Should we perform the serum cryptococcal antigen test in people living with HIV/AIDS hospitalized due to a community-acquired pneumonia episode?

Abstract

Community-acquired pneumonia (CAP) is a common cause of hospitalization among people living with HIV/AIDS (PLWHA), particularly those with severe immunosuppression. Pulmonary disease due to cryptococcosis is uncommonly reported and likely under-diagnosed. There is scarce information cryptococcal antigen (CrAg) prevalence in PLWHA with CAP. The objectives of this study were to identify among PLWHA who were hospitalized with CAP: i) the prevalence of serum CrAg positivity, ii) the proportion with asymptomatic vs. symptomatic cryptococcosis; and iii) the prevalence of serum CrAg positivity in CD4+ T-cell count <100 cells/mm³. We performed a sub-analysis of a prospective cohort of hospitalized adults enrolled into randomized clinical trial testing therapy for CAP. We included 202 participants who had a serum CrAg testing performed. We found a 3.5% prevalence of serum CrAg-positivity overall, being higher (4.8%) in CD4+ T-cell count <100 cells/mm³. Asymptomatic and symptomatic cryptococcosis was present in 2.0% and 1.5% overall, respectively. This study identifies a target population for CrAg testing: PLWHA hospitalized with diagnosis of CAP, particularly those with CD4+ T-cell count <100 cells/mm³ where the number needed to test was 18 to detect 1 CrAg-positive person. This approach may facilitate the detection of asymptomatic cryptococcal infection and allow a timely diagnosis of symptomatic cryptococcal disease.

Introduction

Cryptococcal meningitis is the most common clinical manifestation of cryptococcosis among people living with HIV/AIDS (PLWHA). Cryptococcosis continues to cause high rates of mortality, particularly in low and middle-income countries [1]. Although the respiratory system is the portal of entry of Cryptococcus neoformans and C. gattii, isolated pneumonia without other organ involvement is infrequently reported and diagnosed in PLWHA. In contrast, community-acquired pneumonia (CAP) is a common cause of hospitalization among PLWHA, particularly those with severe immunosuppression [2]. In addition, key strategies in the management of PLWHA with severe immunosuppression include both the timely diagnosis of symptomatic cryptococcosis in order to initiate antifungal treatment and the identification of isolated cryptococcal antigenemia (CrAg) in order to initiate pre-emptive therapy [3,4].

This study’s three objectives were to identify among PLWHA who were hospitalized with CAP: i) the prevalence of serum CrAg positivity, ii) the proportion with asymptomatic vs. symptomatic cryptococcosis; and iii) prevalence of serum CrAg positivity in patients with CD4+ T-cell count <100 cells/mm³.

Methods

This is a sub-analysis of a randomized clinical trial that enrolled adults ≥18 years of age with clinically and radiologically suspected community-acquired pneumonia who required antibiotic treatment (denoting a clinical diagnosis of bacterial pneumonia) and hospitalization. This trial included 224 PLWHA (Brazilian Clinical Trials Registry: RBR-8wtq2b) at Instituto de Infectologia Emílio Ribas, a tertiary teaching infectious diseases hospital in São Paulo, Brazil [5]. An extensive work up for etiology of pneumonia was available, including: blood samples for bacterial, fungal, and mycobacterial cultures; sputum for direct examination for Pneumocystis jirovecii and acid-alcohol-resistant bacilli, and cultures for bacteria, fungi and mycobacteria; blood samples for serology for Chlamydophila pneumoniae and Mycoplasma pneumoniae; blood samples for Streptococcus pneumoniae and Haemophilus influenzae by PCR; urinary antigen test for Legionella pneumophila serogroup 1; nonquantitative PCR methods for C pneumoniae, L pneumophila, M pneumoniae, P jirovecii, and adenovirus in respiratory samples; and nasopharyngeal swabs were tested by real time-PCR for the following agents: parainfluenza viruses 1 to 3, respiratory syncytial virus, influenza viruses A and B, human coronaviruses CoV NL63, HKU1, OC43, and 229E, enterovirus, rhinovirus, adenovirus, bocavirus, human metapneumovirus, C pneumoniae, Bordetella pertussis, and M pneumoniae. In addition, a serum cryptococcal antigen latex agglutination test (CrAg LA) was performed [6].

For this sub-study, we excluded participants missing serum CrAg test results (n=12), with known history of cryptococcal disease (n=5), or non-infectious cause for the pulmonary disease (n=5). Lumbar punctures were routinely performed among CrAg-positive participants to determine presence of central nervous system cryptococcosis. Clinical and laboratorial data, including CD4+ T-cell count were prospectively collected during the parent clinical trial and from the electronic hospital database. We summarized categorical variables with n (%) and summarized quantitative variables with mean and standard deviation (SD), median and interquartile range (IQR) or median and minimum-maximum (min-max), as appropriate.

Results

We included 202 participants admitted with diagnosis of CAP who had a CrAg tested performed. Of participants, 68% (n=137) were men, and the mean (± SD) age was 40 (±12) years. Overall, 80% (161/202) were antiretroviral therapy (ART) experienced, but only 17% (34/161) reported regular ART use. Among 184 participants with CD4+ T-cell count measured, the median (IQR) CD4+ T-cell count was 51 (16–234) cells/mm³ with 57% (106/184) having CD4+ T-cell count <100 cells/mm³. Only 13% (n=27) of hospitalized participants with pneumonia had HIV viral loads of <50 copies/mL.

The prevalence of serum CrAg-positive was 3.5% (7/202; 95%CI, 1.4% to 7.0%). Table 1 reports the main clinical and laboratory characteristics of these CrAg-positive participants. Four (2.0%) of 202 cases were classified as asymptomatic cryptococcal antigenemia, and three (1.5%) of 202 cases had symptomatic cryptococcosis. The medians (min-max) CD4+ T-cell counts were 67 (19–920) cells/mm³ and 12 (4–18) cells/mm³ in asymptomatic cryptococcal antigenemia and symptomatic cryptococcosis, respectively (P = 0,034). Of symptomatic cryptococcosis, one had culture-proven symptomatic cryptococcal meningitis; one had culture-proven asymptomatic cryptococcal meningitis associated with clinical and radiological diagnosis of pneumonia but without bacterial etiologic diagnosis (with pneumonia possibly caused by Cryptococcus); and one had pathology-proven symptomatic cryptococcal pneumonia. The prevalence of serum CrAg-positivity in pneumonia patients with CD4+ T-cell count <100 cells/mm³ was 5.7% (6/106; 95%CI, 2.1% to 11.9%). The one other serum CrAg-positive participant had a CD4+ T-cell count of 920 cells/mm³, HIV viral load of <50 copies/mL, and regular ART use was serum CrAg-positive and was classified as asymptomatic cryptococcosis; however, this may have been a false positive latex agglutination.

Table 1.

Characteristics of people living with HIV/AIDS hospitalized due to community-acquired pneumonia and positive serum cryptococcal antigen (CrAg).

ID	Age / Sex	CD4+ cells/mm3	Clinical presentation	Radiological findings	Community-acquired pneumonia etiology and other diagnosis	Pneumonia treatment	Symptomatic cryptococcosis	CSF culture; CSF CrAg	Fungal blood cultures	Serum CrAg titer	Cryptococcal treatment	Hospital outcome
1	58 / Male	85	Productive cough, pleuritic pain	Interstitial infiltrate and consolidation	Rhodococcus (culture proven in sputum and blood); Candida esophagitis	Ceftriaxone + azithromycin + TMP-SMX	No	No lumbar puncture performed	Negative	1:16	None†	Discharge to home
2	51 / Male	49	Dry cough, dyspnea, pleuritic pain	Interstitial infiltrate	Pneumocystis (PCR-proven in sputum); Oral thrush	Ceftriaxone + azithromycin + TMP-SMX	No	No growth; negative CrAg	Negative	1:16	None	Discharge to home
3	43 / Male	18	Productive cough, dyspnea, pleuritic pain	Interstitial infiltrate	No etiology identified	Ceftriaxone + azithromycin	Yes*	Cryptococcus neoformans and positive CrAg	Negative	1:512	Amphotericin + fluconazole	Death
4	43 / Male	920	Productive cough, dyspnea, pleuritic pain, fever	Consolidation	Rhinovirus (Swab- PCR Film Array)	Ceftriaxone	No	No growth; negative CrAg	Negative	Positive**	None	Discharge to home
5	47 / Male	12	Dry cough, pleuritic pain, mental confusion	Interstitial infiltrate and consolidation	Pneumocystis (PCR-proven in sputum)	Ceftriaxone + azithromycin + TMP+SMX	Yes	Cryptococcus neoformans; positive CrAg	Positive	1:1024	Amphotericin + fluconazole	Discharge to home
6	50 / Male	4	Productive cough, pleuritic pain, fever	Interstitial infiltrate and consolidation	Cryptococcosis (pulmonary pathology-proven)	Ceftriaxone + clarithromycin	Yes*	No growth; negative CrAg	Negative	Positive**	Amphotericin + fluconazole	Discharge to home
7	63 / Male	19	Dry cough, dyspnea, pleuritic pain, fever	Interstitial infiltrate	Pneumocystis (PCR-proven in sputum); Cytomegalovirus esophagitis	Ceftriaxone + azithromycin + TMP-SMX	No	No growth; negative CrAg	Negative	1:4	Amphotericin + fluconazole	Discharge to home

Note. CSF: cerebrospinal fluid; CrAg = cryptococcal antigen latex agglutination test; PCR: polymerase chain reaction; TMP-SMX: trimethoprim-sulfamethoxazole; CAP: community-acquired pneumonia;

^*Absence of neurological complaints.

^**Only qualitative assay was performed.

^†Developed culture-proven cryptococcal meningitis 7 months later.

None of the patients received cryptococcal preemptive treatment during hospitalization. Patient number 1 (Table 1) developed a culture-proven cryptococcal meningitis 7 months later. Unfortunately, long-term follow-up was unavailable for the other participants.

Discussion

We found 3.5% of HIV-infected adults hospitalized with CAP were serum CrAg-positive, and the prevalence of serum CrAg-positive was 5.7% in patients with CD4+ T-cell count <100 cells/mm³.

Recently, the World Health Organization recommended a package of interventions for PLWHA with advanced HIV disease including screening, treatment, and/or prophylaxis for major opportunistic infections, including cryptococcosis and tuberculosis [4]. Routine serum CrAg screening followed by pre-emptive antifungal therapy if CrAg-positive, should be considered in patients with CD4+ T-cell count <100 cells/mm³ to reduce the development of cryptococcal disease [3,4].

Although this recommendation is derived from studies among ART-naïve persons, this strategy should be pursued among PLWHA with advanced disease who are re-engaging into care after a period of ART interruption or who are failing ART and presenting ill [4]. The profiles of patients included in this present study and the results of prior studies performed in Latin America and Africa reinforce this strategy [7–10].

Initial cryptococcal infection is most often asymptomatic, but cryptococcosis can presents as subacute pneumonia. CrAg may precede by several weeks to months the manifestation of meningitis. However, in clinical practice, routine CrAg testingis uncommon when a patient presents with pulmonary disease without neurologic symptoms. Cryptococcal lung involvement can be observed among 38–78% of autopsies of PLWHA with disseminated cryptococcosis [11]. However, isolated cryptococcal pneumonia is less frequent even in autopsy studies (for example, 7% in autopsies of HIV-infected patients with cryptococcosis) [11]. In the present study only one patient had confirmed pulmonary cryptococcosis and all but one of the other 6 CrAg-positive patients had other confirmed etiologies to explain the respiratory findings. Interestingly, the single patient without etiologic diagnosis of pneumonia had cryptococcal meningitis, therefore, pulmonary cryptococcosis cannot be excluded. This patient had no neurologic complaints, and cryptococcal meningitis was confirmed only after lumbar puncture was performed in response to the serum CrAg. In this line, a recent study reported that up to one-third of asymptomatic CrAg-positive patients have concurrent central nervous system cryptococcal involvement [12]. This is particularly true as the CrAg titer rises above 1:160 [13]. These cases of ‘subclinical cryptococcal meningitis’ are likely under-recognized. World Health Organization guidelines do not specifically recommended lumbar punctures among asymptomatic CrAg-positive patients in 2011 [14], but this recommendation updated in 2018 [3]. Thus, lumbar puncture should be used routinely to investigate for cryptococcal meningitis in all CrAg-positive patients, even those without symptoms [3,4,12]. In the same study, blood CrAg titer using lateral flow assay (LFA) (≥ 1:160) was significantly associated with cryptococcal meningitis in both asymptomatic and symptomatic patients [12]. In accordance with this, our patient with cryptococcal meningitis but without neurological manifestations showed CrAg latex agglutination titres of 1:512 equating to an approximate CrAg LFA titer of ~1:1280 [15].

In the present study, two patients with confirmed cryptococcal meningitis had Pneumocystis jiroveci pneumonia, which pulmonary cryptococcosis can mimick [16–18]. There is scarce information about frequency of CrAg in hospitalized PLWHA with PAC, but cryptococcal pneumonia is rarely reported, but also rarely part of diagnostic pulmonary testing [5]. Contrary this general assumption, Harris et al. studied PLWHA in rural Thailand demonstrating among patients hospitalized for an acute respiratory illness, 13.1% were serum CrAg-positive, and 40% may have had cryptococcal pneumonia [19]. In Uganda, among 563 hospitalized PLWHA with cough for ≥2 weeks and suspected tuberculosis, similarly 5.7% were CrAg-positive. In this study, the patients had no clinical suspicion of cryptococcal disease. Interestingly, none of the CrAg-positive patients who underwent bronchoscopy were diagnosed with cryptococcal pneumonia [20].

Our results from Brazil as well as the studies from Thailand and Uganda suggest that when tested for, CrAg-positivity is relatively common among PLWHA with pulmonary disease, particularly in patients with severe immunosuppression. In our study, the population was patients with probable bacterial CAP who were enrolled into a clinical trial on bacterial pneumonia therapy. Thus, we do not contemplate patients with other radiographic patterns (e.g., clustered nodular pattern, solitary pulmonary nodule, mass with/without cavitation, or scattered nodules) where cryptococcal disease is a concern [17,18,21,22]. For future research studies of bacterial pneumonia or tuberculosis in PLWHA, CrAg testing should always be included as part of the initial evaluation.

The limitations of this study correspond to the retrospective design. However, we have been able to retrieve the information needed to achieve the desired goals. In addition, CrAg-positive serum only demonstrates the presence of cryptococcal antigen, but does not necessarily indicate cryptococcal pneumonia. Ideally, all CrAg-positive patients included in this study should have undergone bronchoscopy with bronchoalveolar lavage in order to systematically assess the diagnosis of cryptococcal pneumonia. Finally, considering that LFA performance is more sensitive than any other diagnostic test for cryptococcosis, it is possible that some cases may have been underdiagnosed in this study.

In conclusion, we found a high prevalence (3.5%) of CrAg-positivity in PLWHA admitted with diagnosis of CAP; 2.0% of the cases were classified as asymptomatic cryptococcal antigenemia and 1.5% had symptomatic cryptococcosis. The prevalence of CrAg-positivity was higher (5.7%) in persons with CD4+ T-cell count <100 cells/mm³ where 1 in 18 patients had cryptococcal antigenemia. This study demonstrates another targeted population for CrAg testing: PLWHA hospitalized with diagnosis of CAP, particularly those with CD4+ T-cell count <100 cells/mm³, regardless of ART status. In the absence of timely CD4+ T-cell count, we recommend all PLWHA with pulmonary disease be CrAg screened to detect and preemptively treat cryptococcal infection while asymptomatic or to enable a timely diagnosis of symptomatic cryptococcal disease.