Use of undetectable viral load to improve population-based survey estimates of known HIV-positive status and antiretroviral treatment coverage in Kenya

Introduction

The HIV cascade of care, can be described by how many HIV-infected persons know their status, how many who know their status are on antiretroviral therapy (ART), and how many on ART are virally suppressed. These indicators are used to monitor progress towards controlling the HIV epidemic. Since the 2007 Kenya AIDS Indicator Survey (KAIS), HIV seroprevalence surveys have often included questions on knowledge of HIV status and antiretroviral (ARV) use among HIV-infected respondents, as well as treatment-related biomarkers such as viral load (VL) [1–3]. However, HIV continues to be a highly stigmatized disease in spite of rapid scale up of life-saving treatment [4,5], therefore self-reported knowledge of status and ART can be subject to either positive or negative social desirability bias in some respondents. Some respondents may also have inaccurate understanding and recall of detailed questions about their HIV testing and care histories [6].

Recent surveys have used ARV testing to improve interpretation of self-reported HIV cascade indicators by reclassifying respondents with ARVs detected in their blood as being previously diagnosed and on ART [7,8]. In the 2012 KAIS 46.9% [95% confidence interval (CI) 41.3–52.4] of HIV-infected respondents self-reported that they were aware of their HIV-positive status, but ARVs were also detected in 21.0% of those not reporting prior HIV diagnosis and 19.3% of those reporting no previous HIV test. Younger age and higher household wealth index were independently associated with undisclosed ART [7]. After reclassifying those who had detectable ARVs in their blood knowledge of HIV-positive status rose to 56.2% (95% CI 50.7–61.7). Self-reported population ART coverage was 31.8% (95% CI 27.1–36.5); after adjusting for ARV detection it rose to 42.8% (95% CI 37.9– 47.8) [7]. These increased estimates of both knowledge of HIV-positive status and ART coverage underscore the potential utility of biomarker adjustments on self-reported indicators in population-based surveys.

Antiretroviral testing is both highly sensitive and specific for detection of ARVs in blood [9,10], thus false-positive results are unlikely and anyone testing positive for both HIV and ARVs is likely to be on treatment and previously diagnosed with HIV infection. A negative ARV test in an HIV-positive respondent however may reflect either ARV non-use, poor recent adherence, or non-correspondence between drugs in test panel and those in use in the population. Thus, surveys that include ARV testing usually define ART use as either self-reported use of ART or detection of ARVs in the blood. Antiretroviral testing is also relatively complex, expensive both in terms of up-front investment and ongoing operational costs, and not widely available within national public health laboratories in high-HIV-burden countries.

Viral load monitoring has become a routine part of clinical HIV care given its utility in identifying treatment failure, defined as having a persistent viral load ≥ 1000 copies/mL among patients on treatment [11]. UNAIDS defines viral suppression for surveillance and reporting purposes as VL < 1000 copies/mL [12]. Viral load testing platforms have been implemented widely in high-HIV-burden countries at reasonable cost (9.40 US Dollars per test [13]). The Abbott M2000 Real-Time HIV-1 Assay (Abbott Laboratories, Abbott Park, Illinois, USA) platform used in the KAIS 2012 survey has sensitivity 93.9% (95% CI 88.8–97.2) and specificity of 88.0% (95% CI 82.2– 92.4) for viral detection at a cutoff of 550 copies/mL in dried blood spots (DBS) [14]. While undetectable viral load (UVL) is generally indicative of viral suppression due to treatment, the presence of elite controllers (EC) who have UVL in the absence of treatment could confound its use as a proxy for ART use. The prevalence of EC in the population is not well characterized: in US and European cohorts they are believed to represent 0.15– 1.5% of the HIV-infected population [15], while in East African settings similarly-low prevalence of EC has been observed [16,17], though in one cross-sectional population-based assessment in Uganda 10% of participants with undetectable viral load were not on ART according to self-report, ARV testing and chart review, raising the possibility of higher prevalence of elite control in this population [18].

Although both ARV detection and UVL have the potential to correct for under-reporting of known HIV status and ART use among HIV-positive survey participants, the converse is not true. Survey respondents on ART can be non-adherent to their medication leading to undetectable ARV blood levels. Similarly, patients on treatment may still have detectable viral load, either due to suboptimal drug concentrations, recent treatment initiation, or treatment failure due to drug resistance.

A better understanding of the relationship between UVL and ARV detection is required to assess the utility of VL testing as an alternate or adjunct biomarker for use in survey adjustment. In order to explore the impact of alternative approaches to adjusting HIV diagnosis and ART coverage estimates, we looked at whether adjusting the HIV cascade using a measure of viral load would achieve similar results to adjustments based on detection of ARVs, as well as the impact of simultaneously adjusting based on both biomarkers.

Methods

The 2012 KAIS was a national household survey that included behavioral questions including self-reported HIV and ART status as well as collection of venous blood from which DBS were prepared by field teams and plasma separated and shipped for HIV testing at a national laboratory [2]. After participating in other survey procedures, participants were offered rapid HIV testing by trained HIV counselors in their homes with immediate return of results based on national HIV testing guidelines [19]. Participants testing positive for HIV at the central lab were subsequently tested for viral load using the Abbott M2000 platform on DBS, and DBS were also shipped to the University of Cape Town for ARV testing. Antiretrovirals were considered present if one or more of efavirenz, nevirapine, lopinavir or lamivudine were detected by liquid chromatography tandem mass spectrometry (limit of detection 0.02 µg/mL); these drugs were selected to cover first- and second-line regimens in use in Kenya at the time of the survey [7,20,21].

We retrospectively re-analyzed survey data to compare self-reported and biomarker-adjusted versions of knowledge of status and ART use among HIV-infected respondents aged 15-64 years.

Measures

Self-reported HIV status among respondents testing HIV-positive in the survey was defined as ‘unaware’ of their HIV-positive status, ‘aware, on ART’, or ‘aware, not on ART’. Participants with confirmed HIV-positive status who self-reported being HIV-positive during the interview were categorized as ‘aware’ of their HIV-positive status; all other HIV-positive respondents, including those that refused to answer, were categorized as ‘unaware’ of their HIV-positive status. HIV-positive respondents categorized as ‘aware’ who further responded yes to both the question ‘have you ever taken ARVs, that is, antiretroviral medication, to treat your HIV infection’ and ‘are you currently taking ARVs, that is, antiretroviral medications daily?’ were re-categorized as ‘aware, on ART’, otherwise as ‘aware, not on ART’.

We defined UVL as having a viral load <550 copies/mL on dried blood spots, the limit of detection for the assay used in the study. To calculate UVL-adjusted status, we updated the status for those respondents categorized as ‘unaware’ or ‘aware, not on ART’ with undetectable viral load to ‘aware, on ART’. Similarly, ARV-adjusted status was calculated by updating the status for respondents with ARVs detected in blood to ‘aware, on ART’. For either case, the status for respondents with missing biomarker results were not updated.

We tabulated versions of population ART coverage based on 1) self-report alone, 2) ARV detection alone, 3) self-report adjusted for ARVs, 4) self-report adjusted for UVL and 5) self-report adjusted for both ARVs and UVL. We explored differences in these indicators by age, sex, marital status, educational attainment and mobility. We also tabulated self-reported, ARV- and UVL-adjusted versions of known HIV-positive status, ART coverage among those who knew their HIV-positive status and viral suppression (defined as having <1000 copies/mL in blood) among those on ART, the indicators for which the Joint United Nations Programme on HIV/AIDS (UNAIDS) has set Fast-track targets to ensure 90% of people living with HIV (PLHIV) know their HIV status, 90% of those who know their status are on treatment, and 90% of those on treatment are virally suppressed by 2020 [22]. ARV-only estimates are not presented for the cascade as this adjustment is only meaningful for measures of ART coverage. We calculated measures of accuracy for UVL and UVL & ARV-adjusted versions of knowledge of status and ART use versus ARV-adjusted as a reference standard. Results were analyzed in R version 3.5.0 [23] using the survey package [24]. Resulting indicators were adjusted and weighted to account for the complex survey design, except for measures of accuracy. Wald confidence intervals for survey indicators were calculated on the logarithmic scale and transformed to probability scale using the ‘logit’ method of the svyciprop function in R; confidence intervals previously reported by Kim et al were calculated on the probability scale). Wilson score confidence intervals are reported for measures of accuracy.

Ethical approval

The 2012 KAIS was approved by the Committee on Human Research of the University of California, San Francisco, the Institutional Review Board of the U.S. Centers for Disease Control (CDC) in Atlanta, GA, USA (protocol #6189) and the Ethical Review Committee of the Kenya Medical Research Institute.

Results

In KAIS 2012, among 648 HIV-infected respondents, self-reported status was ‘unaware’ among 343 (53.1%), ‘aware, not on ART’ among 100 (15.1%), and ‘aware, on ART’ among 205 (31.8%) (Table 1). Status varied by age and sex: 120 (62.0%) men versus 223 (47.8%) women were unaware of their HIV-positive status by self-report, and 59 (82.0%) 15–24 year olds versus 121 (58.9%) 25–34 year olds were unaware of their HIV-positive status by self-report. Only 6 (7.1%) of 15–24 year olds were aware and on ART compared with 47 (48.0%) of 50–64 year olds. Among those with undetectable viral load, 65 (28.7%) were unaware of their HIV-positive status by self-report, compared with 55 (25.2%) of those with ARVs detected (Table 1). Of those with UVL and unaware of their HIV-positive status by self-report, 40 also had ARVs detected in blood (Supplemental Table S1).

Antiretroviral treatment coverage among all HIV-infected increased from 31.8% (95% CI 27.3–36.6) based on self-report to 42.5% (95% CI 37.4–47.8) based on ARV detection alone, to 42.8% (95% CI 37.9–47.8) when combining self-report and ARV detection, to 46.2% (95% CI 41.3–51.1) when combining self-report and UVL, and finally to 48.8% (95% CI 43.9–53.8) with self-report, UVL and ARVs combined (Table 2). Changes in ART coverage were fairly consistent across demographic characteristics, typically increasing when measured by ARV only versus self-report only, and increasing further when combining both indicators. The 15–24 year age group saw an increase from 7.1% when self-reported only to 21.6% based on ARVs only and 20.7% when based on self-report and ARVs combined, and up to 29.9% when combining both indicators. The single or never married group also saw larger increases in ART coverage when adjusting. The 25–34 year age group had smaller increases due to adjustments than other groups, as did the no education category (Table SX).

When reviewing the cascade of HIV care based on the UNAIDS Fast-Track 90-90-90 indicators, knowledge of status increased from 46.9% (95% CI 41.4–52.4) based on self-report to 56.2% (95% CI 50.7–61.6) when adjusting with ARVs, to 57.5% (95% CI 51.9–63.0) when adjusting for UVL, and to 59.8% (95% CI 54.2–65.1) when adjusting for both ARV and UVL (Table 3 & Figure 1). Antiretroviral treatment among those previously-diagnosed (the second 90 cascade indicator) increased from 67.9% (95% CI 61.7–73.8) when the numerator and denominator were based on self-report to 76.2% (95% CI 70.3–81.2) when numerator and denominator were ARV-adjusted, to 80.2% (95% CI 75.2–84.4) when UVL-adjusted, and finally to 81.7% (95% CI 76.8–85.8) when adjusted for both ARV and UVL. Viral suppression among those on ART (the third 90 cascade indicator) was 75.4% (95% CI 68.4–81.2) when the denominator was based on self-report alone, 75.6% (95% CI 69.5–80.8) when based on self-report and ARV detection, 83.1 (95% CI 78.2–87.1) when adjusted for UVL, and to 78.6% (95% CI 73.4–83.1) wen adjusting for both ARV and UVL. The youngest age group also saw the biggest impact of adjustment versus self-report in the cascade indicators in both relative and absolute terms.

• Download figure
• Open in new tab

Figure 1. HIV cascade and ART coverage indicators, KAIS 2012

Notes: self-reported: self-report only, self or ARV: either self-reported known-positive/on ART or ARVs detected, self or UVL: either self-reported known-positive/on ART or viral load was undetectable, self-report, ARV or UVL: either self-reported known-positive/on ART, ARVs detected or UVL, 1^st 90: UNAIDS 90-90-90 Fast-Track target of 90% of people living with HIV knowing their HIV-positive status, 2^nd 90: 90% of those who know their status on ART, and 3^rd 90: 90% of those on treatment with viral load < 1000 copies/mL.

When using the ARV-adjusted indicator as a reference standard for knowledge of HIV-positive status, the sensitivity of UVL-adjusted self-reported knowledge of HIV infection was 95.8% (95% CI 93.7–97.9), and specificity was 91.3% (95% CI 88.1–94.6). Positive predictive value was 93.2% while negative predictive value was 94.6% (Table 4). When using the ARV-adjusted indicator as a reference standard for ART coverage, sensitivity of UVL-adjusted self-report was 93.0% (95% CI 89.9–96.0) and specificity was 88.8% (95% CI 85.7– 92.0). Positive predictive value was 85.8% and negative predictive value was 94.6% (Table 5).

We repeated the analysis excluding the respondents for whom either the ARV or UVL biomarkers were not available from all adjustments, findings were generally similar (Supplemental Table S2). However, the differences between ARV adjustments and UVL adjustments were less pronounced. The 15–24 year age group had a disproportionate number of missing ARV testing results (16 out of 75, which was about double the proportion seen in the remaining age groups) which may have contributed to a lower adjustment based on ARVs than on UVL in this age group.

Discussion

In KAIS 2012, UVL-adjusted point estimates were quite similar to, but slightly greater than ARV-adjusted estimates of knowledge of status and ART coverage. This suggests that adjustment with UVL might have been adequate and would be preferable to making no adjustment at all, as the UVL-adjusted indicators are closer to the ARV-adjusted indicators than they are to self-reported status alone. When measuring ART coverage, all of the adjusted estimates (ARV only, UVL only, and ARV/UVL combined) had overlapping confidence intervals, but appear to be significantly higher than estimates based on self-report alone. The change in estimates when adjusting by ARVs and UVL were similar across demographic groups, but 15–24 year olds did see a larger additional increase when adjusting by UVL. This may indicate poor recent adherence in this group leading to non-detection of the ARV biomarker but undetectable viral load (<550 copies/mL in this study). Li et al found that 37% of patients still had a viral load <200 copies/mL four weeks after interrupting ART, with a lower proportion still suppressed among patients on protease inhibitor-based regimens compared with non-nucleoside reverse transcriptase inhibitor-based regimens [25]. Many ARVs reach undetectable levels in blood within several days of treatment interruption, with the exception of efavirenz which can take 12–28 days to reach undetectable levels [7,9,26], thus in populations with poor adherence or high rates of treatment interruption, adjusting based on UVL may result in higher estimated ART coverage than measures incorporating ARV detection. Never married respondents also had larger increases in ART coverage when adjusted, however this was likely due to confounding with age, as never-married persons tend to be younger.

In this analysis we treated combined self-report or ARV detection as a reference standard as it appears to be the current preferred approach for adjusting knowledge of status and ART coverage in population-based surveys [27]. Compared with this standard, combined self-report or UVL performed well with a sensitivity of 95.8% and specificity of 91.3% for knowledge of status, and sensitivity of 93.0% and specificity of 88.8% for ART coverage in this study. The performance of UVL for adjusting ART use will depend on the prevalence of UVL in the population on HIV treatment. In populations with effective ART programs with high rates of viral suppression in the treated population, it may be a relatively sensitive marker for ART use; however, in populations with poor treatment outcomes a larger proportion of patients on treatment would not have UVL.

The prevalence of elite controllers has not been established in Kenya, hence it is not possible to quantify their influence on the UVL-adjusted estimates, but given the similarity between UVL-adjusted and ARV-adjusted estimates, their practical impact was limited in this survey. Although it is possible that estimates adjusted with both the UVL and ARV biomarkers simultaneously may in fact be closest to true population prevalence of the indicators of interest, without better data on prevalence of elite controllers in this population it seems more conservative to use one or the other markers, but not both. In settings with ample evidence of low prevalence of elite control, or where population high ART coverage and immediate treatment initiation means even elite controllers are likely to be on treatment, using the combined indicator would likely represent the most sensitive approach to estimating population-based knowledge of status and ART coverage.

This analysis was subject to several limitations. While adjusting for biomarkers associated with ARV exposure from a single time-point can account for misreporting of status among those on ART, it cannot account for those who misreport their knowledge of HIV-positive status but are not currently on treatment, or those who may be on treatment but transiently non-adherent to medications. This analysis was based on data from a single country with low ART coverage (43.5%, 95% CI 38.5–48.4) and viral suppression among those on treatment (73.9%, 95% CI 67.8–80.0) at the time of the survey compared with current program coverage. The UVL adjustment may perform differently in other populations. Simulation or replication of this analysis in a diverse set of populations could help elucidate the performance of UVL adjustment in different settings. Finally, poor specimen quality could have resulted in false-negative results for both the viral load and ARV biomarkers considered here. In spite of these limitations, this analysis does strongly suggest that use of UVL to adjust self-reported HIV status and ART use should be considered, especially in surveys where the inclusion of the ARV biomarker may be cost-prohibitive or subject to delays.

Conclusion

Undetectable viral load may be a useful biomarker for adjusting self-reported indicators of HIV diagnosis and treatment in cross-sectional surveys in absence of, or in addition to, adjustment based on detected ARVs in blood. Since viral load testing is routinely available within countries which are conducting HIV surveys and is relatively inexpensive due to its widespread use, it is a widely accessible measure which may be suitable for use in population-based surveys if ARV testing is unavailable or cost-prohibitive.

Competing interests

All authors declare no competing interests.

Authors’ contributions

PWY, EZG and KDC conceived the study. PWY conducted the analyses. All authors contributed to drafting and critical review of the manuscript.

Additional files

Additional file 1. Supplementary tables and figures

Word document containing additional supplementary analyses referenced in text.