FormalPara Key Summary Points

Why carry out this study?

Screening for cryptococcal antigenemia at the time of HIV diagnosis is recommended for people living with HIV who present with low CD4 lymphocyte counts

Uptake of cryptococcal antigen screening and prevalence of cryptococcosis in Taiwan in the era of universal antiretroviral therapy are unknown

What was learned from the study?

Prevalence of cryptococcal antigenemia at the time of HIV diagnosis was 7.8% and 2.6% among people living with HIV who presented with CD4 lymphocyte counts < 100 and 100–199 cells/µl, respectively

Uptake of cryptococcal antigen screening among HIV late presenters was improving, but still suboptimal and delayed, and this delay was associated with a higher mortality

Delayed cryptococcal antigen screening was associated with mortality and hospitalization among people living with HIV who were diagnosed in the outpatient settings and presented with initial CD4 lymphocyte counts < 200 cells/µl

Digital Features

This article is published with digital features, including a summary slide, to facilitate understanding of the article. To view digital features for this article go to https://doi.org/10.6084/m9.figshare.14459733.

Introduction

Cryptococcal disease remains one of the leading causes of deaths among people living with HIV (PLWH) who are late presenters for care [1], and globally cryptococcal meningitis accounts for 15% of AIDS-related deaths [2]. Despite the state-of-the-art treatments, including optimized antifungal therapy and antiretroviral therapy (ART) and measures to control intracranial pressure, mortality from cryptococcal meningitis remains high [3, 4], and long-term disabilities are common [5].

Preventive interventions, such as screening of cryptococcal antigen (CrAg) before initiation of ART and pre-emptive administration of fluconazole, provide survival benefits [6, 7] and are cost-effective [8, 9] among PLWH with a very low CD4 lymphocyte count. The World Health Organization guidelines recommend this preventive intervention strategy among PLWH who present with CD4 < 100 cells/μl and provide a conditional recommendation of the strategy for those with CD4 100–199 cells/μl [10].

Burden of HIV-associated cryptococcosis varies geographically, with the highest prevalence and mortality occurring in sub-Saharan Africa and several Asia-Pacific countries [2]. In other parts of the world where cryptococcosis is non-endemic, such as Europe and North America, the incidence of cryptococcal meningitis has decreased dramatically in the era of effective ART [11, 12]. However, despite the rollout of universal ART programs, the burden of cryptococcal disease among PLWH remains substantial in the endemic area for cryptococcosis where HIV diagnosis and linkage to care are often delayed [13].

Taiwan, sitting along the west Pacific Island chain, has a similar prevalence of cryptococcal infection compared to those of Southeast Asian countries [14]. In a single-center study in Taiwan, the prevalence of HIV-associated cryptococcosis among 2022 newly diagnosed PLWH was 1.4% between 2004 and 2015 [15]. Routine CrAg screening was not included in the national HIV treatment guidelines, and the uptake of CrAg testing at the time of HIV diagnosis was unknown. In this study, we aimed to investigate the uptake of CrAg testing among newly diagnosed PLWH in Taiwan between 2009 and 2018, the prevalence of cryptococcal antigenemia, and their temporal changes. We also explored the impact of the timing of CrAg testing on all-cause mortality and hospital admission among those who presented with a CD4 lymphocyte count of < 200 cells/μl.

Methods

Study Setting and Patients

In this retrospective, multicenter study, consecutive PLWH who were newly diagnosed with HIV infection and presented for care to six medical centers around Taiwan for the first time between 2009 and 2018 were included. Those who had received their HIV diagnosis outside of these six participating hospitals and those who refused HIV care or were lost to follow-up shortly after a confirmed HIV diagnosis were excluded. Information on the demographics, co-infections, and results of laboratory tests at the time of HIV diagnosis was collected, as well as the first serum cryptococcal antigen titer within 6 months before and after the HIV diagnosis. Clinical outcomes, including retention in care, hospital admission, and mortality, were tracked up to 12 months after the HIV diagnosis.

The study was approved by the Research Ethics Committees and Institutional Review Boards of the participating hospitals (National Taiwan University Hospital, registration number 201003112R]; Far Eastern Memorial Hospital [105040-F]; Chung Shan Medical University Hospital [CS14034], Changhua Christian Hospital [160408]; National Cheng Kung University Hospital [B-BR-105-038]; Kaohsiung Medical University Hospital [KMUH-IRB-20110040]), and the informed consent was waived. The study was carried out according to the principles expressed in the Declaration of Helsinki.

Routine baseline laboratory investigations for newly diagnosed PLWH in Taiwan included CD4 lymphocyte count, plasma HIV RNA load (PVL), complete blood-cell count, liver and renal function tests, and serology for hepatitis A virus (HAV), hepatitis B virus (HBV), hepatitis C virus (HCV), and syphilis according to the national HIV treatment guidelines. The above laboratory investigations and prescription of ART are reimbursed by the Taiwan Centers for Disease Control. During the study period, ART initiation was recommended when CD4 was < 200 cells/µl in 2009, when CD4 was < 350 cells/µ, in 2010–2012, and when CD4 was < 500 cells/µl in 2013–2015; after 2016, newly diagnosed PLWH were provided with ART regardless of CD4 count at diagnosis.

Study Outcomes and Definitions

In this study, the burden of cryptococcosis was represented by the prevalence of cryptococcal antigenemia, defined by positive CrAg irrespective of the titers, and that of CrAg titer ≥ 1:8, which indicated active cryptococcosis [16]; both prevalences were stratified according to CD4 lymphocyte count categories (< 100, 100–199, and ≥ 200 cells/μl). Diagnosis of cryptococcal meningitis was confirmed by isolation of Cryptococcus species in the cerebrospinal fluid (CSF) or a CSF CrAg titer ≥ 1:8 [17].

To investigate the association between timing of CrAg testing and mortality in PLWH who presented with low CD4 counts, the included PLWH were categorized in the “late CrAg” group if CrAg was determined 14 days or later after HIV diagnosis [18], in the “early CrAg” group if CrAg was determined no later than 14 days after HIV diagnosis, and in “CrAg before HIV” group if CrAg was determined before confirmed HIV diagnosis. PLWH presented as “inpatients” at the time of HIV diagnosis if they were hospitalized when they received the HIV diagnosis; otherwise, they were classified as “outpatients.”

Laboratory Investigations

Throughout the study period, Latex-Cryptococcus Antigen Test (IMMY, Oklahoma, USA) was employed to determine CrAg positivity and CrAg titers with dilution up to 1:1024 [19]. The procedure was carried out at each participating hospital according to the manufacturer’s instructions. Determinations of PVL, CD4 lymphocyte count, and serological tests for syphilis and viral hepatitis (HAV, HBV, and HCV) were performed using certified commercial kits at the six participating hospitals.

Statistical Analysis

Proportions of newly diagnosed PLWH who underwent CrAg testing, as well as the prevalences of cryptococcal antigenemia and CrAg titer ≥ 1:8 among PLWH who tested for CrAg, were calculated annually and examined by Cochran-Armitage test for trend. Factors associated with CrAg testing among PLWH with CD4 < 100 cells/μl were examined in univariate analysis, with categorical variables analyzed by chi-square test or Fisher’s exact test, and continuous variables by Mann-Whitney U tests. Variables with P-value < 0.1 in the univariate analysis were then entered in multivariate logistic regression models with missing values treated by exclusion to identify independent association factors. The same procedure was carried out to identify factors associated with positive CrAg among those who tested for it.

Among PLWH with CrAg testing and baseline CD4 < 200 cells/µl, the 6- and 12-month all-cause mortality was shown on Kaplan-Meier plots, with stratification by timing of CrAg testing and inpatient/outpatient status at the time of HIV diagnosis. The association of timing of CrAg testing with 12-month all-cause mortality among PLWH who were outpatients at the time of HIV diagnosis was examined in a Cox proportional hazards model after adjusting for age, risk of HIV acquisition, baseline PVL and CD4 lymphocyte count, and presence of any opportunistic infection other than cryptococcosis. The association of timing of CrAg testing with 12-month all-cause hospital admission was also explored in the same manner. Statistical analyses were performed using the R statistics software (version 3.6.1), and P < 0.05 was deemed statistically significant throughout the analyses.

Results

Characteristics of the Patients

During the 10-year study period, 5372 newly diagnosed PLWH who sought HIV care at the six participating hospitals for the first time were included, which accounted for 25.2% of 21,321 newly diagnosed PLWH in Taiwan from 2009 to 2018 [20]. The included PLWH were predominantly male (97.3%) and their median age was 29 years (interquartile range [IQR], 25 to 36) (Table 1). The proportion of PLWH who presented with CD4 < 100 and those with < 200 cells/μl was 21.4% and 34.7%, respectively, both of which had decreased over the 10-year study period (P = 0.032 and 0.039, respectively) (Fig. S1a, b). More than half (51.9%) of the PLWH with baseline CD4 < 100 cells/μl received their HIV diagnosis as outpatients.

Table 1 Baseline characteristics of 5372 newly diagnosed people living with HIV included in the study
Full size table

Uptake of Cryptococcal Antigen Screening and its Trend

Overall, 2252 (41.9%) of 5372 PLWH underwent CrAg testing around the time of the HIV diagnosis (Fig. 1). Among PLWH with CD4 < 100 cells/μl, the uptake of CrAg testing increased significantly, from 65.7% in 2009 to 78.0% in 2018 (P = 0.002) (Fig. 2). The increasing trends were also observed among those with CD4 100–199 and those with CD4 > 200 cells/μl (both with P < 0.001). Among PLWH with CD4 < 100 cells/µl, after exclusion of 289 PLWH who had CrAg testing before their HIV diagnosis, 401 of 861 (46.6%) underwent CrAg testing within 14 days after confirmed HIV diagnosis, 101 (11.7%) had CrAg testing ≥ 14 days after HIV diagnosis, and 359 (41.7%) did not have CrAg testing.

Fig. 1
figure 1

Prevalence of cryptococcal meningitis and cryptococcal antigenemia. *Among the 5372 newly diagnosed people living with HIV, 2971 (55.3%) did not have serum cryptococcal antigen measured, 11 (0.2%) had serum CrAg measured > 6 months before HIV diagnosis, and 138 (2.6%) had serum CrAg measured > 6 months after HIV diagnosis

Full size image
Fig. 2
figure 2

Annual uptake of serum cryptococcal antigen (CrAg) testing among newly diagnosed people living with HIV according to CD4 lymphocyte count category from 2009 to 2018

Full size image

Among PLWH who presented with CD4 < 100 cells/μl, those being outpatients at the time of diagnosis were less likely to have CrAg testing (adjusted odds ratio [aOR], 0.313 [95% confidence interval [CI] 0.225–0.432]), while those who were older (aOR 1.038; 95% CI, 1.021–1.055), having CD4 lymphocyte counts < 50 cells/μl (aOR 1.643; 95% CI 1.202–2.245), and with leukocytosis (aOR 2.136; 95% CI 1.167–4.158) were more likely to have CrAg testing in the multivariate logistic regression model (Table 2 and univariate analysis in Table S1).

Table 2 Factors associated with testing of serum cryptococcal antigen among newly diagnosed people with HIV with nadir CD4 lymphocyte count < 100 cells/μl in multivariate logistic regression model
Full size table

Burden of Cryptococcal Disease and its Trend

Among PLWH who tested for CrAg, the prevalence of cryptococcal antigenemia was 7.8%, 2.6%, and 0.5% in those with CD4 < 100, 100–199, and ≥ 200 cells/μl, respectively (Fig. 1). The prevalence of cryptococcal antigenemia did not change significantly over the study period, except for the group with CD4 100–199 cells/μl, in which the prevalence decreased over time (P = 0.006) (Fig. S2a, b). Among 76 PLWH who tested positive for CrAg, 65 (85.5%) had CrAg titers ≥ 1:8, indicating active cryptococcal disease, and 65 (85.5%) required hospitalization. In a multivariate model, lower CD4 lymphocyte counts, anemia, and being an inpatient at the time of HIV diagnosis were independent factors associated with cryptococcal antigenemia (P < 0.001, P = 0.012, and P < 0.001, respectively) (Table S2). Among all included PLWH, confirmed cryptococcal meningitis occurred in 30 of 1150 (2.6%) PLWH with CD4 < 100 cells/µl, in 3 of 708 (0.4%) with CD4 100–199 cells/µl, and in 1 of 3504 (0.03%) with CD4 ≥ 200 cells/µl (Fig. 1). There were no temporal changes in the prevalences of cryptococcal meningitis between 2009 and 2018 (P = 0.624).

Timing of Cryptococcal Antigen Testing and Mortality

At 6 months of HIV diagnosis, the all-cause mortality rate was 11.8% among PLWH with positive CrAg compared to 3.7% among those with negative CrAg (unadjusted odds ratio 3.516, 95% CI 1.488–7.417). The overall mortality was highly associated with the CD4 category at the time of HIV diagnosis (Fig. S3). In the Kaplan-Meier plot of PLWH with CD4 < 200 cells/µl, the 6-month mortality was significantly higher among PLWH who were hospitalized at the time of HIV diagnosis compared to those who were not, irrespective of the timing of CrAg testing (P < 0.001, Figs. 3a and S4). Among those who received HIV diagnosis as outpatients, the mortality in the “late CrAg” group continued to accumulate with time, lasting up to 12 months, and exceeded that in the “early CrAg” and “CrAg before HIV” groups after 2 months of HIV diagnosis (P = 0.027, Fig. 3b). When the analysis was limited to PLWH presented with CD4 < 100 cells/μl, the findings were similar (Fig. S5a, b).

Fig. 3
figure 3

Kaplan-Meier plot demonstrating. a Six-month all-cause mortality by timing of cryptococcal antigen (CrAg) testing and outpatient/inpatient status at the time of HIV diagnosis, and b 12-month all-cause mortality by timing of CrAg among people living with HIV who tested for CrAg and presented with initial CD4 lymphocyte count less than 200 cells/µl

Full size image

In the Cox proportional hazards model, late CrAg testing was associated with increased risk of 12-month mortality compared to early CrAg testing (adjusted hazard ratio [aHR] 2.028, 95% CI 1.109–3.708) after adjusting for age, risk group of HIV acquisition, year of HIV diagnosis, outpatient/inpatient status and presence of opportunistic infections at the time of HIV diagnosis, baseline CD4 lymphocyte count, and PVL (Table 3). However, of the 12 PLWH with positive CrAg in the “late CrAg” group, only one with concurrent Burkitt’s lymphoma died of enterococcal bacteremia at 145 days after HIV diagnosis.

Table 3 Factors associated with 12-month all-cause mortality after HIV diagnosis in the Cox proportional hazards model including 1104 people living with HIV who presented with baseline CD4 lymphocyte count < 200 cells/µl and tested for cryptococcal antigen
Full size table

Timing of Cryptococcal Antigen Testing and Hospitalization

Among those who presented as outpatients at the time of HIV diagnosis, hospitalizations were also more commonly observed in the “late CrAg” group after 1 month of HIV diagnosis (P = 0.024, Fig. S6). In a Cox proportional hazards model, late CrAg testing was associated with increased risk for all-cause hospitalization compared to early CrAg testing after adjustments (aHR 1.550, 95% CI 1.100–2.184).

Discussion

In this multicenter study, 1 out of 13 newly diagnosed Taiwanese PLWH with a CD4 lymphocyte count < 100 cells/µl had cryptococcal antigenemia, and this prevalence did not decrease over the 10-year study period. The uptake of CrAg testing among newly diagnosed PLWH was increasing; however, at the end of the study period in 2018, > 20% of newly diagnosed PLWH with CD4 < 100 cells/µl still missed CrAg testing, while some tested late. Among PLWH with a baseline CD4 count < 200 cells/µl, late CrAg testing was associated with increased mortality 2 months after the HIV diagnosis.

Compared with other developed countries that have witnessed a decreasing incidence of HIV-associated cryptococcosis [13], the prevalence of cryptococcal meningitis in Taiwan remained stable despite improved access to HIV care and coverage of ART [21]. This was probably driven by the persistently high number of late presenters throughout the years. In our study, more than a third of newly diagnosed PLWH still presented with CD4 < 200 cells/µl and more than 20% with CD4 < 100 cells/µl by 2018 (Fi. S1a, b). Compared to the last surveillance data in Taiwan (data up to May 2016) [22], our findings demonstrated that the proportion of late presenters did not decline in the past several years. Ongoing efforts to remove the barriers to early HIV diagnosis and to improve linkage to care are needed to reduce the mortality and morbidity related to cryptococcosis and AIDS [1]; meanwhile, in countries with high cryptococcal endemicity like Taiwan [14], other measures to reduce the complications of cryptococcosis in PLWH should be pursued.

The strategy of screening coupled with pre-emptive antifungal therapy has been shown to prevent cryptococcal meningitis [7] and is cost-effective among PLWH with CD4 < 100 cells/µl [23]. Recently, growing evidence suggests that the risk of cryptococcal disease among PLWH with CD4 100–199 cells/µl was also substantial and should not be overlooked [24,25,26]. Furthermore, the REMSTART trial demonstrated that mortality could be reduced with CrAg screening and adherence support among PLWH with CD4 < 200 cells/µl [27]. In the present study, the prevalence of cryptococcal antigenemia among newly diagnosed PLWH with CD4 100–199 was 2.6%, and about 12% of cryptococcal meningitis developed among PLWH with CD4 ≥ 100 cells/µl. More studies are needed to investigate the benefit and cost-effectiveness of screening and a pre-emptive treatment strategy in this population with higher CD4 lymphocyte counts.

Our data showed a substantial gap in the uptake of CrAg testing among asymptomatic, newly diagnosed PLWH with advanced immunocompromised state. Among those who presented with CD4 < 100 cells/µl and an opportunity for the screening and a pre-emptive therapy approach (not yet testing for CrAg at the time of HIV diagnosis), half of them did not test for CrAg or did so within 14 days of their HIV diagnosis, especially those who presented as outpatients. The data from this real-world setting identified a critical gap in terms of HIV care, which has prompted the inclusion of a recommendation regarding CrAg screening among PLWH with a low CD4 count in the revised Taiwanese national HIV treatment guidelines in 2020 (data not shown). Other innovative strategies, such as reflexive laboratory testing, may address this problem if it can be successfully integrated into the clinical care [6, 26, 28].

Even with appropriate CrAg screening and pre-emptive fluconazole therapy, the mortality related to HIV-associated cryptococcosis was still higher than that of PLWH without cryptococcosis [7, 29]. This is not a phenomenon uniquely observed in the resource-limited countries. In Taiwan, where healthcare was generally considered accessible and affordable, our study revealed that the overall mortality of PLWH with cryptococcal antigenemia was more than three times higher compared to that of PLWH who were CrAg negative. Of note, a third of asymptomatic PLWH who screened positive for CrAg had confirmed cryptococcal meningitis [7, 30]. Post-screening lumbar puncture among those who screen positive for CrAg is a critical step to identify these patients and to guide optimal treatment. It has been shown that CrAg screening with pre-emptive fungal therapy failed to provide survival benefit when lumber puncture was not provided, especially among patients with high serum CrAg titers [31]. Further research is urgently needed to identify optimal management of this subgroup of patients [32]. In this study, we also found an association of delayed CrAg testing with increased mortality and hospitalization later after HIV diagnosis. The explanations for the association could be complex. PLWH with delayed CrAg testing could represent those whose ARTs were initiated late or whose adherence to ART was poor, who developed symptoms that triggered diagnostic testing later in the course, and who had the worst outcome, or it could indicate failure of a healthcare system to identify individuals with risks for opportunistic infection, which resulted in increased mortality. As detailed clinical information, such as timing of ART initiation, use of antifungal agents, response to ART, and others, was not included in the analyses in this study, we were unable to better define the reasons behind this observation. Nevertheless, caution should be taken that failure to recognize cryptococcosis early may result in unmasking immune reconstitution inflammatory syndrome after ART initiation, which may link to high mortality [18]. In the era of early and same-day ART, timely screening of CrAg become even more crucial to reduce cryptococcosis-related detrimental outcomes.

The study had several limitations. The retrospective study design introduced information and selection biases, which may affect the results of the study. For instance, about 30% of included PLWH with CD4 < 100 cells/µl did not test for CrAg, and the prevalence of cryptococcal antigenemia in this study could be over-estimated. However, even if all of them had negative CrAg, the corrected prevalence would be 5.3%, which would still be comparable to those of other high-burden countries [2]. Outpatient status at the time of HIV diagnosis was used as a surrogate marker for asymptomatic HIV infection. However, to interpret the findings related to this surrogate marker, it would be important to clarify that some of the patients with AIDS or other co-infection could be managed in the outpatient setting.

While the concept of care continuum models has been successfully introduced in HIV, tuberculosis, and HCV care [33], a care cascade for HIV-associated cryptococcosis, including timely screening, performance of lumbar puncture for those who were screened positive, and initiation of antifungals for prevention or treatment depending on the results, could also be implemented to improve clinical care. Like HIV care, the cascade starts with screening. Nowadays, point-of-care lateral flow assay, with better sensitivity and short turn-around time [34, 35], has made the screen and pre-emptive strategy more feasible. A newer, semi-quantitative, lateral flow test is currently under development and, if proven in well-designed clinical studies, has a potential to further improve the care cascade [36].

Conclusions

The burden of cryptococcosis among newly diagnosed PLWH with advanced immunosuppression remains high in Taiwan. The uptake of CrAg testing is still suboptimal and sometimes delayed. Delayed CrAg testing is associated with increased mortality later in the course of HIV care. Measures should be taken to ensure that people in HIV care also receive timely screening for cryptococcosis and subsequent interventions to reduce the related mortality and morbidity.