Recent clinical trials have demonstrated that HIV transmission can be reduced in discordant couples by control of viremia in the infected partner.1–5 Ecologic analyses suggest that the concept can be extended to whole populations,6 and mathematical models predict that a “test-and-treat” strategy incorporating universal testing, immediate initiation of antiretroviral therapy (ART), and population-level virologic suppression would ultimately eliminate incidence.7–10 A “real-world” evaluation of these predictions would require following patients for periods consistent with the expected survival of persons living with HIV, including patients at all stages of disease, not just those currently eligible for and receiving treatment,1,7,8,11,12 and looking beyond the research setting, where proactive follow-up, monitoring and adherence support are customary. Despite their limitations, routinely collected public health surveillance data are uniquely positioned to provide this type of analysis. They are population based, standardized, transparent, and reproducible. They include all laboratory reports on all diagnosed patients. They assume that the majority of undetectable viral loads (VL) are treatment mediated and that all undiagnosed and out-of-care patients are viremic. Therefore, they can establish baselines and estimate what might realistically be expected from the treatment component of a local test and treat strategy.13,14 We used the New York City HIV/AIDS Reporting System (NYC HARS) to evaluate viral suppression among persons newly diagnosed with HIV between 2006 and 2009.
The NYC HARS is a population-based registry that has existed since 1981. It contains named reports of all diagnoses of HIV and AIDS, positive Western blot tests for HIV antibody, all VL and CD4 counts, and HIV genotypes. It is continuously updated with newly confirmed diagnoses and laboratory results, and vital status on new and existing case records. As of June 30, 2011, NYC HARS contained a cumulative total of 217,283 cases and more than 6 million laboratory tests.
The population analyzed included all persons newly diagnosed with HIV between January 1, 2006, and December 31, 2009, and reported to HARS by June 30, 2011 (N = 16,038). We excluded persons who died within 1 month of HIV diagnosis (n = 240), who were younger than 18 years (n = 275), or not NYC residents (n = 1237) at the time of diagnosis, the last on the premise who were likely to obtain care in their jurisdiction of residence, and we would not receive their laboratory reports. To minimize misclassification of new diagnoses, we further excluded those with (1) no Western blot test within 3 months of HIV diagnosis (n = 1459), (2) a detectable VL (>50 copies/mL) more than 1 month earlier than HIV diagnosis (n = 9), and (3) viral suppression within 1 month of HIV diagnosis (n = 696). All patients who were known to have acute HIV infection at the time of diagnosis were included regardless of the date and result of their Western blot and VL assay(s). These inclusion and exclusion criteria resulted in a final analytic population of 12,122.
We defined achievement of viral suppression as the first HIV RNA level of <400 copies per milliliter and time to viral suppression as the number of days between the initial diagnosis of HIV and first VL <400 copies per milliliter. We classified CD4 count at diagnosis into 4 intervals: <200, 200–349, 350–499, and ≥500 cells per cubic millimeter, based on the first CD4 count after the initial diagnosis and before viral suppression.11 For patients whose first CD4 count was reported more than 1 year after the diagnosis but before viral suppression, CD4 count at diagnosis was imputed by adding 50 cells per cubic millimeter per 12-month interval between diagnosis date and first CD4 date15; for patients who had no CD4 count documented in HARS or whose first CD4 test in HARS was after viral suppression, CD4 count at diagnosis was not imputed, and a missing value was entered. We “started the clock” on achievement of viral suppression on the date of diagnosis as opposed to the date of the nadir CD4 because the purpose of this analysis was to gauge how rapidly new cases achieve suppression.
We defined virologic failure as any VL ≥ 1000 copies per milliliter after 30 days of viral suppression (VL < 400 copies/mL).16 Duration of suppression was defined as time in days between date of first VL <400 copies per milliliter and date of first VL ≥ 1000 copies per milliliter.1,16 Surveillance, which does not ascertain the initiation of ART, presumes that all persons in care will have ART initiated at the Department of Health and Human Services CD4 count threshold in effect during the year of analysis. Thus, nadir CD4 count before viral suppression was classified into 4 intervals: < 200, 200–349, 350–499, and ≥500 cells per cubic millimeter. Patients lacking a CD4 count value before viral suppression were assigned missing values. We started the clock on maintenance of viral suppression on the date of the first VL value that was <400 copies per milliliter.
We compared the demographic characteristics, risk factors, and CD4 count of new HIV diagnoses by the year of diagnosis, using the Pearson χ2 test for categorical variables (sex, race, and transmission risk), analysis of variance for age, and the Kruskal–Wallis test for CD4 count at diagnosis. We calculated the mean, median, and interquartile ranges of the continuous variables.
We described the proportion of new diagnoses achieving viral suppression at 6 and 12 months after the diagnosis by sex, race, age, transmission risk, year of diagnosis, and CD4 count at diagnosis. The Kaplan–Meier product limit method was used to estimate the cumulative proportion achieving viral suppression after HIV diagnosis, stratified by year of diagnosis, and CD4 count at diagnosis, with log rank tests to assess differences across groups. Patients were censored on the date of death, on the date of viral suppression, or at the end of the analysis period, whichever came first. Patients who never had a VL test after diagnosis or who had at least one VL test but never had one <400 copies per milliliter by the time of death or end of the analysis period were considered to be not virologically suppressed. Multivariate Cox proportional hazard regression with 6 independent variables (sex, race, age, transmission risk, year of diagnosis, and CD4 count at diagnosis) was used to identify factors associated with time to achieve viral suppression and to calculate the respective hazard ratios (HRs).
We described the proportion of patients maintaining viral suppression at 6 and 12 months after the first VL <400 copies per milliliter, by sex, race, age, transmission risk, year of diagnosis, and nadir CD4 count before viral suppression, among those who ever achieved viral suppression. The Kaplan–Meier product limit method was used to estimate the cumulative proportion maintaining viral suppression, stratified by the year of diagnosis and nadir CD4 count before viral suppression, with log rank tests to assess differences across groups. Patients were censored on their date of death, on the date of virologic failure, or at the end of the analysis period, whichever came first. Patients who never had a VL ≥ 1000 copies per milliliter after viral suppression by the time of death or at the end of the analysis period were considered to have maintained suppression. Multivariate Cox proportion hazard regression analysis with 6 independent variables (sex, race, age, transmission risk, year of diagnosis, and nadir CD4 count before viral suppression) was used to identify factors associated with the time to virologic failure and to calculate the respective HRs. HRs for maintaining viral suppression were calculated and presented by taking the inverse of HRs for virological failure. All analyses were conducted using the Statistical Analysis System (SAS 9.2; SAS Institute, Cary, NC).
Demographic Characteristics, VL, and CD4 Count at Diagnosis
The population was 74.3% male, 15.4% white, 56.8% aged 25–44 years, 41.7% men who have sex with men (MSM), and 6.4% injecting drug users (Table 1). From 2006 to 2009, the population contained progressively more men and more MSM. Median (interquartile range) CD4 count (in cells per cubic millimeter) at diagnosis was 326 (135–503) in 2006, 324 (140–505) in 2007, 332 (151–522) in 2008, and 346 (168–525) in 2009 (P = 0.007). Applying the pre-2011 ART initiation threshold of CD4 <350 cells per cubic millimeter,17 43.6% would have been eligible to begin ART based on the first CD4 after the initial diagnosis; applying the January 10, 2011, threshold of CD4 <500 cells per cubic millimeter,11 60.8% would have been eligible to begin ART.
Achievement of Viral Suppression
Overall, 7663 of 12,122 patients (63.2%) ever achieved viral suppression by the end of the analysis period; 26.6% achieved suppression within 6 months of diagnosis, and 39.8% within 12 months (Table 2). More than half (50.9%) of patients with an initial CD4 <200 cells per cubic millimeter achieved suppression within 6 months of diagnosis; in comparison, only 9.2% of patients with an initial CD4 count ≥500 cells per cubic millimeter achieved suppression within 6 months. The proportion of patients achieving viral suppression within 12 months was 36.3% among patients diagnosed in 2006; it increased to 45.4% among patients diagnosed in 2009.
Figure 1 shows that the proportion achieving viral suppression after diagnosis improved over the 4 years (P < 0.0001) and that patients with a lower initial CD4 count achieved suppression more rapidly (P < 0.0001). Stratified by initial CD4 count, annual improvement in time to suppression was seen in all groups except those with an initial CD4 count <200 cells per cubic millimeter (P = 0.15).
On Cox proportional hazards regression, achievement of suppression was significantly associated with female sex, white vs. black or Hispanic race/ethnicity, increasing age, MSM risk, more recent year of initial diagnosis, and low initial CD4 count (Table 3). After controlling for other factors, persons with initial CD4 count between 200 and 349 cells per cubic millimeter were 35% less likely, persons with CD4 count from 350 to 499 cells per cubic millimeter were 62% less likely, and those with CD4 count ≥500 cells per cubic millimeter were 74% less likely than those with CD4 count <200 cells per cubic millimeter to achieve suppression. Persons diagnosed in 2009 were 49% more likely than those diagnosed in 2006 to achieve suppression by the end of the analysis period (P < 0.0001).
Maintenance of Suppression
Overall, 89.2% of the 7663 persons who ever achieved viral suppression maintained their suppression for 6 months and 81.9% for 12 months (Table 2). The proportion maintaining suppression at 12 months was 79.3% among patients diagnosed in 2006% and 86.4% among patients diagnosed in 2009.
Figure 2 shows the improvement in maintenance of suppression with each successive year of diagnosis. It also shows that those least likely to maintain suppression were those with highest nadir CD4 counts (≥500 cells/mm3) before achieving suppression, that is, the most immunocompetent in our population. When stratified by nadir CD4 before suppression, only the group with CD4 < 200 cells per cubic millimeter showed improvement in maintenance of suppression over the 4 years.
On Cox proportional hazards analysis, maintenance of suppression was significantly associated with white vs. black or Hispanic race/ethnicity, increasing age, MSM risk, and successive year of diagnosis (Table 3). For example, persons older than 60 at the time of diagnosis were 79%, and persons aged 45–49 were 73% more likely than persons aged 18–24 to maintain suppression. Persons diagnosed in 2009 were 52% more likely to maintain suppression overall than those diagnosed in 2006. An interaction between sex and nadir CD4 count was found to be statistically significant (P = 0.03), and HRs by nadir CD4 count are reported separately for men and women. Those with the highest nadir CD4 counts before suppression (≥500 cells/mm3) were least likely to maintain it (men: HR = 0.72; 95% confidence interval, 0.61 to 0.82 and women: HR = 0.67; 95% confidence interval, 0.53 to 0.87). Women with CD4 counts <350 cells per cubic millimeter and men with CD4 200–499 cells per cubic millimeter did best.
Three important observations emerge from this analysis. The first is that persons belonging to successively later diagnostic cohorts consistently improved in proportion suppressed, time to suppression, and the duration of suppression. The finding reinforces data showing that belonging to a more recent diagnostic cohort provides an advantage in entry to care18 and suggests that this advantage might extend to initiation and outcome of ART, even during the 4 years during which the CD4 <350 threshold was in effect. Ecologic changes that may have influenced these outcomes included the (1) growing number of classes, gene targets, and drugs in the AIDS formulary, (2) better bioavailability, (3) simpler dosing schedules, (4) increasing physician experience, and (5) changes in ART guidelines and prescribing practices, as well as scale-up of testing and new legislation requiring proactive linkage to care.19,20
The second is that persons with more advanced disease were significantly more likely to achieve viral suppression. The finding suggests that during the analysis period, ART was still being deferred until Department of Health and Human Services thresholds (<200 cells/mm3 in 2006, <350 cells/mm3 in 2007–2009)17,21 were reached. It also suggests the possibility that the achievement of suppression may be tied to the greater likelihood of symptomatic or clinical disease that accompanies advanced immunosuppression, a phenomenon that may motivate the physician to prescribe ART and the patient to comply.22–24
The third is that persons with the highest nadir CD4 count (≥500 cells/mm3) were least likely to maintain viral suppression. The absence of clinical disease may also play a role in the higher risk of failure experienced by persons with CD4 count ≥500 cells per cubic millimeter who initially achieved suppression because it may compromise adherence. Setting realistic goals for test and treat also requires that we estimate the degree of viral suppression that would be sufficient to achieve widespread disease control in individuals and reduced incidence in populations. The ongoing transmission of HIV in NYC (3700 new diagnoses in 2010) and the existence of a large pool of HIV-infected persons among whom more than 55% have detectable VL, and more than 36% lack laboratory evidence of care in the past year are sober reminders that the city faces many challenges in providing ongoing care and controlling VL.25,26 Achieving 70% suppression among newly diagnosed persons is only a beginning and probably represents a best case scenario—78% of the newly diagnosed initiate care within 3 months of diagnosis and 85% within the first year; 91% of those initiating care are retained for at least 1 subsequent year.27 Engagement in care and adherence to ART over the long term is a much more formidable challenge.28–30 However, our analysis suggests what is possible with new cohorts of infected persons who initiate care and start treatment earlier each year and take advantage of simpler, less toxic, and more effective ART. If the majority is able to achieve and maintain suppression without special assistance beyond their routine primary care visits, the city could consider offering directly observed or assisted therapy, case management, and other support to the smaller, more manageable number of persons who cannot.
Our analysis has the advantage of large numbers representing a heterogeneous epidemic. However, it has significant limitations. Although electronic laboratory reporting is mandatory and considered to be complete, not all laboratory results may have been reported, resulting in underascertainment of viral suppression and failure, and not all reports may be matched to the correct case record. For consistency, we retained <400 copies per milliliter as the undetectable threshold, despite the many improvements in sensitivity of the VL assays and the changes in the lower limits of detection that occurred during the period. We allowed patients an isolated “blip” ≤1000 copies, as permitted in other test-and-treat analyses that found no transmission at low detectable VL.1,4 We classified all invalid and missing VLs as detectable, and we assumed that all patients with no VL reports (ie, not in care and therefore not on ART) would be viremic. Patients who had 1 suppressed VL and no further data were presumed to remain suppressed. Risk factor was missing in 29% of patients; CD4 count at diagnosis was missing in 17% and resulted in an inability to classify these patients by CD4 threshold.
Surveillance does not collect data on ART, and therefore, this analysis cannot estimate ART coverage or efficacy. It reflects only the proportion of the population that was diagnosed and reported and achieved or maintained viral suppression during the analysis period. The absence of data on ART is the most significant limitation associated with the use of surveillance data. It stems at least in part from the fact that the primary and sometimes sole field investigation involving medical record review takes place immediately after the diagnosis, before ART has begun, and sometimes even before the patient has been linked to care. All subsequent data received on a case are obtained via the electronic transmission of reportable laboratory events (CD4, VL, and genotype). However, if viral suppression is viewed as an indicator of the combined effects of provider practice, patient compliance, and natural nonprogression in a population and not only as a treatment outcome, it is a useful measure of that population's health and the prevention challenges that face it.
Similar thinking underlies the proposal to use viral suppression as a population outcome indicator in the nation's strategy to control HIV/AIDS.26 Because the sole population-based data source for HIV/AIDS is surveillance, the Centers of Disease Control and Prevention and the states have been charged with making the enhancements to the national and local systems that will allow them to monitor this and other outcomes as promptly and accurately as possible. The new guidelines will compensate to some extent for the absence of data on ART uptake because all diagnosed persons should be started on ART as soon as they initiate care and complete their baseline evaluations.17,20,21
Finally, this analysis included new diagnoses only. New diagnoses represent a small (∼4%) fraction of the total prevalence pool, whose numbers now exceed 111,000 and whose narrative contains events and characteristics that can compromise virologic control, for example, diagnoses that are many years' old, complicated treatment and resistance histories, aging, and comorbidities. Moreover, we do not include undiagnosed individuals (currently estimated at 14% of NYC's prevalence pool),31 the overwhelming majority of whom will be viremic and thus an ongoing source of HIV transmission. We justified restricting this initial effort to newly diagnosed persons in view of the many changes in ART and prescribing practices and attitudes that have occurred since the first generation of highly active ART. Data continue to accumulate and influence the recommendations regarding optimum time to initiate ART, acceptable levels of RNA suppression, strategies for early diagnosis and intervention in pregnancy and acute HIV infection, and strategies for retention in care and achievement and maintenance of suppression. By design, we began with a clean slate of diagnoses lacking historical burdens so that we could gauge what might be possible with a new generation of people with HIV/AIDS.
More than 70% of persons newly diagnosed with HIV in between 2006 and 2009 achieved virologic suppression, 40% within 1 year of diagnosis. The majority achieving suppression maintained it for at least 8.4 months over a follow-up period ranging from 1 month to 4.7 years. Our results represent a broad population that is heterogeneous with respect to race and risk factor, which includes both treated and untreated individuals and persons receiving adherence support ranging from extremes of no service to clinical trial follow-up and/or incentives to remain on ART. Therefore, the observations establish a baseline that represents what the minimum achievable population outcome might be if a test-and-treat model were implemented with newly diagnosed persons in NYC. The achievements diverge from the ideals promulgated by models that project eventual elimination of the epidemic if 100% virologic control were achieved. However, they show us both what is already possible and what might be achieved with the diagnostic cohorts of the future, for whom treatment will be initiated immediately after the diagnosis.
Establishing baselines and targets for local test-and-treat strategies will require future analyses incorporating a variety of factors not analyzed here, including (1) adjustment for the potential effect of changes in the percent diagnosed and timing of diagnosis (including improved ascertainment and early treatment of acute HIV) resulting from the expansion of HIV testing and linkage to care, (2) ongoing changes in the guidelines for the initiation of ART, (3) changes in the efficacy, availability, and administration of antiretroviral agents, (4) changes in the virulence of HIV, (5) mediators of transmission other than plasma VL, (6) deployment of preexposure prophylaxis to high-risk populations, (7) behavioral disinhibition secondary to preexposure prophylaxis and ART, and (8) modeling of the effect on incidence of different levels of viral suppression in the population.
1. Cohen MS, Chen YQ, McCauley M, et al.. Prevention of HIV-1 infection with early antiretroviral therapy. N Engl J Med. 2011;365:493–505.
2. Donnell D, Baeten JM, Kiarie J, et al.. Heterosexual HIV-1 transmission after initiation of antiretroviral therapy: a prospective cohort analysis. Lancet. 2010;375:2092–2098.
3. Attia S, Egger M, Muller M, et al.. Sexual transmission of HIV according to viral load and antiretroviral therapy: systematic review and meta-analysis. AIDS. 2009;23:1397–1404.
4. 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.
5. Fideli US, Allen SA, Musonda R, et al.. Virologic and immunologic determinants of heterosexual transmission of human immunodeficiency virus type 1 in Africa. AIDS Res Hum Retroviruses. 2001;17:901–910.
6. Das M, Chu PL, Santos GM, et al.. Decreases in community viral load are accompanied by reductions in new HIV infections in San Francisco. PloS One. 2010;5:e11068.
7. 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.
8. Dodd PJ, Garnett GP, Hallett TB. Examining the promise of HIV elimination by ‘test and treat' in hyperendemic settings. AIDS. 2010;24:729–735.
9. Eaton JW, Johnson LF, Salomon JA, et al.. HIV treatment as prevention: systematic comparison of mathematical models of the potential impact of antiretroviral therapy on HIV incidence in South Africa. PLoS Med. 2012;9:e1001245.
10. Wagner BG, Blower S. Universal access to HIV treatment versus universal ‘test and treat': transmission, drug resistance & treatment costs. PloS One. 2012;7:e41212.
11. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Washington, DC: Department of Health and Human Services; 2011:1–161.
12. Jia Z, Ruan Y, Li Q, et al.. Antiretroviral therapy to prevent HIV transmission in serodiscordant couples in China (2003-11): a national observational cohort study. Lancet 2012 [epub ahead of print; doi: 10.1016/S0140-6736(12)61898-4].
13. Folkers GK, Fauci AS. Controlling and ultimately ending the HIV/AIDS pandemic: a feasible goal. JAMA. 2010;304:350–351.
14. Fairchild AL, Bayer R. HIV surveillance, public health, and clinical medicine—will the walls come tumbling down? N Engl J Med. 2011;365:685–687.
15. Rodriguez B, Sethi AK, Cheruvu VK, et al.. Predictive value of plasma HIV RNA level on rate of CD4 T-cell decline in untreated HIV infection. JAMA. 2006;296:1498–1506.
16. Do T, Duncan J, Butcher A, et al.. Comparative frequencies of HIV low-level viremia between real-time viral load assays at clinically relevant thresholds. J Clin Virol. 2011;52(suppl 1):S83–S89.
17. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Washington, DC: Department of Health and Human Services; 2007.
18. Torian LV, Wiewel EW, Liu KL, et al.. Risk factors for delayed initiation of medical care after diagnosis of human immunodeficiency virus. Arch Intern Med. 2008;168:1181–1187.
19. State of New York Laws. HIV Testing and Counseling. Amendment to New York State Public Health Law Article 21, Amendment of Part 63 of Title 10, Codes, Rules and Regulations of the State of New York (HIV/AIDS Testing, Reporting and Confidentiality of HIV-Related Information). Chapter 308. Albany, NY: State of New York; 2010.
20. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Washington, DC: Department of Health and Human Services; 2012.
21. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Washington, DC: Department of Health and Human Services; 2006.
22. Giordano TP, Gifford AL, White AC Jr, et al.. Retention in care: a challenge to survival with HIV infection. Clin Infect Dis. 2007;44:1493–1499.
23. Fleishman JA, Gebo KA, Reilly ED, et al.. Hospital and outpatient health services utilization among HIV-infected adults in care 2000-2002. Med Care. 2005;43(suppl 9):III40–52.
24. Meyerson BE, Klinkenberg WD, Perkins DR, et al.. Who's in and who's out: use of primary medical care among people living with HIV. Am J Public Health. 2007;97:744–749.
26. The White House Office of National AIDS Policy. National HIV/AIDS Strategy for the United States. Washington, DC, 2010.
27. 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.
28. 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.
29. Andrews JR, Wood R, Bekker LG, et al.. Projecting the benefits of antiretroviral therapy for HIV prevention: the impact of population mobility and linkage to care. J Infect Dis. 2012;206:543–551.
30. Burns DN, Dieffenbach CW, Vermund SH. Rethinking prevention of HIV type 1 infection. Clin Infect Dis. 2010;51:725–731.
31. Eavey JJ, Torian LV, Jablonsky A, et al.. Undiagnosed HIV infection in a New York City emergency room. Presented at: XIX International AIDS Conference; July 22-27, 2012; Washington, DC. Abstract #TUPE282.