groups. Severe anemia appeared to occur more commonly in the early ART group; however, this association was calculated to not be statistically significant. The adverse consequences related to early ART initiation after cryptococcosis diagnosis may involve unrecognized increases in inflammatory responses in the CNS due to ART-associated immune reconstitution in the first month after ART initiation. Interestingly, we observed no significant increase in specific neurological events in the early ART group. These results concur with those of a past randomized controlled study that compared early-vs. deferred-ART in patients with HIV-associated tuberculous meningitis (20). In contrast to these results, however, early ART initiation is considered to be beneficial in other non-central nervous system (non-CNS) opportunistic infections, and has been demonstrated to reduce AIDS-defining events and deaths on these cohorts (1, 21). Thus, the consequences of early ART initiation in patients with CNS infections, such as cryptococcal and tuberculous meningitis, seem to differ from those of early ART initiation in non-CNS opportunistic infections.

Early initiation of ART most likely resulted in the increased rates of CM-IRIS and the excess mortality observed in the early ART group. Interestingly, however, there was no statistically significant increase in paradoxical CM-IRIS incidence in the present study after early ART initiation (6.4%) when compared to deferred ART (7.3%) (p = 0.872). Moreover, the overall incidence of IRIS was low, despite most subjects having an advanced immunodeficient immunological state. Our observation that the timing of ART initiation is positively associated with the incidence of CM-IRIS is consistent with some (1, 22), but not other (23–25) studies. One possible reason for this is that the case definition of paradoxical CM-IRIS relies on the requirement for previous clinical improvement of CM (13). In our study, clinicians were often unable to accurately distinguish whether progressive clinical deterioration in the early ART group was secondary to CM complications or to CM-IRIS, which might have resulted in a degree of ascertainment bias, with under-detection of early IRIS events. In contrast, the CM-IRIS case definition performed well for later-occurring IRIS events, in cases where the patients experienced clinical improvement, started to receive antiretroviral treatment, and subsequently showed clinical deterioration. In addition, participants who succumbed in both the early and the deferred ART groups were more likely to have developed IRIS, and therefore our calculations of CM-IRIS incidence may not be precisely accurate.

The risk factors associated with CM-IRIS development include an advanced state of immunosuppression (CD4+ T-cell count less than 50 cells/μL, CD4/CD8 ratio <0.15), high cryptococcal fungal burden at ART initiation, rapid decrease in viral load after ART initiation, and early ART initiation (26–28). Low CSF cellular counts, low inflammatory marker levels, and low IFN-γ response at baseline were also associated with an increased risk of paradoxical IRIS (29). A study by Jarvis et al., observed that IRIS development was associated with excessive polarization toward a T-helper type 1 (Th1) response against cryptococcal antigens after ART initiation, via baseline CSF cytokine analysis (19). Paucity of CSF immune responses during CM has been identified as a major risk factor for CM-IRIS. An inability to mount an effective interferon-driven antiviral immune response, accompanied by a systemic granulocyte pro-inflammatory signature prior to ART initiation, predisposes patients to the development of CM-IRIS (29). Early ART initiation in CM increases CSF cellular infiltrate, macrophage/microglial activation, and T-helper 2 responses within the central nervous system. This suggests that the increased mortality from early ART in the COAT trial was likely to be immunologically mediated (30).

Our study has several limitations. First, unmeasured study-specific effects due to timing or geographic differences between multicenter study sites may have caused residual confounding. Fortunately, these residual confounding effects were minimized by collection of extensive baseline data. Second, the absence of data on outcomes and key predictors, such as cryptococcal antigen titer, early fungicidal activity, biomarkers at baseline, and CSF opening pressures restricted our ability to accurately use these indices to extract important information relevant to our study. Third, our data were extracted from analysis of a fairly small cohort size, with a relatively narrow follow-up duration, which consequently limits the power of the subgroup analyses to prescribe customized guidance for care of individual patients with CM. Consequently, studies based on larger sample sizes and extended observation durations are warranted in the future.

In summary, our study provides evidence that early ART initiation during hospitalization for CM does not improve overall survival in HIV-positive patients with comorbid CM. Moreover, a trend toward increased all-cause mortality was observed with early initiation of ART in subgroup analysis. Furthermore, early ART initiation appears to be associated with an increase in the frequency of grade 4 adverse events, suggesting that it may be safer to defer initiation of ART in HIV-associated CM patients. Our findings emphasize the need for early diagnosis and treatment of HIV infection, preferably before patients present with advanced disease and life-threatening opportunistic infections such as CM.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by the Ethics Committee of Chongqing Public Health Medical Center (2019-003-02-KY). The patients/participants provided their written informed consent to participate in this study.

The study was conceived and designed by Y-KC and ML. ML, JY, J-MN, J-HY, T-TY, G-QZ, JL, and Y-MQ performed the component clinical trials. HC and TZ analyzed data. TZ and X-lX wrote the manuscript, with editorial input from all authors. VH, Y-qL, and TZ edited and revised the article. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling editor declared a shared affiliation with one of the authors HC at the time of review.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.