Baseline C-reactive Protein as a Risk Factor for Cryptococcal Meningitis and Death in HIV-associated Cryptococcal Antigenemia With CrAg Titer as an Effect Modifier

Abstract

Background: Persons with HIV and cryptococcal antigenemia are at high risk of progression to cryptococcal meningitis or death. Baseline cryptococcal antigen (CrAg) plasma titer ≥1:160 is a known risk factor for poor outcomes, but other risk factors are unknown. In HIV-associated cryptococcal meningitis, baseline serum C-reactive protein (CRP) concentrations are positively associated with increased mortality. We hypothesized that CRP might also be associated with meningitis or death in persons with cryptococcal antigenemia.

Methods: We measured plasma CrAg titers and CRP concentrations on cryopreserved serum from prospectively enrolled persons with HIV and cryptococcal antigenemia. Using time-to-event analyses, we compared 24-week meningitis-free survival in persons with normal CRP (<8 mg/L) and elevated CRP (≥8 mg/L). Logistic regression was used to assess how CRP concentration and CrAg titer might interact as covariates.

Results: Of the 94 persons with elevated CRP, 19 (20.2%) developed meningitis or death, whereas of the 88 persons with normal CRP, 8 (9.1%) developed meningitis or death (P = .035). Persons with CrAg titer <1:160 and normal CRP had an ∼5% (3/61) event rate, whereas those with CrAg titer <1:160 but elevated CRP had an ∼20% (12/59) event rate. Importantly, we identified a statistically significant interaction effect between CrAg titer and CRP groups, in which elevated CRP increased risk in the low CrAg titer group (odds ratio, 1.54; 95% confidence interval, 1.16-2.04), but this effect was not present in high CrAg titer group (odds ratio, 0.78; 95% confidence interval, .53-1.15).

Conclusions: Our findings demonstrate that CrAg titer may modify the direction of effect of CRP with meningitis-free survival; future studies should account for this interaction.

Keywords: HIV; c-reactive protein; cryptococcal antigen titers; cryptococcal antigenemia; cryptococcal meningitis.

Figure 1.

A, Cumulative meningitis-free survival at 24 wk by C-reactive protein grouping. Figure 1A demonstrates a Kaplan Meier curve for 24-wk meningitis-free survival by CRP grouping in persons with HIV and asymptomatic cryptococcal antigenemia. The primary event was the development of symptomatic meningitis or death. Elevated serum CRP (>8 mg/L) was associated with higher probability of meningitis or death as compared to a normal CRP (<8 mg/L). B, Cumulative meningitis-free survival at 24 wk by C-reactive protein and CrAg titer grouping. Figure 1B demonstrates a Kaplan Meier curve for 24-week meningitis-free survival by serum CRP and plasma CrAg titer grouping among those with HIV-associated cryptococcal antigenemia. Persons with normal CRP (<8 mg/L) and low CrAg titer (<1:160) had an event rate of ∼5%. Persons with elevated CRP (>8 mg/L) or high CrAg titer (>1:160) had an approximately 4-fold higher event rate of ∼20%. Although normal CRP had a protective effect among participants with low CrAg titer, this benefit was not observed among participants with high CrAg titer, where participants from both normal and elevated CRP groups experienced approximately the same event rate. Supplementary Table 1 breaks down the composite event outcome to either meningitis or death.

Figure 2.

Logistic regression model measuring goodness of fit using area under the curve (AUC) for predicting future meningitis-free survival through 24 wk. (A) log2 CRP as lone predicator. (B) log2 CRP and CrAg titer >1:160. (C) log2 CRP and CrAg titer >1:160 with addition of an interaction term (log2 CRP * CrAg titer >1:160).

Figure 3.

Log odds of clinical event per log2 CRP increase by CrAg titer grouping. Log odds of clinical events (meningitis or death) by increasing log₂ CRP, stratified by CrAg titer grouping. Events are defined by progression to meningitis or death. In the group with CrAg titer <1:160, the log odds of an event increases as CRP increases. On the contrary, in the CrAg titer ≥1:160 group, the log odds of an event decreases as CRP increases. We speculate that this may indicate the immune system is still capable of generating an inflammatory response in the setting of systemic infection, where elevated CRP in low CrAg titer persons may represent undiagnosed AIDS-related pathology while elevated CRP in high CrAg titer persons may represent a commensurate immune response to a large burden of cryptococcal disease.