Association of Low CD4/CD8 Ratio With Adverse Cardiac Mechanics in Lymphopenic HIV-Infected Adults.

Abstract

To the Editors: INTRODUCTION People with HIV (PWH) have elevated risks for heart failure (HF),1,2 which can manifest during various stages of HIV infection. Among PWH, a lower nadir CD4 count (<200 cells/mm3) is associated with a higher HF risk.3 However, less is known regarding cardiac dysfunction in PWH with severe CD4 lymphopenia starting antiretroviral therapy (ART); such data are derived primarily from historical case series of AIDS cardiomyopathy and preceded detailed hemodynamic and tissue characterization on cardiac imaging.4–6 Of interest are PWH with immune reconstitution inflammatory syndrome (IRIS), who experience a dramatic increase in systemic and tissue inflammation after initiation of ART. Up to one-third of US PWH, and a larger proportion worldwide, present with CD4 counts <200 and are at a heightened risk for IRIS7; therefore, there is clinical relevance in determining factors underlying cardiac dysfunction in lymphopenic patients initiating ART. Given the dearth of modern data on cardiac function in PWH presenting with severe immunosuppression, we investigated associations of immunologic and inflammatory markers with cardiac function and intracardiac pressures in a cohort of PWH presenting with severe lymphopenia. METHODS Study Design We performed a nested study within a prospective cohort of PWH at the National Institute of Allergy and Infectious Diseases (NIAID) HIV clinic between January 1, 2007, and July 24, 2019. The cohort included patients from the following 2 clinical protocols: (1) Immune reconstitution syndrome in HIV-infected patients taking ART (IRIS, NCT00286767) and (2) positron emission tomography Imaging and Lymph Node Assessment of IRIS in People with AIDS (PANDORA, NCT02147405). Inclusion criteria included age ≥18 years, documented HIV infection, CD4 count <100 cells/µL, no previous ART, and willingness to start therapy. Exclusion criteria were pregnancy and active substance use. All study participants signed informed consent and were followed prospectively from ART initiation (week 0) to weeks 2, 4, 8, 12, 24, 36, and 48. ART regimens and initiation time were chosen according to local treatment guidelines and clinicians' preferences. The clinical team at the study site identified IRIS events as previously published.8,9 For this analysis, data elements derived from the IRIS and PANDORA studies were paired with echocardiographic data from the subset of IRIS and PANDORA participants who had clinically indicated echocardiography performed in the course of clinical care. The primary analyses focused on echocardiographic parameters after ART initiation and included participants from IRIS and PANDORA with echocardiography performed within 100 days after ART initiation. Secondarily, we investigated associations of pre-ART immune markers with pre-ART cardiac markers in participants from IRIS and PANDORA who had echocardiography performed within 14 days before initiating ART. Laboratory Evaluations Plasma HIV viral load, CD4, and CD8 T-cell counts were performed using standardized assays at enrollment (baseline/week 0—before ART), after ART initiation (week 2), and within 10 days of the first echo (up to 100 days post-ART). Batched cryopreserved plasma samples from participants at baseline and week 2 were tested using electrochemiluminescence (Meso Scale Discovery, Rockville, MD) for C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α); enzyme-linked fluorescent assay on a VIDAS instrument (bioMerieux, Marcy-l'Etoile, France) for D-dimer levels; and enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN) for soluble tissue factor and soluble CD14 (sCD14). Echocardiographic Measurements Left ventricular ejection fraction (LVEF), right ventricular systolic pressure (RVSP), diastolic mitral inflow velocities, left ventricular end diastolic dimension, and left atrial end systolic diameter were measured. RVSP is a validated correlate of pulmonary arterial pressure,10 which is a marker of cardiac congestion and worsening HF.11 For the primary analyses, we used the first available data after starting ART [median time to echo 18 days [interquartile range (IQR) 7, 42]. For secondary analyses, data were derived from the latest pre-ART echocardiogram. Statistical Analysis Wilcoxon rank-sum and Fischer exact tests were used to compare demographics, ART categories, echo parameters, immunologic markers, and inflammatory markers. We measured associations of immune cell markers at baseline, week 2, and at the time of first post-ART echo as well as inflammatory markers at baseline and week 2, with echo measures post-ART. Secondary analyses compared pre-ART laboratory markers and echocardiographic data. Correlations were measured using Pearson's correlation coefficients. To account for multiple testing of correlations, Bonferroni correction was used (significance threshold of P < 0.0007). For the between-group comparisons, P < 0.05 threshold was used. The co-primary echo end points measured were LVEF and RVSP, both of which were log transformed for the analyses. RESULTS There were 91 participants with echocardiography performed within 100 days after ART initiation, of a total of 270 total IRIS and PANDORA participants. Seventy-three participants in IRIS and PANDORA had echocardiography performed within 14 days before ART initiation and were eligible for secondary analyses. Of note, only 13 study participants had both pre-ART and post-ART echocardiography, precluding adequately powered longitudinal comparisons of pre-ART to post-ART changes in echocardiographic parameters. The mean age of the group analyzed for primary analyses (N = 91) was 40.7 ± 11.4 years, 59% were men, and 75% were non-Hispanic Blacks. The baseline median CD4 count was 19 T cells/mm3 (IQR, 8–44), and the median viral load was 193,184 copies/mL (IQR, 88,911–456,875). ART class prevalence was as follows: 53% on nonnucleoside reverse transcriptase inhibitors, 32% on protease inhibitors, and 56% on integrase inhibitors. Thirty-one of the included patients (34.1%) developed IRIS, 19% of the overall cohort had clinical HF at baseline, 13% had Kaposi sarcoma, 7% had Cryptococcus, 12% had active tuberculosis, and 33% had pneumocystis pneumonia. Patients with vs. without IRIS had no differences in demographics, baseline HF, ART type, or opportunistic infections (Table 1). There was no difference in echo parameters or immune cell counts and viral load at baseline, week 2, or time of echo. Patients with IRIS had higher baseline levels of D-dimer and TNF-α and higher levels of D-dimer, TNF-α, CRP, and IL-6 at 2 weeks (Table 1). Of the 31 patients, 7 patients (22.6%) in the IRIS group had adjudicated HF vs. 10 of the 60 patients in the non-IRIS group (16.7%; P = 0.69). Patients with vs. without baseline HF had no significant demographic differences and no difference in baseline immune cell counts or viral load but had higher baseline IL-6 levels [median = 4.3 pg/mL (IQR, 3.6–8.5)] compared with those without {median = 2.4 pg/mL [(IQR, 1.7–4.6), P = 0.035]}. TABLE 1. - Associations of Demographics, Clinical Variables, and Laboratory Markers With Cardiac Function and the Presence of IRIS Right Ventricle Systolic Pressureb Left Ventricle Ejection Fractionb IRIS absent (n = 60) IRIS Present (n = 31) P Statistica P Statistica P Age, yr (mean ± SD) 0.22 0.145 −0.01 0.921 41.92 ± 11.92 38.45 ± 10.14 0.171 Male sex (%) 43 (71.7) 16 (51.6) 0.095 BMI, kg/m2 (mean ± SD) 0.08 0.647 −0.02 0.922 22.44 ± 4.83 24.65 ± 9.52 0.327 Race-ethnicity (%) 0.299 Hispanic 3 (10.3) 4 (18.2) Non-Hispanic Black 21 (72.4) 17 (77.3) Non-Hispanic White 1 (3.4) 1 (4.5) 4 (13.8) 0 (0.0) Heart failure (%) 10 (16.7) 7 (22.6) 0.688 NNRTI (%) 30 (50.0) 18 (58.1) 0.611 PI (%) 20 (33.3) 9 (29.0) 0.857 Integrase (%) 34 (56.7) 17 (54.8) 1 Comorbid infections Active TB (%) 4 (8.2) 5 (20.0) 0.272 Pneumocystis pneumonia (%) 17 (35.4) 6 (27.3) 0.69 Left ventricular ejection fraction, % (mean ± SD) 52.52 (16.61) 53.79 (16.53) 0.796 Right ventricular systolic pressure, mm Hg (mean ± SD) 29.81 (10.15) 29.60 (11.24) 0.948 Mitral E/A ratio (mean ± SD) 1.16 (0.61) 1.39 (0.92) 0.352 LA systolic diameter, mm (mean ± SD) 34.00 (7.30) 35.16 (6.01) 0.575 LV end diastolic diameter, mm (mean ± SD) 49.04 (8.73) 47.93 (5.90) 0.676 Baseline markers CD4+ T cells/µL (median [IQR])b −0.03 0.848 −0.02 0.903 17.00 [8.00–41.00] 28.00 [8.00–47.50] 0.544 CD8+ T cells/µL (median [IQR])b 0.27 0.067 −0.13 0.387 373.00 [245.50–496.50] 393.00 [257.50–795.50] 0.306 CD4/CD8 ratio (median [IQR])b −0.52 0.0003 0.23 0.136 0.05 [0.03–0.10] 0.05 [0.03–0.08] 0.76 HIV viral load, copies/mL (median [IQR])b −0.07 0.656 0.04 0.810 156,623 [55,177–400,265] 242,793 [120,962–500,000] 0.105 d-dimer, ng/mL (median [IQR])b 0.13 0.394 0.03 0.865 1373.30 [856–2674] 2400 [1,434–3546] 0.027 CRP, ng/mL (median [IQR])b 0.30 0.042 0.04 0.774 5200 [1,223–17,432] 5869 [2,748–18,415] 0.388 IL-6, pg/mL (median [IQR])c −0.15 0.332 0.17 0.274 2.35 [1.59–5.00] 3.41 [2.13–5.93] 0.105 TNF-α, pg/mL (median [IQR])c −0.16 0.301 −0.10 0.539 9.08 [6.54–14.05] 13.26 [8.30–18.52] 0.026 Soluble CD14, ng/mL (median [IQR]) 0.11 0.481 0.02 0.903 2,610,237 [2,021,491–3,024,374] 2,505,229 [2,197,651–3,339,005] 0.519 Tissue factor, pg/mL (median [IQR]) 0.17 0.292 −0.23 0.155 80.55 [67.71–105.42] 94.15 [65.02–113.42] 0.874 Week 2 markers Change in CD4 from baseline to week 2 (median [IQR]) −0.18 0.240 0.15 0.329 24 [7–60] 57 [10–106] 0.087 CD4+ T cells/µL (median [IQR])b 0.03 0.824 −0.10 0.494 53.5 [23–100.5] 82 [30–153] 0.11 CD8+ T cells/µL (median [IQR])b 0.16 0.295 −0.01 0.935 417 [280–653] 492 [193–793] 0.983 CD4/CD8 ratio (median [IQR])b −0.15 0.326 0.09 0.564 0.10 [0.06–0.22] 0.16 [0.09–0.30] 0.148 HIV viral load, copies/mL (median [IQR])b −0.35 0.017 0.23 0.127 458 [233–1513] 699 [299–2910] 0.28 D-dimer, ng/mL (median [IQR])b 0.07 0.689 −0.03 0.864 969 [629–1836] 2354 [797–3920] 0.009 CRP, ng/mL (median [IQR])b 0.06 0.691 −0.06 0.695 6564 [1396–30,722] 32,795 [9,343–78,900] 0.001 IL-6, pg/mL (median [IQR])c −0.14 0.381 0.00 1.000 2.32 [1.22–5.13] 7.47 [2.63–16.52] <0.001 TNF-α, pg/mL (median [IQR])c −0.10 0.550 0.02 0.918 7.51 [4.76–13.08] 16.34 [8.42–25.66] 0.001 Soluble CD14, ng/mL (median [IQR]) 0.09 0.576 −0.04 0.813 2,257,003 [1,813,221–2,852,754] 2,742,045 [2,185,802–3,116,137] 0.086 Tissue factor, pg/mL (median [IQR]) 0.16 0.332 −0.30 0.067 84.16 [61.92–96.43] 86.23 [59.68–97.65] 0.96 Markers closest in time to echocardiogram (within 10 days) CD4+ T cells/µL (median [IQR])b −0.26 0.084 0.13 0.368 30.50 [11.00–57.50] 36 [14.00–73.50] 0.412 CD8+ T cells/µL (median [IQR])b 0.29 0.047 −0.10 0.491 371 [246.75–531.75] 455 [291.50–866.00] 0.197 CD4/CD8 ratio (median [IQR])b −0.52 0.0002 0.23 0.120 0.08 [0.03–0.15] 0.09 [0.03–0.20] 0.815 HIV viral load, copies/mL (median [IQR])b 0.25 0.096 −0.24 0.102 83,448 [1,825–194,913] 109,246.00 [623–300,598] 0.903 Bold P<0.05 significant for pairwise comparisons; Bonferroni-corrected P<0.0007 significant for multiple comparisons.aPearson correlation coefficients for continuous variable and Beta coefficients from linear regression for categorical variables, with P value calculated using the Wald test.bLog transformation performed because of nonnormal distribution of the variable.cSquare-root transformation performed because of nonnormal distribution of the variable.BMI, Body-mass index; LA, left atrial; LV, left ventricular; NNRTI, nonnucleoside reverse transcriptase inhibitor; PI, protease inhibitor; TB, tuberculosis. In the overall study population included in primary analyses, a lower CD4/CD8 ratio at baseline and time of echo (post-ART) was associated with higher RVSP (Table 1). HIV viral load at baseline, week 2, and near the time of echo did not have a significant association with LVEF. The inflammatory markers at baseline or week 2 and the change in CD4 count from baseline to week 2 were not associated with LVEF or RVSP. There were no significant interactions between immune or inflammatory markers and sex with respect to echocardiographic parameters. In secondary analyses of associations of pre-ART laboratory and pre-ART echocardiographic measures (N = 73 with pre-ART echocardiography), higher CRP levels were associated with higher RVSP (r = 0.46, P = 0.00004) and a lower CD4/CD8 ratio was associated with higher RVSP (r = −0.39, P = 0.0009), although this did not meet significance after Bonferroni correction. DISCUSSION We investigated associations of immune and inflammatory markers with cardiac function among PWH who presented with advanced AIDS in the modern ART era and were followed prospectively after ART initiation. This provided the opportunity to understand the relationship between immune activation that occurs in patients with advanced AIDS before and after ART initiation, including those with IRIS and myocardial dysfunction that may contribute to a downstream risk of HF observed in PWH. We found that lower CD4/CD8 ratios pre-ART and after ART initiation were associated with higher RVSP on post-ART echocardiography.11 This finding is consistent with recent studies from the broader HIV population that demonstrate clear associations of lower CD4 counts with a higher HF risk.2,3 Whether HIV-related CD4 lymphopenia itself is a cause of cardiac dysfunction and HF or a proxy marker of other factors (virologic, inflammatory, or otherwise) that lead to HF merits investigation in future mechanistic and clinical studies. Furthermore, it is possible that relative depletion of CD4 compared with CD8 lymphocytes and a related shift toward amplified CD8-predominant immune activation may predispose to heightened inflammation (as observed in studies of immune-progressed HIV) and HF.1,12–18 Yet, interestingly, although pre-ART CRP levels were associated with higher pre-ART RVSP, the presence of IRIS or inflammatory markers at baseline and week 2 was not associated with post-ART LVEF or RVSP. This may reflect the limited sample size, early intervention with anti-inflammatory therapies, or other residual confounding and warrants further investigation. The limitations of our study include the sample size and possible selection bias because echocardiography was performed as clinically indicated. Nevertheless, this represents the largest study investigating cardiac function among PWH presenting with advanced AIDS in the modern ART era. Our finding that, among PWH presenting with very low CD4 counts, a lower CD4/CD8 ratio is associated with higher RVSP warrants further study and may have implications for HIV-associated HF.