In recent years, a test and treat approach to HIV prevention has garnered considerable interest. After HIV diagnosis (“test”), timely initiation and uninterrupted receipt of antiretroviral therapy (ART, “treat”), taken at a high level of adherence, can suppress plasma HIV viral load (VL) thereby significantly reducing the likelihood of transmission.1–3 Implicit to the success of a test and treat strategy is the need for prompt linkage and sustained retention in HIV medical care after diagnosis. Currently, barriers exist at each step of the treatment cascade, from HIV testing and awareness of serostatus to VL suppression, that limit the effectiveness of test and treat in local communities and nationally.4,5 An estimated 21% of HIV-infected individuals in the United States are unaware of their HIV serostatus and 31% of newly diagnosed persons delay linkage to outpatient HIV medical care for 6 months or longer.6,7 Even with initial linkage into care, subsequent retention in HIV care is required for full access to treatment benefits. Early missed visits are common and have been associated with delayed initiation of ART in the short term and adverse clinical outcomes in the long term.8,9 One-year attrition has been observed in upwards of 25%–30% of patients after initial enrollment in outpatient HIV care, compromising the potential real world impact of treatment as prevention strategies.10,11
Although numerous initiatives have focused on expanding HIV testing, and a brief case-management intervention has proven efficacious for linking newly diagnosed patients to care,11,12 limited evidence, both observational and interventional, exists for the period immediately after entry to care.9,10,13,14 The year after initial linkage to outpatient HIV medical care is a dynamic and formative time for patients adjusting to a life changing diagnosis. It is also a time of considerable vulnerability. Patients, many of whom have limited experience navigating the health care system and may not be fully prepared for a long-term commitment to taking prescribed medications, are called upon to attend frequently scheduled medical visits and initiate treatment with ART regimens that require high levels of sustained adherence to achieve success (eg, viral suppression) and avoid harm (eg, resistance mutations).9,10 Suboptimal early retention in care represents a formidable obstacle to achieving HIV VL suppression, which has important individual and public health implications, yet little is known about its impact on the ultimate goal of test and treat strategies, viral suppression.
Here, we address this gap in the literature through the careful evaluation of the effect of early retention in HIV care on VL suppression. In addition to evaluating time to virologic suppression, we evaluate cumulative VL burden over the first 2 years in care. Beyond any single time point, longitudinal plasma viremia has important implications for patient outcomes and for HIV transmission. We use a novel measure, viremia copy-years, an area under the curve estimate of cumulative plasma VL exposure,15 to characterize the relation between early retention in care and longitudinal VL burden. We hypothesized that early missed HIV care visits would be associated with delayed VL suppression, and that patients demonstrating poorer retention in care as measured by visit adherence would have higher cumulative viremia copy-years over the first 2 years of HIV care.
Sample and Procedures
HIV-infected patients initiating outpatient HIV medical care at 2 academically affiliated HIV treatment centers at the University of Alabama at Birmingham (UAB 1917 Clinic) and the University of Washington (UW Harborview Clinic) between January 2007 and September 2010 were included in the overall study sample. Patients who had received outpatient HIV medical care at another facility before enrollment at the 2 study sites were excluded. To allow for a complete 2-year observation period, only the subset of patients initiating care before September 2008 were included in analyses of cumulative VL burden during the first 2 years after entry into outpatient HIV medical care. Data were captured through the Centers for AIDS Research Network of Integrated Clinical Systems (CNICS), a nationally distributed HIV clinical cohort that has been described in detail previously.16 Briefly, every 2 months, sites transmit comprehensive and well-defined data elements captured from point-of-care electronic health record systems using standardized terminology and format. Systematic and rigorous processes for data verification and quality assurance are in place to generate a centralized high quality clinical database. This study and the CNICS protocol were approved by local institutional review boards at both study sites.
Outcome Measures: Plasma VL Suppression After Entry into Outpatient HIV Medical Care
Plasma HIV VL suppression after entry into care was evaluated using 2 distinct measures: time to suppression of plasma HIV RNA <50 copies per milliliter and cumulative VL burden measured over the 2 years after the initial medical visit. Viremia copy-years, a time-varying measure of cumulative plasma HIV burden calculated using methods described in detail previously, was used to evaluate longitudinal VL burden.15 Briefly, the trapezoidal rule is used to approximate the integral representing the area under each patient's longitudinal VL curve. VL burden for each segment (time interval between 2 consecutive VL values) is calculated by multiplying the mean of the 2 values by the time interval. The copy × years per milliliter for each segment of a patient's VL curve are summed to calculate viremia copy-years. Formally, viremia copy-years is the number of copies of HIV RNA per milliliter of plasma over time. For example, 10,000 copy-years of viremia equals having a VL of 10,000 copies per milliliter for 1 year, or alternatively, a VL of 1000 copies per milliliter for 10 years. For the current study, viremia copy-years was calculated for each patient for 2 years starting from each patient's initial clinic visit date. Each patient's baseline VL was assigned to the initial medical visit date, and all VL values over the subsequent 2 years (730 days) were used in the viremia copy-years calculation (Fig. 1). Patients with a gap of >365 days between VL measures (n = 78) were excluded due to concerns about the reliability of calculating VL segments separated by long periods of time. Plasma VL values >1,000,000 copies per milliliter were administratively assigned a value of 1,000,000 copies per milliliter, and values below the limits of assay detection (49 copies/mL) were assigned a value of 24 copies per milliliter. Two-year viremia copy-years were calculated for all patients using natural nontransformed plasma HIV VL values. Subsequently, the 2-year viremia copy-year measure was log10 transformed for each patient. The 2-year log10 viremia copy-years measure was used for all analyses.
Principal Exposure of Interest: Early Retention in Care
Early retention in care was measured as a time-varying count of “no show” visits for time to VL suppression analyses (<50 copies/mL), and as visit adherence, the proportion of scheduled visits that were attended, for the evaluation of 2-year viremia copy-years.17 Different measures were selected to best capture short-term and long-term retention in care in accordance with the outcome measures and analytic plan. During shorter observation periods fewer visits are available analytically, as observed for the time to VL suppression analyses (median = 5 visits, interquartile range = 3–9 visits), guiding our a priori selection of the “no show” count measure for these analyses. Additionally, the “no show” count measure functions well as a time-varying covariate in survival analyses.17 During the 2-year observation period for the cumulative viremia copy-years analyses, more HIV medical visits were scheduled (median = 10 visits, interquartile range = 8–16 visits), allowing a sufficient denominator to draw meaningful inference from the visit adherence measure. The visit adherence measure allows for more detailed evaluation of exposure–response relationships between retention in care and 2-year VL burden across a wider range (proportion bound by 0, 1) relative to the “no show” count measure and may be preferable for longer observation periods.17 Only scheduled visits with a primary HIV medical care provider were included. In accordance with prior studies,17–20 visits cancelled in advance by the patient and visits cancelled by the clinic and/or medical provider were excluded. In separate analyses, visit adherence was evaluated as a continuous measure and as a categorical variable; 0%–79%, 80%–99%, and 100% adherence, which roughly represented tertiles of the study sample.
Additional measures selected a priori included age, sex, race/ethnicity, health insurance, baseline CD4+ T-lymphocyte count, and baseline plasma HIV VL. Baseline plasma HIV VL and CD4+ T-lymphocyte count measurements were the value on the date nearest the initial medical visit date within a window of −180 to +7 days; median ± interquartile range of −3 (−23, 0) and −5 (−23, 0) days, respectively, from initial visit date.
Descriptive statistics including means, standard deviations, medians, interquartile range, counts, and percents were calculated for all study variables as appropriate. Survival methods including Kaplan–Meier plots and Cox proportional hazards models were used to evaluate factors associated with time to first VL suppression (<50 copies/mL). Analysis of variance, linear regression, and analysis of covariance were used to evaluate factors associated with 2-year log10 viremia copy-years. All adjusted models controlled for covariates as outlined in the previous section. Sensitivity analyses included patients with a baseline CD4+T-lymphocyte count <350 cells per milliliter in accordance with the recommended threshold for ART initiation during the majority of the study period. Additional sensitivity analyses replaced the visit adherence measure with the “no show” count measure in the evaluation of 2-year viremia copy-years. All statistical analyses were performed using SAS version 9.2 (SAS Institute Inc, Cary, NC).
Among 676 patients, the mean ± standard deviation age at entry into care was 36 ± 11 years, 44% were non-white males, 13% were non-white females, 38% were white males, and 36% were uninsured (Table 1). The mean log10 baseline VL was 4.56 ± 0.97 log10 copies per milliliter, and 33% and 43% of patients had baseline CD4+ T-lymphocyte counts <200 cells per milliliter and >350 cells per milliliter, respectively. Twenty-five percent of patients had 2 or more no show visits, 79% initiated ART (median = 35 days from initial visit, interquartile range: 14–105 days), and 63% of the overall sample achieved VL suppression (<50 copies/mL) in a median 308 days (Kaplan–Meier estimate, interquartile range: 268–343 days) from entry into care. In adjusted Cox proportional hazards analysis, patients with private insurance [hazard ratio (HR) = 1.36 versus uninsured; 95% confidence interval (CI) = 1.08 to 1.73] and lower baseline CD4+ T-lymphocyte counts (<200 cells/mL: HR = 3.73; 95% CI = 2.82 to 4.93, and 200−350 cells/mL: HR = 2.96; CI = 2.26 to 3.86 versus >350 cells/mL) at entry into care experienced shorter times to VL suppression below 50 copies per milliliter (Table 2). Patients with higher baseline VLs (HR = 0.97 per additional 50,000 copies per milliliter; 95% CI = 0.96 to 0.99) and with more no show clinic visits (HR = 0.84 per additional no show visit; 95% CI = 0.76 to 0.92) experienced significantly longer time to VL suppression.
Two hundred fifty-eight patients with 2 years of follow-up from clinic entry were included in the evaluation of cumulative VL burden measured by viremia copy-years, as a function of early retention in care measured by visit adherence. In general, sociodemographic characteristics of these patients were comparable with the overall sample (Table 1). Over the first 2 years in outpatient HIV medical care, nearly half of patients (48%) had 2 or more no show clinic visits. The mean ± standard deviation visit adherence was 0.84 ± 0.16, and the following visit adherence categories were observed among patients: 0%–79% adherence in 83 patients (32%), 80%–99% adherence in 95 patients (37%), and 100% adherence in 80 patients (31%). Eighty-seven percent of patients started ART during the 2-year period, with 81% achieving a VL <50 copies per milliliter on at least 1 occasion. The mean ± standard deviation 2-year log10 viremia copy-years was 4.3 ± 0.8 copy x years per milliliter. Across increasing visit adherence categories, lower 2-year log10 viremia copy-years (mean ± standard deviation log10 copy × years/mL) was observed: 0%–79% adherence: 4.6 ± 0.8 (β = 0.47; 95% CI = 0.24 to 0.70), 80%–99% adherence: 4.3 ± 0.7 (β = 0.18; 95% CI = −0.04 to 0.40), and 100% adherence: 4.1 ± 0.7 log10 copy × years per milliliter (referent), respectively (Table 3). In multivariable linear regression, lower visit adherence (β = −0.11 per 10% visit adherence; 95% CI = −0.17 to −0.04) was associated with greater 2-year viremia copy-years.
Sensitivity analyses restricted to patients with baseline CD4+ T-lymphocyte counts <350 cells per milliliter on clinic entry (n = 372) yielded comparable findings to primary analyses. Ninety-seven percent of patients (n = 361) started ART with 76% (n = 283) achieving a VL <50 copies per milliliter in a median 192 days (Kaplan–Meier estimate, interquartile range: 175–211 days) from entry to care. In adjusted Cox proportional hazards analysis, patients with more no show clinic visits (HR = 0.81 per additional no show visit; 95% CI = 0.71 to 0.92) experienced significantly delayed time to VL suppression. As observed in primary analyses, visit adherence demonstrated an inverse relationship with 2-year log10 viremia copy-years in multivariable linear regression (β = −0.14 per 10% visit adherence; 95% CI = −0.19 to −0.08). Finally, when replacing the visit adherence measure of early retention in care with the “no show” count measure, a comparable relationship was observed with greater 2-year log10 viremia copy-years among patients with more missed visits (β = 0.11 per “no show” visit; 95% CI = 0.08 to 0.14).
Early retention in HIV care was associated with time to VL suppression (<50 copies/mL) and 2-year cumulative VL burden among patients newly initiating outpatient HIV medical care. Each “no show” clinic visit conveyed a 17% increased risk of delayed VL suppression. Significantly greater viremia copy-years, an estimate of cumulative HIV burden, were accumulated among patients with poorer visit adherence over the first 2-years in care (4.6 ± 0.7 vs. 4.1 ± 0.8 log10 copy × years per milliliter for 0%–79% vs. 100% visit adherence). These findings have implications for patient outcomes, as recent studies have identified increased risk of deleterious clinical events among patients experiencing greater cumulative VL burden over time.21,22
Beyond individual health consequences, these findings have implications for HIV prevention efforts, particularly in the context of a test and treat approach to HIV prevention.1,2,5 Failure to achieve and sustain VL suppression poses increased risk of HIV transmissibility.3,23 In the current study, we identified poor early retention in HIV care as a barrier to timely VL suppression and a factor associated with greater cumulative VL burden over the first 2 years after entry into care. A recent study by Metsch et al identified reductions in sexual risk transmission behaviors among recently diagnosed patients with better early retention in HIV care relative to patients who were poorly retained.24 Taken together, these 2 studies highlight the vital role of early retention in HIV care as it relates to sexual risk behavior and VL suppression—critical elements in secondary HIV transmission, making early retention in care an important target for prevention interventions.
Previously, we identified significant independent associations between viremia copy-years and clinical outcomes among HIV-infected patients. Among Multicenter AIDS Cohort Study HIV-uninfected participants who seroconverted, viremia copy-years was associated with AIDS-defining clinical events and mortality in the precombination ART era (1984–1996).15 Among CNICS patients initiating modern ART regimens (nonnucleoside reverse transcriptase inhibitor–based or protease inhibitor/ritonavir-based regimens) in recent years (2000–2008), viremia copy-years was independently associated with all-cause mortality when controlling for traditional cross-sectional VL measures and time-updated CD4+ T-lymphocyte count.25
Here, we extend the use of this novel, area under the curve estimate of cumulative VL burden by evaluating viremia copy-years as an outcome variable for the first time. Beyond single cross-sectional measures of VL suppression, we posit longitudinal measures of cumulative viremia may better predict individual risk of HIV transmission and may augment the evaluation of treatment as prevention interventions, the success of which is predicated on not merely achieving, but subsequently sustaining VL suppression. Moreover, we suggest viremia copy-years should be explored as a population-level indicator of transmission risk compared to single-measure indicators, such as community VL or the cross-sectional proportion of patients suppressed at the most recent measure or other single time point.
Our study has limitations. As with all observational studies, we can identify associations but not ascribe causality because our findings are subject to possible uncontrolled confounding. Patients had variable numbers and timing of VL measures, reflecting the realities of HIV clinical care. This may impact the precision of the viremia copy-years estimate of cumulative HIV burden. Patients with >365 days between VL measures were excluded to address this limitation. Although conducted at 2 sites in distinct regions of the United States, study findings may not be generalizable to other regions or settings. Future studies are planned to evaluate the generalizabilty of these findings to the larger nationally distributed CNICS cohort (8 sites). In this initial analysis, we did not formally evaluate the timing of ART receipt, ART persistence (or durability), and ART adherence in multivariable analyses, as these factors are on the causal pathway between early retention in care and VL suppression.4,10 Here we sought to first establish the relationship between early retention in outpatient HIV care and VL outcomes. However, in exploratory analyses, the relationship between the number of “no show” visits and VL suppression persisted after controlling for initiation of ART as a dichotomous time-varying covariate (data not shown). Future studies using methods of causal inference (eg, structural equation models or marginal structural models) are planned to evaluate the intricacies of the dynamic relationships among variables on the causal pathway from entry to HIV care to longitudinal VL suppression, including early retention in care, initial ART receipt, ART persistency, and ART adherence. Finally, we did not systematically capture substance abuse, mental illness, and other psychosocial factors at care initiation such that we could not evaluate the role of these important and prevalent factors on early retention in care and VL suppression.
In conclusion, we identified significant associations between early retention in care and VL suppression among patients initiating outpatient HIV medical care. Beyond delayed time to VL suppression (<50 copies/mL) among patients with more “no show” visits, we demonstrated the importance of early visit adherence as it relates to cumulative HIV burden, measured by viremia copy-years. Longitudinal measures of cumulative VL burden, like viremia copy-years, may significantly contribute to the evaluation of test and treat HIV prevention interventions, the success of which are predicated on both rapidly achieving, and also longitudinally sustaining VL suppression.
We thank the patients, providers, clinical and research personnel at the UAB 1917 Clinic and the UW Harborview Clinic and the CNICS administrative and data management teams. We thank Stephen Cole, Sonia Napravnik, Joseph Eron, and Bryan Lau for their collaboration in developing the viremia copy-years measure.
1. Dieffenbach CW, Fauci AS. Universal voluntary testing and treatment for prevention of HIV
transmission. JAMA. 2009;301:2380–2382
2. Granich RM, Gilks CF, Dye C, et al. Universal voluntary HIV
testing with immediate antiretroviral therapy as a strategy for elimination of HIV
transmission: a mathematical model. Lancet. 2009;373:48–57
3. . Initiation of antiretroviral treatment protects uninfected sexual partners from HIV
infection (HPTN Study 052) May 12, 2011 Available at: http://www.hptn.org/web%20documents/PressReleases/HPTN052PressReleaseFINAL5_12_118am.pdf
. Accessed: May 17, 2011
4. Mugavero MJ, Norton WE, Saag MS. Health care system and policy factors influencing engagement in HIV
medical care: piecing together the fragments of a fractured health care delivery system. Clin Infect Dis. 2011;52(suppl 2):S238–S246
5. Gardner EM, McLees MP, Steiner JF, et al. The spectrum of engagement in HIV
care and its relevance to test-and-treat strategies for prevention of HIV
infection. Clin Infect Dis. 2011;52:793–800
6. Campsmith ML, Rhodes PH, Hall HI, et al. Undiagnosed HIV
prevalence among adults and adolescents in the United States at the end of 2006. J Acquir Immune Defic Syndr. 2010;53:619–624
7. Marks G, Gardner LI, Craw J, et al. Entry and retention in medical care among HIV
-diagnosed persons: a meta-analysis. AIDS. 2010;24:2665–2678
8. Giordano TP, White AC Jr, Sajja P, et al. Factors associated with the use of highly active antiretroviral therapy in patients newly entering care in an urban clinic. J Acquir Immune Defic Syndr. 2003;32:399–405
9. Mugavero MJ, Lin HY, Willig JH, et al. Missed visits and mortality among patients establishing initial outpatient HIV
treatment. Clin Infect Dis. 2009;48:248–256
10. Ulett KB, Willig JH, Lin HY, et al. The therapeutic implications of timely linkage and early retention in HIV
care. AIDS Patient Care STDS. 2009;23:41–49
11. Gardner LI, Metsch LR, Anderson-Mahoney P, et al. Efficacy of a brief case management intervention to link recently diagnosed HIV
-infected persons to care. AIDS. 2005;19:423–431
12. Craw JA, Gardner LI, Marks G, et al. Brief strengths-based case management promotes entry into HIV
medical care: results of the antiretroviral treatment access study-II. J Acquir Immune Defic Syndr. 2008;47:597–606
13. Olatosi BA, Probst JC, Stoskopf CH, et al. Patterns of engagement in care
-infected adults: South Carolina, 2004–2006. AIDS. 2009;23:725–730
14. Torian LV, Wiewel EW. Continuity of HIV
-related medical care, New York City, 2005-2009: Do patients who initiate care stay in care? AIDS Patient Care STDS. 2011;25:79–88
15. Cole SR, Napravnik S, Mugavero MJ, et al. Copy-years viremia as a measure of cumulative human immunodeficiency virus viral burden. Am J Epidemiol. 2010;171:198–205
16. Kitahata MM, Rodriguez B, Haubrich R, et al. Cohort profile: the Centers for AIDS Research Network of Integrated Clinical Systems. Int J Epidemiol. 2008;37:948–955
17. Mugavero MJ, Davila JA, Nevin CR, et al. From access to engagement: measuring retention in outpatient HIV
clinical care. AIDS Patient Care STDS. 2010;24:607–613
18. Keruly JC, Conviser R, Moore RD. Association of medical insurance and other factors with receipt of antiretroviral therapy. Am J Public Health. 2002;92:852–857
19. Melnikow J, Kiefe C. Patient compliance and medical research: issues in methodology. J Gen Intern Med. 1994;9:96–105
20. Robbins GK, Johnson KL, Chang Y, et al. Predicting virologic failure in an HIV
clinic. Clin Infect Dis. 2010;50:779–786
21. Cole SR, Napravnik S, Mugavero MJ, et al. Copy-years viremia as a measure of cumulative human immunodeficiency virus viral burden. Am J Epidemiol. 2010;171:198–205
22. Marconi VC, Grandits G, Okulicz JF, et al. Cumulative viral load
and virologic decay patterns after antiretroviral therapy in HIV
-infected subjects influence CD4 recovery and AIDS. PLoS One. 2011;6:e17956
23. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load
and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med. 2000;342:921–929
24. Metsch LR, Pereyra M, Messinger S, et al. HIV
transmission risk behaviors among HIV
-infected persons who are successfully linked to care. Clin Infect Dis. 2008;47:577–584
25. Mugavero M, Napravnik S, Cole S, et al.CFAR Network of Integrated Clinical Systems. Viremia copy-years predicts mortality among treatment-naïve HIV
-infected patients initiating antiretroviral therapy. Clin Infect Dis. [Epub ahead of print September 2, 2011] doi:10.1093/cid/cir526