In total, 490 articles were obtained from 4 databases, among which 276 were from Pubmed, 106 were from Web of Science, 12 were from Cochrane Library, and 96 were from MEDLINE/EMBASE. 82 of the 490 articles were RCTs or cohort studies. Additional 12 articles (RCTs or cohort studies) were extracted from references of previous meta-analyses and their included studies.
All the 94 RCTs or cohort studies were included for screening. Firstly, 9 articles (6 from Web of Science and 3 from MEDLINE/EMBASE) were reduplicative and were therefore excluded from the 94 articles. Next, after screening titles and abstracts, 43 of the remaining 85 articles were excluded. Then, 25 articles were excluded from the remaining 42 articles after screening full-text, among which 1 article was news report, 3 articles reported patients with cryptococcal disease, 2 articles reported HIV-negative patients with cryptococcal antigenemia, 4 articles reported patients with CM or asymptomatic CM, 6 articles reported HIV-infected patients with negative CrAg, 1 article reported the epidemiology of cryptococcosis and 8 articles did not report CD4 cell counts or primary outcomes. Finally, 17 articles were included in our meta-analysis.
The characteristics of the included 17 studies were shown in Table 1. The assessment of quality and potential risk bias showed that the following may contribute to clinical and methodological heterogeneity, including: (1) the confounding factors or subject recruiting or incomplete follow-up in 7 of the 16 cohort studies, (2) the unclear risk of attrition in the RCT, and (3) the unclear risk of reporting and other bias in the RCT, as shown in Supplementary Table 1 and Supplementary Table 2.
Twelve of the 17 included studies reported the prevalence of CrAg positivity (1735 persons with CD<200 cells/μL in 3 studies; 783 persons with CD<150 cells/μL in 1 study; 3876 persons with CD<100 cells/μL in 8 studies). The pooled CrAg positivity prevalence in 6394 HIV infected persons with CD<200 cells/μL was 5% (95%CI: 3-6, I2=85.5%) (Figure 2).
Thirteen studies reported the incidence of CM among CrAg+ persons (1789 persons with CD<200 cells/μL in 3 studies; 312 persons with CD<150 cells/μL in 2 study; 248 persons with CD<100 cells/μL in 8 studies), and 4 studies reported the incidence of CM among CrAg- persons (54 persons with CD<200 cells/μL in 1 studies; 2107 persons with CD<100 cells/μL in 3 studies). The incidence of CM in 2349 CrAg+ persons was 7% (95%CI: 4-10; P=0.000; I2=74.6%), whereas the incidence of CM in 2161 CrAg- persons was 1% (95%CI: 0-1; P=0.343; I2=10.0%). The former is 7 times that of the latter (Figure 3 A and Table 2).
Nine studies reported the incidence of CM among persons who received antifungal therapy (909 persons with CD<200 cells/μL in 3 studies; 166 persons with CD<200 cells/μL in 2 studies; 88 persons with CD<100 cells/μL in 2 studies), and 9 studies reported the incidence of CM among persons who received placebo or no intervention (946 persons with CD<200 cells/μL in 3 studies; 146 persons with CD<150 cells/μL in 1 study; 160 persons with CD<100 cells/μL in 5 studies). The incidence of CM of 1163 persons receiving antifungal therapy was 2% (95%CI: 0-3; P=0.008; I2=61.1%), whereas the incidence of CM of 1252 persons in 9 studies who received placebo or no intervention was 9%, a 78% reduction (95%CI: 5-13; P=0.000; I2=73.7%) (Figure 3 B and Table 2).
Four studies compared the incidence of CM between 1081 persons receiving azoles and 1072 persons receiving placebo or no intervention (1768 persons with CD<200 cells/μL in 2 studies; 295 persons with CD<150 cells/μL in 1 study; 90 persons with CD<100 cells/μL in 1 study). We found that the risk ratio of CM events among persons who received placebo or no intervention was 6.03 times of those who received antifungal therapy (95%CI: 2.74-13.24; P<0.00001; I2=17%).
Thirteen of the 17 included studies reported all-cause mortality among CrAg+ persons (1936 persons with CD<200 cells/μL in 5 studies; 312 persons with CD<150 cells/μL in 2 studies; 255 persons with CD<100 cells/μL in 6 studies), 6 studies reported all-cause mortality in CrAg- persons (54 persons with CD<200 cells/μL in 1 study; 2087 persons with CD<100 cells/μL in 5 studies). The all-cause mortality of 2503 CrAg+ persons was 18% (95%CI: 11-25; P=0.000; I2=94.0%), 1.06 times that of 2141 CrAg- persons (17%, 95%CI: 3-31; P=0.000; I2=99.0%), which is shown in Figure 4A and Table 2.
Nine studies reported all-cause mortality in persons who received antifungal therapy (230 persons with CD<200 cells/μL in 4 studies; 166 persons with CD<150 cells/μ in 2 studies L; 88 persons with CD<100 cells/μL in 3 studies), 7 studies reported all-cause mortality in persons receiving placebo or no intervention (946 persons with CD<200 cells/μL in 3 studies; 146 persons with CD<150 cells/μL in 1 study; 110 persons with CD<100 cells/μL in 3 studies). The all-cause mortality of 484 persons receiving antifungal therapy was 16% (95%CI: 10-22; P=0.002; I2=67.3%), which was same to that of 1202 CrAg- persons receiving placebo or no intervention (16%, 95%CI: 7-24; P=0.000; I2=92.4%) (Figure 4A and Table 2).
Five studies (1897 persons with CD<200 cells/μL in 3 studies; 295 persons with CD<150 cells/μL in 1 study; 90 persons with CD<100 cells/μL in 1 study) compared all-cause mortality between persons who received azole antifungal therapy and persons who received placebo or no intervention. No significant difference was found in all-cause mortality (risk ratio: 0.82, 95%CI: 0.42-1.600; P=0.55; I2=61%) between 1144 CrAg+ persons who received an azole drug and 1138 CrAg+ persons who received placebo or no intervention (Figure 5).
In addition, we estimated and compared the prevalence of CrAg positivity, the incidence of CM and all-cause mortality between persons with CD4<100 and persons with CD4 100-200 cells/μL. The results showed that the risk ratio of CrAg positivity prevalence among HIV-infected persons with CD4<100 cells/μL was 2.69 times that of those with 100-200 cells/μL (95%CI: 1.48-4.88; p=0.001, I2=34%; 2 studies; 1672 persons) (Table 3). The risk ratio of the incidence of CM among HIV-infected persons with CD4<100 cells/μL was 4.96 times that of those with CD4 100-200 cells/μL (95%CI: 1.94-12.68, p=0.23, I2=31%; 3 studies, 1943 persons). The risk ratio of the all-cause mortality among HIV-infected persons with CD4<100 cells/μL was 4.15 times that of those with CD4 100-200 cells/μL (95%CI: 0.89-19.42, p=0.07, I2=0%; 2 studies, 1552 persons) (Table 3). Further, the risk ratio of the incidence of CM among persons with CD4 100-200 cells/μL receiving antifungal treatment was 0.90 compared to those receiving placebo or no intervention (95%CI: 0.06-13.89, p and I2 not applicable; 2 studies; 135 persons) (Table 3). The risk ratio of the all-cause mortality among persons with CD4 100-200 cells/μL receiving antifungal treatment was 0.27 compared to those receiving placebo or no intervention (95%CI: 0.01-4.93, p and I2 not applicable; 1 study; 7 persons) (Table 3).