Clinical progression of severely immunosuppressed HIV-infected patients depends on virological and immunological improvement irrespective of baseline status

Abstract

Introduction

Current ART guidelines recommend early initiation of therapy regardless of CD4 count.^1,2 However, even among patients presenting for care in developed countries, a proportion seek medical care only after advanced immunosuppression has developed.^3,4 This may vary in different population groups and with geographic origin.⁵ Several papers have suggested that patients who initiate ART with advanced HIV may be at greater risk of disease progression compared with those who initiate ART with less immunosuppression.^6,7 In particular, discordant patients (those with a persistently low CD4 count despite adequate virological response) may progress to AIDS or death more frequently.⁸ However, few studies have assessed the interaction between immunological and virological status at baseline (BL) and follow-up and the progression of the disease.^8,9

The aim of this study was to analyse the role of CD4 and viral load (VL) at BL and at follow-up in severely immunosuppressed HIV-infected patients initiating their first ART regimen.

Methods

The PISCIS Cohort is a multicentre, observational study of HIV-infected individuals aged >18 years followed at 14 participating hospitals in Catalonia and the Balearic Islands (Spain). The ethics committees of the participating hospitals approved the study and patients gave their signed informed consent to participate. Patient follow-up is performed according to published clinical guidelines. Clinical and laboratory parameters were assessed every 3–6 months during follow-up. Between 1 January 1998 and 31 December 2011, 14 673 HIV-seropositive patients were recruited (73 726 person-years of follow-up). During that period, patients were considered lost to follow-up when clinical and laboratory parameters were not reported for >12 months since the date of administrative censoring (31 December 2011).

We included all ART-naive patients with CD4 <200 cells/mm³ at enrolment and who initiated ART consisting of two nucleoside analogues plus either a PI or an NNRTI.

BL for this analysis was defined as the date of the first study visit. In the survival analysis of AIDS/death, a patient was considered as a progressor to AIDS/death if presenting an AIDS event [a first episode or a new AIDS-defining disease in those with a BL diagnosis of AIDS, according to the CDC 2003 definition adapted for Europe (CD4 <200 cells/mm³ was not considered an AIDS event)] or death. In the overall survival analysis (event, death), patients with or without AIDS during follow-up were included.

Immunological response (IR) was defined as reaching a threshold of CD4 ≥200 cells/mm³ on two separate occasions ≥3 months apart and with no further results below this value, before presenting an AIDS event. Virological response (VR) was defined as having an undetectable VL (either <200 or <500 copies/mL depending on the detection limit of the assay) in ≥80% of determinations after the first 6 months of follow-up, before presenting an AIDS event.

Statistical analysis

Descriptive results of continuous variables were expressed as medians and IQRs and the results of qualitative variables were presented as frequencies and percentages.

We focused the analyses on the possible effect of CD4 and VL, both at BL and follow-up, in the risk of progression to an AIDS event or death. Kaplan–Meier survival curves were used to determine the probability of the combined endpoint of AIDS/death or death only, using the log-rank test to compare groups stratified by combined BL CD4/VL and whether or not the patient reached IR and/or VR, before the event AIDS/death occurred. Regarding the survival analysis, patients not reaching the events (AIDS/death or death only) were censored at the end of the available follow-up. A Bonferroni correction was applied to adjust the significance level for multiple pairwise comparisons. Cox regression models were used to assess the effect of CD4 and VL on the risk of progression to AIDS/death and death. Three different approaches were considered: (i) stratification of patients according to BL CD4/VL and whether or not IR/VR was reached; (ii) the time needed to achieve immune restoration (CD4 ≥200 cells/mm³ before an AIDS event) in <6 months, 6–12 months, 12–24 months, >24 months or never; and (iii) separately assessing the BL CD4 counts, BL VL and both CD4 counts (≥200 or <200 cells/mm³) and VL (detectable or undetectable) over time included as time-dependent variables. All models were adjusted for age (>50 or ≤50 years), sex, transmission group (intravenous drug use), hepatitis C virus status, AIDS at BL, ART initiation year (range 1998–2011) and ART regimen (NRTI + PI or NRTI + NNRTI).

For all statistical analyses, P values ≤0.05 were considered significant. Analyses were performed using SPSS, version 19.0 (SPSS, an IBM company, Chicago, IL, USA).

Results

The analysis included 2295 patients. BL characteristics are detailed in Table 1. The median follow-up time was 5 years (IQR 2.4–9.1) and at the end of follow-up, 69.9% had reached a CD4 count ≥200 cells/mm³, 64.4% reached VR and 482 (21%) patients had progressed to AIDS/death.

Figure 1 shows the percentages of patients with progression to AIDS/death stratified by BL CD4/VL and whether or not IR or VR were achieved during follow-up. At 5 years, the highest disease progression (77.8% AIDS/death) was observed in patients in the worst scenario, i.e. those with BL CD4 count ≤100 cells/mm³ and VL >5 log copies/mL and who did not achieve either IR or VR during follow-up. Patients with a better BL situation, but poor virological or immunological response, also had a poor outcome, with 59.9% having disease progression. Patients with any BL situation and VR during follow-up (persistent CD4 count <200 cells/mm³) experienced reduced disease progression (35.4% AIDS/death), whereas only 18.1% of patients reaching IR, but not VR, had disease progression. The lowest rate of disease progression was found in patients who reached both IR and VR during follow-up, regardless of their BL situation (1.9% in BL CD4 >100 cells/mm³ and VL <5 log copies/mL; 2.3% in BL CD4 ≤100 cells/mm³ and/or VL >5 log copies/mL). Overall, statistically significant differences were found between groups (P < 0.001). These results were confirmed after adjusting for BL variables and confounders in a multivariate Cox regression model (Model 1, Table 2). Similar results were found in the analysis of overall survival (Figure 1b) and also were confirmed by the multivariate Cox regression model (data not shown).

Figure 1.

(a) Progression to AIDS/death and stratified by BL CD4/VL, reaching IR and VR during follow-up (censored at 5 years). (b) Overall survival stratified by BL CD4/VL, reaching IR and VR during follow-up (censored at 5 years). FU, follow-up. ^aLog-rank test to compare amongst groups. Pairwise comparisons in progression to AIDS/death: after adjusting by Bonferroni correction, statistically significant differences were found between all pairs of Groups a, b, c, d, e and f (P values <0.05), except in the comparison between Groups e and f. Pairwise comparisons in death-only analysis: after adjusting by Bonferroni correction, statistically significant differences were found between most pairs of Groups a, b, c, d, e and f (P values <0.05), except in the comparison between the following pairs of groups: a–b, b–c, c–d and e–f.

Open in new tabDownload slide

Since reaching CD4 ≥200 cells/mm³ during follow-up plays an important role in progression to AIDS/death, we studied the possible effect on disease progression of the time taken to reach CD4 ≥200 cells/mm³. In a multivariate Cox regression model (Model 2, Table 2), not reaching immune restoration during follow-up was independently associated with a higher risk of disease progression, compared with reaching CD4 ≥200 cells/mm³ in <6 months (HR = 4.52, 95% CI = 3.29–6.19, P < 0.001), although a non-significant trend was observed when comparing this with 6–12, 12–24 and >24 months to immune restoration (HR = 1.46, 95% CI = 0.98–2.18, P = 0.064; HR = 1.45, 95% CI = 0.98–2.14, P = 0.061; and HR = 1.34, 95% CI = 0.95–1.90, P = 0.093, respectively).

Finally, in order to assess the possible effect in the risk of progression to AIDS/death of CD4 and VL separately, both at BL and follow-up, a multivariate Cox regression model approach was used (Model 3, Table 2). The following variables were independently associated with progression to AIDS/death: BL CD4 (51–100 cells/mm³: HR = 1.36, 95% CI = 1.06–1.75 and 0–50 cells/mm³: HR = 1.25, 95% CI = 0.99–1.58 versus 101–199 cells/mm³: P = 0.044), BL VL >5 log copies/mL (HR = 1.51, 95% CI = 1.23–1.86, P < 0.001), IR during follow-up as time dependent (not reaching CD4 ≥200 cells/mm³: HR = 4.86, 95% CI = 3.78–6.23, P < 0.001) and VR during follow-up as time dependent (detectable VL: HR = 1.99, 95% CI = 1.63–2.42, P < 0.001).

In all three multivariate Cox regression model approaches, age (>50 years) and AIDS at BL were significantly associated with disease progression (see Table 2).

All previous analyses were replicated for patients without prior AIDS (n = 1433), obtaining similar results (data not shown). In the multivariate Cox regression model, CD4 count <200 cells/mm³ during follow-up as a time-dependent variable (HR = 3.55, 95% CI = 2.57–4.92, P < 0.001) was also the strongest factor associated with AIDS/death.

Discussion

Our data show that in severely immunosuppressed patients initiating ART, reaching CD4 ≥200 cells/mm³ (immunological response) and undetectable VL (virological response) during follow-up is associated with a very low rate of disease progression, regardless of BL status. In contrast, a majority of patients not reaching this virological or immunological outcome will experience disease progression. Intermediate between these extremes are reaching CD4 ≥200 cells/mm³, but not undetectable VL, and achieving virological suppression, but not CD4 restoration, although the former has a better clinical outcome. Similar results were obtained when only death was considered as the endpoint.

Our study partially supports previous data highlighting the value of CD4 increases by 6 months.¹⁰ Disease progression was significantly reduced in patients reaching CD4 count ≥200 cells/mm³ before 6 months of follow-up compared with patients not reaching CD4 ≥200 cells/mm³ during the study period. A non-significant trend was observed when comparing reaching CD4 count ≥200 cells/mm³ in <6 months versus after 6, 12 and 24 months.

Several variables were independently associated with disease progression, including age, VL at BL and follow-up and CD4 count <200 cells/mm³ during follow-up, but the latter had the greatest impact with HR of 4.86 and 3.55 in patients with or without a diagnosis of AIDS at BL, respectively.

Antiretroviral regimens achieving more rapid viral suppression and improved CD4 recovery may be associated with reduced disease progression. However, it is difficult to confirm this important observation in clinical trials because international guidelines in recent years recommend early ART initiation and the number of enrolled patients with CD4 counts <200 cells/mm³ is very low.^11,12

A lengthy, outstanding paper has recently evaluated the long-term mortality of HIV-infected patients with incomplete CD4 recovery, a proportion of whom remained with a CD4 count <200 cells/mm³.¹³ Our sample is smaller, but by virtue of the different design, we add further information not provided by that study. As with other studies,¹⁴ they only included patients with viral suppression and did not study the role of the combined effect of CD4 and VL changes or measure outcomes according to different scenarios related to BL CD4 and VL. Our design allows us to evaluate different situations that occur in real life and helps to elucidate the expected morbidity and mortality in patients who do not achieve viral suppression and/or CD4 recovery due to poor adherence, older age, etc.

Although the role and periodicity of CD4 monitoring in the management of HIV-infected patients has become controversial, it has been recognized that in late presenters CD4 count may be helpful for several reasons such as prophylaxis of opportunistic infections.¹⁵ Our data support the use of CD4 determinations in severely immunosuppressed patients as a prognostic marker of disease progression and survival, importantly contributing to the planification of health resources and clinical management of this special population.

Our study has several important limitations. Therapy discontinuation or changes as well as adherence were not recorded. Until recently, integrase inhibitors were not given as first-line therapy in Spain. Several studies have shown better virological and/or immunological response as well as better tolerability—a crucial aspect in advanced patients—with integrase inhibitor-containing regimens compared with PI- or NNRTI-containing regimens.^16,17 Thus, our data might be further improved in patients treated with this new ART class. Eighty percent of patients were men, impeding the extrapolation of our results to women.

As we have only analysed progression to AIDS-defining diseases and death, important outcomes in immunosuppressed HIV-infected patients such as non-AIDS-defining comorbidities are not known in our cohort.¹⁸ In addition, causes of AIDS/death are not provided in this study.

Despite these limitations, our study provides important data on the role of CD4 count and VL at BL and during follow-up and other clinical factors in disease progression. These results, prospectively obtained in routine clinical practice, may contribute to the knowledge of the expected clinical course and antiretroviral response in patients who are severely immunosuppressed at the time of ART initiation in the real world.

In summary, in severely immunosuppressed patients starting ART, achieving a CD4 ≥200 cells/mm³ is the main predictor of decreased progression to AIDS/death. However, in those not reaching this CD4 threshold, virological suppression is sufficient to reduce disease progression by half. Finally, even in the worst scenario of CD4 ≤100 cells/mm³ and/or high VL at BL, virological suppression and immune reconstitution are associated with a dramatic decrease in progression to AIDS/death. The current availability of newer, more effective and better-tolerated ART regimens may even improve upon the present results. In addition, therapies restoring the number and function of CD4 cells are needed to further increase the survival of patients presenting to medical care at advanced stages of HIV disease. More importantly, great efforts should be made to increase early diagnosis of HIV infection.

PISCIS Investigators

Coordination: J. Casabona (Centre d'Estudis Epidemiològics sobre les Infeccions de Transmissió Sexual i Sida de Catalunya: CEEISCAT) and Jose M. Miró (Hospital Clinic-IDIBAPS, Universitat de Barcelona). Field coordinator: C. N. J. Campbell (CEEISCAT). Steering committee: J. Casabona, A. Esteve, C. N. J. Campbell (CEEISCAT), J. M. Miró (Hospital Clinic-IDIBAPS, Universitat de Barcelona), D. Podzamczer (Hospital Universitari de Bellvitge-IDIBELL) and J. Murillas (Hospital Son Espases de Mallorca). Scientific committee: J. M. Gatell, C. Manzardo (Hospital Clinic-IDIBAPS, Universitat de Barcelona), C. Tural, B. Clotet (Fundació Lluita contra la Sida, Fundació irsicaixa, Hospital Universitari Germans Trias i Pujol, Universitat Autónoma de Barcelona), E. Ferrer (Hospital Universitari de Bellvitge-IDIBELL), M. Riera (Hospital Son Espases de Mallorca), F. Segura, G. Navarro (Corporació Sanitària i Universitària Parc Taulí, Universitat Autónoma de Barcelona), L. Force (Hospital de Mataró, Consorcio Sanitario del Maresme), J. Vilaró (Hospital General de Vic), A. Masabeu (Hospital de Palamós), I. García (Hospital General d'Hospitalet), J. Mercadal (Hospital Comarcal de l'Alt Penedès), C. Cifuentes, F. Homar (Hospital Son Llàtzer), D. Dalmau, À. Jaen (Hospital Universitari Mútua de Terrassa), P. Domingo (Hospital de la Santa Creu i Sant Pau), V. Falcó, A. Curran (Hospital Universitari Vall d'Hebron) and C. Agustí (CEEISCAT). Data management and statistical analysis: A. Esteve, A. Montoliu (CEEISCAT), I. Pérez (Hospital Clinic-IDIBAPS, Universitat de Barcelona) and J. Curto (Hospital Universitari de Bellvitge-IDIBELL). Informatics: F. Gargoulas (Hospital Son Espases y Hospital Son Llàtzer), A. Gómez (Hospital Comarcal de l'Alt Penedès) and J. C. Rubia (Hospital General d'Hospitalet). Participating physicians and centres: L. Zamora, J. L. Blanco, F. Garcia-Alcaide, E. Martínez, J. Mallolas (Hospital Clinic-IDIBAPS, Universitat de Barcelona), J. M. Llibre, G. Sirera, J. Romeu, A. Jou, E. Negredo (Fundació Lluita contra la Sida, Hospital Universitari Germans Trias i Pujol, Universitat Autónoma de Barcelona), M. Saumoy, A. Imaz, F. Bolao, C. Cabellos, C. Peña, S. DiYacovo, E. Van Den Eynde (Hospital Universitari de Bellvitge-IDIBELL), M. Sala, M. Cervantes, M. J. Amengual, M. Navarro, V. Segura (Corporació Sanitària i Universitària Parc Taulí, Universitat Autónoma de Barcelona,) P. Barrufet (Hospital de Mataró, Consorcio Sanitario del Maresme), J. Molina, M. Alvaro, M. Orriols (Hospital Alt Penedès de Vilafranca), T. Payeras (Hospital Son Llàtzer) and Mª. Gracia Mateo (Hospital de la Santa Creu i Sant Pau). Civil society representative: J. Fernández (Comitè 1er de Desembre).

Funding

The PISCIS Cohort was supported by the Fundación para la Investigación y Prevencion del Sida en España (FIPSE, Madrid Spain, National AIDS Plan Secretariat of the Spanish Ministry of Health), grants 3084/99, 36354/02, 36488/05. The study received support from the Fondo de Investigaciones Sanitarias (FIS), Spain, grant 1520/97 and was partially funded by the RD12/0017/0013 project as part of the Plan Nacional R+D+I and cofinanced by the ISCIII, Subdirección General de Evaluación y el Fondo Europeo de Desarrollo Regional (FEDER). The PISCIS Cohort is currently supported by the Agencia de Salut Publica de Catalunya (ASPC), Generalitat de Catalunya, Spain.

Transparency declarations

None to declare.

Author contributions

Study design, data interpretation, writing and literature search: E. F., J. Curto, A. E. and D. P. Data analysis and figure: J. Curto and A. E. Study design and data interpretation: J. Casabona. Data collection: E. F., J. M. M., C. T., J. M., F. S., P. B. and D. P. Review of the manuscript draft: E. F., J. Curto, A. E., J. M. M., C. T., J. M., F. S., P. B., J. Casabona and D. P.

Acknowledgements

Presented in part at the Eleventh International Congress on Drug Therapy in HIV Infection, Glasgow, 2012 (Poster P-169).

We are indebted to the study participants and to the staff of the HIV units at the study centres for retrieving detailed data on HIV infection.

References

Author notes