Durable Viral Suppression among Persons with HIV in the Deep South: An Observational Study

Abstract

This study assessed predictors of stable HIV viral suppression in a racially diverse sample of persons living with HIV (PWH) in the southern US. 700 PWH were recruited from one of four HIV clinics in Metro Atlanta, GA. Data were collected from September 2012 to July 2017, and HIV viral loads were retrieved from EMR for 18 months. The baseline visits and EMR data were used for current analyses. Durable viral suppression was categorized as: 1. Remain suppressed, 2. Remain unsuppressed, and 3. Unstable suppression. The number of antiretroviral medications and age were significantly associated with durable viral suppression. Older age, fewer ART medications and availability of social support were positively associated with durable viral suppression over the 18-month observation period. Findings suggest that regimen complexity is potentially a better predictor of viral suppression than self-reported medication adherence. The need for consensus on the definition of durable viral suppression is also urged.

Introduction

Suppression of peripheral HIV viral load is essential to reduce the spread of HIV infection and to optimize the health of persons currently living with this infection (Cohen et al., 2016; Samji et al., 2013). If taken as prescribed, antiretroviral medications (ART), reduce and maintain HIV viral loads at undetectable levels. However, stable viral suppression that is durable over the long term is not achievable for all persons living with HIV/AIDS. Prior studies have indicated certain groups may be at risk for fluctuations in HIV viral suppression rates over time. One of the largest studies of HIV viral load stability was conducted by Marks et al., (2016). This study included over 10,000 patients with at least two viral load tests within a 12-month period and found that 65.9% of participants were virally suppressed at all observation points. Moreover, more patients exhibited improving viral suppression than worsening viral suppression over the 12-months of observation. The authors noted that a single viral load measure over-estimated durable viral suppression by 16%. Additionally, Marks et al., (2016) reported that those 40 years old and older, males, persons of Hispanic or “other” race/ethnicity, and individuals with a CD4 count < 200 cells/μL at entry into the study cohort were more likely to have stable viral suppression.

Few other studies have monitored viral stability and factors associated with fluctuating viral loads. Thakarar et al., (2016), among a sample of 138 persons engaged in “health care for the homeless” program, reported that incomplete viral suppression was more likely among those who were homeless. Among homeless persons, taking ART, there was 3.54 greater odds of incomplete viral suppression compared to housed individuals. A history of illicit drug use was associated with lower HIV viremia (likely due to engagement with drug treatment services); fluctuations in viral stability over time were not measured. These findings suggest, however, that this vulnerable group may be more likely to have unstable or incomplete viral suppression.

PWH face challenges that increase their vulnerability to suboptimal health outcomes. The region of the US with the largest increase in new HIV infections is the southern United States, where many systemic barriers are especially prevalent. Our work and other’s has shown that low health literacy is a risk for ART non-adherence and lower rates of attendance to scheduled HIV care visits (Anderson et al., 2020; Osborn et al., 2007; Waldrop-Valverde et al., 2010; Waldrop-Valverde et al., 2018). Other psychosocial indicators including social support (Waldrop-Valverde et al., 2014) and provider relationships (Valverde et al., 2018) have been shown to be associated with retention in HIV care. Internalized HIV stigma may also reduce important health behaviors linked to viral suppression (Valverde et al., 2018). Although several vulnerabilities common among persons with HIV have been associated with poor health behaviors, few studies to date have linked these risks with stable viral suppression.

The purpose of the present study, therefore, was to understand the association of demographic factors and psychosocial vulnerabilities to stable viral suppression over an 18-month period among a racially diverse sample of men and women with HIV in the southern US.

Methods

Data were collected from September 2012 to July 2017. PWH who were enrolled in one of four outpatient HIV clinics in the urban metro-Atlanta area were recruited. Study flyers, referrals from clinic providers/staff and word-of-mouth were the methods of recruitment. Those who tested positive for HIV-infection, prescribed antiretroviral therapy for a minimum of 6 months, had a HIV medical visit within the prior 9 months, and were fluent in English were eligible. Consent to participate and to verify information by medical records review were elicited verbally and written consent was obtained at the baseline visit. The study included face-to-face and computer-administered surveys at baseline and again at 6-months post-baseline. Baseline study visit data were used for all variables and HIV viral load values were additionally collected from electronic medical records (EMR) for 18 months post baseline. The study was approved by appropriate institutional human subjects oversight committees.

Measures

Demographics:

Basic demographic characteristics such as race, age, gender, and education were included. Health insurance was used as a proxy for socioeconomic status (SES) (Anderson et al., 2019). Insurance data were collected from the electronic medical record (EMR) that corresponded closest to the baseline interview date. Participants’ SES was categorized as “not low SES” if using private or commercial insurance or self-pay; “low SES” if receiving Ryan White (income eligibility for Ryan White is less than or equal to 400% of the federal poverty level); “very low SES” if receiving Medicare or Medicaid services (income eligibility requirement for Medicaid in Georgia is less than or equal to 133% of the federal poverty level) (Anderson, et al., 2019)

Internalized HIV Stigma.

Participants rated their agreement with statements reflecting internalized HIV stigmas on the Internalized HIV Stigma scale (Kalichman et al., 2009). This scale has 6-items describing ways persons with HIV (PWH) may internalize feelings of stigma related to HIV. Responses are rated “agree” or “disagree” and summed. In order to identify participants with high levels of internalized HIV stigma, the scale was dichotomized at the median score of 4 (< 4 = low stigma & ≥ 4 = high stigma)

Health Literacy.

Health literacy was measured using the HIV-Health Literacy (HIV-HL) measure (Ownby et al., 2013). This 19-item measure was developed to be administered on a touchscreen computer. The HIV-HL includes items assessing interpretation of prescription instructions, a video vignette of an encounter with a patient and provider, and other items assessing skills taking medications.

Social Support.

Social Support was measured using the Social Support Questionnaire (Zich & Temoshok, 1987). Four types of support that one may have received since becoming diagnosed with HIV are measured: whether support is “desirable”, “available” whether it has been “experienced” and if experienced, how “useful” it was. Eight items assess each of the 4 types of support and each item is rated from 1 “not at all” to 5 “very much, constantly”. The “available” social support subscale was used for this study.

Illicit Drug Use:

Participants responded “yes” or “no” to use of marijuana, cocaine, crystal meth, opiates or tranquilizers in the past six months. Responses were categorized as a single binary item coded “yes” if any of the substances were used in the past six months or “no” if they were not used in the past six months.

Trust in HIV Provider.

A 5-item scale was created to assess patient’s trust in their provider to “…offer quality care”, “…know the best treatment”, “provide enough information to me”, “keep personal information confidential”, and “…treat me in a non-judgmental way”. Items were rated on a 5-point Likert scale from 1=not at all true to 5= completely true. Higher scores reflect greater trust. Responses were summed and total scores used in the analyses. Cronbach’s alpha for this sample was 0.932.

Durable Viral Suppression.

The health outcome of interest was suppressed HIV viral load over time. Single HIV-1 RNA viral load values and clinical lab collection date were gathered from the EMR for each study participant for an 18-month observational period. The 18-month observation window was calculated from 6-months prior to the participant’s baseline visit to the 12-month post-baseline visit. Viral load was categorized as virologically suppressed or virologically non-suppressed using the definition of HIV-1 RNA less than 2.3 log₁₀, which corresponds to a viral load of 200 copies/mL. This definition of viral suppression is consistent with the Centers for Disease Control and Prevention (Centers for Disease Control and Prevention, 2018). Durable viral suppression was categorized into three groups: Remain Suppressed – HIV viral load < 200 copies/mL for the entire 18-months; Remain Unsuppressed – HIV viral load ≥ 200 copies/mL for the entire 18-months; Unstable Suppression – HIV viral load transitioned between < & ≥ 200 copies/mL during the 18 months.

Self-Report ART Adherence.

Self-reported adherence to ART medications in the previous four days was measured using the ACTG adherence questionnaire (Chesney et al., 2000). Percent of scheduled doses taken across all ARTs was calculated.

Statistical Methods

Summary statistics are provided for variables of interest including subjects’ relevant demographic and clinical characteristics, health literacy scores, and the outcomes of interest. Descriptive statistics were computed for the overall sample (Table 1) and compared between the three viral suppression groups (Table 2).

Table 1:

Demographic and clinical characteristics for study participants (N=700)

Variable	Level	N (%) = 700
VL Status Change	Remain Suppressed	339 (50.2)
	Remain Unsuppressed	148 (21.9)
	Unstable Suppression	188 (27.9)
	Missing	25
Age	Less than 50	351 (50.2)
	50 years or more	348 (49.8)
	Missing	1
SES	Very Low	421 (66.3)
	Low	69 (10.9)
	Not Low	145 (22.8)
	Missing	65
Gender	Cisgender Male	483 (69.1)
	Cisgender Female	197 (28.2)
	Transgender Female	16 (2.3)
	No response/Missing	4 (0.4)
Race (Black vs. non-Black)	Non-Black	276 (39.4)
	Black	424 (60.6)
Drug use past 6 months	No	447 (65.4)
	Yes	237 (34.6)
	Missing	16
Internalized AIDS Stigma	Less Stigmatized	491 (70.5)
	More Stigmatized	205 (29.5)
	Missing	4
Health Care Provider Distrust Total Score	Mean	11.26
	SD	17.96
	Missing	14.00
Total number of antiretroviral medications	Mean	2.33
	SD	1.06
	Missing	1.00
Health Literacy	Mean	16.70
	SD	3.55
	Missing	1.00
Self-reported ART adherence	Mean	0.93
	SD	0.18
	Missing	9.00
Available social support	Mean	29.72
	SD	7.38
	Missing	16.00

Table 2:

Univariate Association Analysis for evaluating association between viral suppression and the covariates. N = 675

			VL Suppression
Covariate	Level	Remain Suppressed N=339	Remain Unsuppressed N=148	Unstable Suppression N=188	Test Statistic (df)	Parametric P-value*
N (Col %) or Mean (SD)
Drug use past 6 months	No	230 (68.86)	85 (60.71)	117 (63.24)	χ2=3.51 (2)	0.173
	Yes	104 (31.14)	55 (39.29)	68 (36.76)
Age	Less than 50	161 (47.63)	94 (63.51)	85 (45.21)	χ2=13.24 (2)	0.001
	50 years or more	177 (52.37)	54 (36.49)	103 (54.79)
Internalized AIDS Stigma	Low Stigma	244 (72.4)	90 (61.22)	140 (74.87)	χ2=8.40 (2)	0.015
	High Stigma	93 (27.6)	57 (38.78)	47 (25.13)
SES	Very Low	201 (65.05)	92 (68.66)	116 (68.64)	χ2=1.17 (4)	0.883
	Low	35 (11.33)	13 (9.7)	19 (11.24)
	Not Low	73 (23.62)	29 (21.64)	34 (20.12)
Gender	Cisgender Male	235 (69.53)	107 (72.3)	125 (66.49)	---	0.388*
	Cisgender Female	94 (27.81)	34 (22.97)	60 (31.91)
	Transgender Female	8 (2.37)	5 (3.38)	3 (1.6)
	No response	1 (0.3)	2 (1.35)	0 (0)
Race	Non-Black	142 (41.89)	54 (36.49)	71 (37.77)	χ2=161 (2)	0.448
	Black	197 (58.11)	94 (63.51)	117 (62.23)
Health Care Providers Distrust total score.		10.88 (15.39)	11.92 (21.90)	11.54 (19.42)	F = 0.19 (2,671)	0.825
Total number of antiretroviral medications		2.19 (1.02)	2.56 (1.11)	2.39 (1.06)	F = 6.63 (2,671)	0.001
Health Literacy		16.85 (3.43)	16.49 (3.71)	16.45 (3.68)	F = 0.98 (2,671)	0.378
Self-Report ARV adherence		0.95 (0.16)	0.91 (0.21)	0.93 (0.17)	F = 1.93 (2,663)	0.146
Available social support		30.26 (7.38)	28.05 (7.64)	30.07 (7.05)	F = 4.97 (2,669)	0.007

^*Fisher’s Exact Test employed to assess association between gender and VL suppression.

Univariate association analyses were performed to evaluate the relationship between selected covariates of interest and the outcome. For continuous covariates, ANOVA was conducted to evaluate these associations. For each categorical covariate, either a Chi-square or Fisher’s Exact test was conducted to evaluate each association. For multivariate analysis, a generalized (i.e., multinomial) logistic regression model was fit to assess the association between chosen covariates and viral suppression status. Age, internalized stigma, number of ARTs, and social support availability had a p< 0.10 in univariate tests and were entered into the multivariate model. Multicollinearity was assessed among covariates by examining the logistic model design matrix and by examining associations among covariates. All analyses were conducted using SAS version 9.4 (Cary, NC) and statistical significance was determined at the α = 0.05 type-I error rate.

Results

A total of 700 participants enrolled in the study. Of those enrolled, EMR data were extracted for 687 participants, and 673 of those participants had EMR data for more than one time point. About half of the study participants (N=339, 50.2%) fell into the Remain Suppressed group, while approximately 22% (N=148) were in the Remain Unsuppressed group. Of the study participants, approximately 35% (N=237) reported drug use in the past 6 months, 70% (N=483) were cisgender men, half (N=348) were over the age of 50, and 60% (N=424) were Black. On average, participants were prescribed 2.33 (SD=1.06) ART medications. Scores for healthcare provider distrust, social support, and internalized stigma exhibited good reliability with Cronbach’s alpha of 0.93, 0.96, and 0.91, respectively.

Results from univariate tests for association (Table 2) showed statistically significant associations between four covariates and durable viral suppression: age, total number of ARTs, internalized AIDS stigma, and availability of social support. With respect to age, there was a significantly greater proportion of persons less than 50 years old in the Remain Unsuppressed group. That group also included a significantly larger proportion of persons who were prescribed a higher number of ARTs. The largest proportion of persons with high perceived internalized stigma were in the ‘Remain Unsuppressed’ group. In addition, the Remain Unsuppressed group contained the lowest mean scores on available social support.

Results of the multivariate analysis are shown in Table 3. Age, number of ARTs, social support, ART adherence, and the interaction between number of ARTs and ART adherence were each significantly associated with viral suppression while internalized stigma was not. Multicollinearity was not found among these covariates. Participants older than 50 years of age were more than twice as likely to remain suppressed rather than unsuppressed when compared to those 50 years or younger (OR = 2.19, p < 0.001). Availability of social support was positively associated with the odds of remaining suppressed; participants with a score one point higher than others were 0.97 times as likely to remain unsuppressed as opposed to suppressed (OR = 0.97, p = 0.009). For each 1-unit increase in the number ART medications, at the mean ART adherence rate of 93%, the odds of remaining unsuppressed as opposed to suppressed were about 1.4 times higher (OR = 1.43, p <0.05). No significant differences were found between the remain suppressed and unstable suppression groups.

Table 3:

Multivariate Association Analysis for evaluating association between viral suppression and the covariates. N=646

			VL Status Change=Remain Suppressed as the Reference
Covariate	Level	VL Status Change	Odds Ratio (95% CI)	OR P-value	Type3 P-value
Age (Years)	50 or older	Remain Unsuppressed	0.457 (0.299,0.697)	<.001	<.001
	50 or older	Unstable Suppression	1.033 (0.711,1.501)	0.865
Available social support		Remain Unsuppressed	0.963 (0.936,0.992)	0.011	0.028
		Unstable Suppression	0.999 (0.973,1.027)	0.970
Internalized AIDS Stigma		Remain Unsuppressed	0.978 (0.886,1.080)	0.662	0.736
		Unstable Suppression	1.022 (0.933,1.21)	0.638
Total number of ART medications	93% ART adherence	Remain Unsuppressed	1.427 (1.172,1.737)	0.071	0.048
	93% ART adherence	Unstable Suppression	1.203 (1.004,1.442)	0.321
Self-reported ART adherence	2 ART medications	Remain Unsuppressed	0.289 (0.081,1.029)	0.008	0.009
	2 ART medications	Unstable Suppression	1.097 (0.226, 5.319)	0.584
Total number of ART medications * ART adherence		Remain Unsuppressed	-	0.013	0.012
		Unstable Suppression	-	0.459

Note: Number of observations in the original data set = 700. Number of observations used = 646.

Discussion

Maintaining suppressed viral load is essential for the health of persons living with HIV and for reducing future transmission. This study, conducted in the southern United States, assessed predictors of durable viral suppression (< 200 copies/ml over an 18 month time period) in persons who may have a number of barriers to achieving this goal. Although participants in this study reported high ART adherence, only half demonstrated durable viral suppression. Our analyses show that, when controlling for other factors, the most prominent risk factor for failing to achieve viral suppression was number of ART medications. Age was the most prominent protective factor. The greatest differences were noted between the group of participants whose viral load remained suppressed and those who remained unsuppressed over the 18-month period. In addition, those with greater social support were more likely to maintain viral suppression over 18 months than persons who reported less support from friends and family. Psychosocial issues included in the study were not associated with durable viral suppression.

Research has routinely demonstrated that complex or multi-drug ART medication regimens reduce effective management of HIV and viral suppression (Altice et al., 2019; Stone et al., 1998) and has led to improved medication formulations and simplified regimens. However, the average number of medications taken by participants in this study was over two pills, indicating that some PWH continue to have complex regimens. Although, multiple medications are required for a variety of reasons, early detection and treatment of HIV infection play a significant role in reducing the number of medications needed for durable viral suppression and better outcomes for PWH (Pilcher et al., 2017; Zolopa, 2010). Durable viral suppression and diminished transmission is dependent on adherence to treatment with ART (Eisinger et al., 2019) and this study, and others (Chen et al., 2017), indicate that a ≥ 2 ART medications regimen could be a marker for poor adherence and non-durable viral suppression. Given limitations of self-report medication adherence, regimen complexity may be a better indicator of durable viral suppression than self-reported medication adherence since, in our study, average self-reported medication adherence was 93% at study outset, but only 50% remained suppressed throughout the study period. Innovations in HIV treatment such as long acting injectables may eventually help to alleviate the risks associated with multiple pill ART regimens.

As in the present study, prior research supports the relationship of older age and durable viral suppression (Marks et al., 2016; Tanner et al., 2016). Older age reduced non-adherence by a relative risk of 27% greater in a meta-analysis (Ghidei et al., 2013), supporting the notion that older adults living with HIV are more adherent than their younger counterparts. This is further supported among diverse groups of PWH and across several countries (Mizuno et al., 2017; O’Neil et al., 2012). In the present study, those < 50 years of age were more likely to experience viral rebound during the 18-month observation period. As has been identified previously, interventions targeting medication adherence likely need to emphasize strategies that are relevant to younger persons living with HIV.

Additionally, our findings indicate that social support may be an important factor in viral suppression. The broader research on the relationship of social support with viral suppression, however, is equivocal. Burgoyne (2005) showed that perceptions of available emotional, informational, and interpersonal support were associated with achieving undetectable viral load over time (Burgoyne, 2005). However, other studies have not supported a relationship among available social support and viral suppression (Kemp et al., 2019) or have found variable findings for men and women (Maragh-Bass et al., 2018; Maragh-Bass et al., 2021). An earlier study from our team demonstrated that social support was associated with increased retention in HIV care (Waldrop-Valverde et al., 2014). It may be that equivocal findings on HIV viral suppression and social support are due to additional care-related factors (e.g., retention or other access issues) that are not accounted for in some studies. Additional research is needed to better understand if social support may predictably improve rates of durable viral suppression and through what mechanisms.

Other variables that we examined (low socioeconomic status, internalized HIV stigma, health literacy, drug use, and provider relationships) were not associated with durable viral suppression. Earlier studies by this team and others (Anderson et al., 2019; Osborn et al., 2007; Waldrop-Valverde et al., 2010; 2018) have shown that health literacy is associated with reduced ART adherence, retention in HIV care, and a one-time measure of viral suppression (Anderson et al., 2020). Similarly, other research has shown SES (Ghiam et al., 2017), AIDS-related stigma (Christopoulos et al., 2019), provider relationships (Anderson et al., 2020) and drug use (Sheehan et al., 2020) to be associated with viral suppression using various definitions. Sample characteristics and use of durable viral suppression, rather than a one-point-in-time measure, may explain the differences in our findings compared to others.

Importantly, a definition of durable viral suppression has yet to be established by the scientific community (Diepstra et al., 2020) and varying operationalizations of the construct likely contribute to varying findings across studies. The more viral load tests a patient receives within a given time period, the greater the likelihood a patient’s status as “durably suppressed” may change (Diepstra et al., 2020). As our knowledge of the relationship between sustained HIV viral suppression, HIV transmissibility, and optimum health for PWH expands, scientific consensus on the definition of durable viral suppression will be key to understanding its full impact and consistent barriers and contributors to durable viral suppression.

Limitations

Although the study sample was purposely recruited from multiple HIV clinics in a metro-area in the southern US where the epidemic is focused, the sample is not representative of many patients in the US south, including those from rural areas. Factors associated with durable viral suppression in rural areas may be unique from those reported herein. As noted above, no consensus on the definition of durable viral suppression exists and comparison across studies should account for variable definitions. Our observation period of 18 months, spanning pre- and post-study enrollment, may also limit comparison to other studies with shorter monitoring periods. Data collection for this study was completed in 2017 and may not capture contemporary improvements in medication regimens, however, sustained suppression over time remains essential for the health of PWH and to eliminate future infections.

Despite limitations, this study contributes to our understanding of factors associated with durable viral suppression over an 18-month time-period. Older age, simple ART regimens and social support were associated with durable viral suppression in this sample. Future studies that address risks (e.g., targeting younger PWH) and bolster facilitators of durable suppression are warranted.