Introduction

There is increasing interest in accurately monitoring antiretroviral therapy (ART) adherence for people with HIV (PWH). Tenofovir disoproxil fumarate (TDF) is included in fixed-dose combinations alongside emtricitabine and efavirenz, or lamivudine and dolutegravir, which are used by over 95% of people receiving ART in low-income and middle-income countries (LMICs) [1,2]. Therefore, objective tenofovir measurements could identify poor adherence. TDF is converted to tenofovir (TFV), which is metabolized intracellularly to tenofovir diphosphate (TFV-DP). TFV is excreted in urine and correlates with short-term adherence as it has a 12–15 h terminal half-life [3], whereas TFV-DP accumulates in red blood cells and correlates with medium-term adherence as it has a longer half-life of 17 days [4]. Studies have shown that qualitative urine TFV [5–8] and quantitative dried blood spot (DBS) TFV-DP levels [9–11] are associated with viral suppression in PWH receiving ART, but none have compared the two measures, or determined thresholds that best predict viral suppression. Furthermore, dolutegravir has a higher genetic barrier to resistance than efavirenz, meaning that measures of adherence should be more closely associated with viral suppression, unlike efavirenz where resistance can cause viraemia despite good adherence.

Therefore, we aimed to compare the association between urine TFV, and DBS TFV-DP concentrations, with viraemia among PWH receiving dolutegravir and efavirenz-based ART. In post hoc analyses, we also aimed to estimate optimal urine TFV and DBS TFV-DP thresholds to detect viraemia, and to assess associations between TFV levels and both HIV drug resistance (HIVDR), and self-reported adherence.

Methods

We conducted a cross-sectional analysis at enrolment into a randomised study of point-of-care HIV viral load testing (POwER) [12]. We included consecutively enrolled POwER participants receiving TDF as part of dolutegravir or efavirenz-based first-line ART. Eligible PWH had a preenrolment viral load greater than 1000 copies/ml in the past 6 weeks, without having received enhanced adherence counselling. At enrolment, participants self-reported adherence, and had urine, DBS and plasma samples taken and stored at −80 °C, for retrospective testing.

We quantitated urine TFV and DBS TFV-DP concentrations using liquid chromatography and dual tandem mass spectrometry (LC-MS/MS). We tested viral load using the cobas 6800 platform (Roche, Basel, Switzerland), and attempted drug resistance testing for all samples with viral load at least 1000 copies/ml (see testing details supplement, https://links.lww.com/QAD/D80).

We used logistic regression models to assess the relationship between the exposure of either urine TFV concentrations, or DBS TFV-DP concentrations, and the outcome of viraemia. To determine whether associations differed by ART regimen, we included a variable for ART regimen (dolutegravir versus efavirenz) in the model, with an interaction term between ART regimen and urine TFV, or DBS TFV-DP concentrations. We fitted separate models for the outcomes of viraemia at least 1000 copies/ml, and at least 50 copies/ml, as these thresholds are used in WHO guidelines [1]. We compared the Nagelkerke pseudo-R2 of the urine TFV and DBS TFV-DP models to determine which measure was more strongly associated with viraemia [13]. In exploratory, post hoc analyses, we used receiver-operating curves (ROCs) to estimate urine TFV and DBS TFV-DP thresholds that maximize sensitivity and specificity to predict concurrent viraemia. Lastly, we described urine TFV and DBS TFV-DP levels among people with and without HIVDR, and compared self-reported short-term and longer term adherence with urine TFV and DBS TFV-DP levels using logistic regression and linear regression models, respectively. Sample size was determined by the number of participants enrolled into POwER and receiving TDF.

We analysed data using R 4.2.0 (R Foundation for Statistical Computing, Vienna, Austria). The University of KwaZulu-Natal Biomedical Research Ethics Committee (BREC 00000836/2019) and the University of Oxford Tropical Research Ethics Committee (OxTREC 66-19) approved the study.

Results

Between August 2020-March 2022, we enrolled 124 PWH. 68 (54.8%) were women, the median age was 39 years [interquartile range (IQR) 34–45] and 74 (59.7%) were receiving efavirenz versus 50 (40.3%) receiving dolutegravir (Table S1, https://links.lww.com/QAD/D83). 23.4% self-reported missing a dose in the past 4 days, and 62.9% reported last missing a dose over 4 weeks before enrolment. Median time since the preenrolment viraemic viral load was 15 days (IQR 13–21). In December 2020, we discovered that 45 participants had preenrolment viral loads measured on a faulty analyser, with potentially false viraemic preenrolment results. This did not affect the enrolment viral loads used in this analysis but meant that not all participants had recent viraemia, and so there were a higher number of participants with viral suppression at enrolment than anticipated. Therefore, enrolment, viral loads were at least 1000 copies/ml in 44 of 124 participants (35.5%), 50–999 copies/ml in 23 of 124 (18.5%), and suppressed less than 50 copies/ml in 57 of 124 (46.0%). Among the 43 with successful HIVDR testing, 24 of 43 (55.8%) had mutations conferring resistance to their current regimen. Among those receiving efavirenz, 23 of 27 (85.2%) had resistance to their current regimen, versus 1 of 16 (6.3%) of those receiving dolutegravir (one person with M184V mutation alone). Median urine TFV concentration was 20 000 ng/ml (IQR 7280–33 625), and median TFV-DP concentration was 734 fmol/punch (IQR 471–1015).

Higher concentrations of urine TFV [1000 ng/ml increase, odds ratio (OR) 0.97 95% CI 0.94–0.99, P = 0.005) and DBS TFV-DP (100 fmol/punch increase, OR 0.76, 95% CI 0.67–0.86, P < 0.001) were associated with lower odds of viraemia at least 1000 copies/ml, with similar results at at least 50 copies/ml (Table 1a). There was some evidence that the association between urine TFV and viraemia at 1000 copies/ml (LRT for interaction P = 0.072), and between DBS TFV-DP and viraemia at both 1000 (P = 0.003) and 50 (P = 0.068) copies/ml, was stronger among people receiving dolutegravir than among people receiving efavirenz. There was no evidence of a difference by ART regimen in the association between urine TFV and viraemia at 50 copies/ml (P = 0.797, Table 1a). Overall, at both 1000 and 50 copies/ml thresholds, the Nagelkerke pseudo-R2 for the DBS TFV-DP models was higher than for the urine TFV models, meaning there was a stronger relationship between DBS TFV-DP and viraemia (Table 1a).

Table 1 - (a) Logistic regression models of the association between urine tenofovir concentrations, and dried blood spot tenofovir diphosphate concentrations, and viraemia; (b) diagnostic accuracy thresholds and areas under the curve for urine tenofovir and tenofovir diphosphate concentrations to detect viraemia. Overall Dolutegravir Efavirenz (a) Logistic regression models Odds ratio (95% CI) P Nagelkerke pseudo-R 2 Odds ratio (95% CI) P Odds ratio (95% CI) P Viraemia ≥1000 copies/ml  Urine TFV concentration (increase of 1000 ng/ml) 0.97 (0.94–0.99) 0.005 0.107 0.94 (0.89–0.98)a 0.007 0.98 (0.95–1.01)a 0.250  DBS TFV-DP concentration (increase of 100 fmol/punch) 0.76 (0.67–0.86) <0.001 0.274 0.53 (0.35–0.71)b <0.001 0.85 (0.74–0.95)b 0.009 Viraemia ≥50 copies/ml  Urine TFV concentration (increase of 1000 ng/ml) 0.98 (0.96–1.00) 0.041 0.097 0.98 (0.95–1.00)c 0.099 0.98 (0.96–1.01)c 0.225  DBS TFV-DP concentration (increase of 100 fmol/punch) 0.86 (0.78–0.93) <0.001 0.188 0.74 (0.58–0.89)d 0.010 0.90 (0.81–0.99)d 0.040 (b) Thresholds to detect viraemia and associated areas under the curve (AUCs) Thresholde (sens, spec [%]) AUC [%] (95% CI) Thresholde (sens, spec) AUC [%] (95% CI) Thresholde (sens, spec [%]) AUC [%] (95% CI) Viraemia ≥1000 copies/ml  Urine TFV concentration (ng/ml) 8340 (85.0, 54.5) 69.3 (58.6–80.0) 3495 (76.5, 93.9) 83.2 (68.3–98.1) 20 100 (59.3, 57.4) 59.2 (44.6–73.8)  DBS TFV-DP concentration (fmol/punch) 639 (68.2, 76.3) 77.9 (68.9–87.0) 483 (88.2, 84.8) 90.1 (80.2–100) 816 (74.1, 70.2) 74.7 (62.5–87.0) Viraemia ≥50 copies/ml  Urine TFV concentration (ng/ml) 18 350 (56.7, 63.2) 62.4 (52.4–72.3) 16 000 (54.5, 70.6) 66.8 (51.9–81.8) 20 100 (58.8, 60.0) 59.5 (46.0–73.0)  DBS TFV-DP concentration (fmol/punch) 795 (74.6, 68.4) 76.4 (68.0–84.8) 686 (81.8, 76.5) 84.4 (73.1–95.7) 816 (64.7, 70.0) 68.3 (55.8–80.8)

AUC, area under the curve; CI, confidence interval; DBS TFV-DP, dried blood spot tenofovir diphosphate; DTG, dolutegravir; EFV, efavirenz; TFV, tenofovir.

aLRT for interaction P = 0.072.

bLRT for interaction P = 0.003.

cLRT for interaction P = 0.797.

dLRT for interaction P = 0.068.

eCalculated as the point, which minimizes (1 − Sens)2 + (1 − Spec)2.

Overall, AUCs for concurrent viraemia at 1000 copies/ml were modest (urine TFV 69.3%, DBS TFV-DP 77.9%, Table 1b). However, AUCs were higher among people receiving dolutegravir compared with efavirenz, using both urine TFV (0.83, 95% CI 0.68–0.98 versus 0.59, 95% CI 0.45–0.74, P = 0.026 Fig. 1a) and DBS TFV-DP (0.90, 95% CI 0.80–1.00 versus 0.75, 95% CI 0.62–0.87, P = 0.059, Fig. 1b). From the ROC curves among people receiving dolutegravir, a urine TFV concentration threshold of 3495 ng/ml would have 77% sensitivity and 94% specificity to predict concurrent viraemia at least 1000 copies/ml (Table S2, https://links.lww.com/QAD/D84). A DBS TFV-DP concentration of 483 fmol/punch would have 88% sensitivity and 85% specificity to predict concurrent viraemia at least 1000 copies/ml. When using a threshold of at least 50 copies/ml, AUCs were similar to more than 1000 copies/ml, except with urine TFV among people receiving dolutegravir, which performed less well (Table 1a, Figure S1, https://links.lww.com/QAD/D81), compared to a threshold of 1000 copies/ml.

Fig. 1:

Receiver operating characteristic curves of urine tenofovir and dried blood spot tenofovir diphosphate concentrations to predict viraemia at least 1000 copies/ml, by antiretroviral therapy regimen.

Among 43 people with viraemia greater than 500 copies/ml and successful HIVDR testing, median urine TFV (17 300 ng/ml, IQR 1120–29350 versus 343 ng/ml, 0–20 950) and DBS TFV-DP levels (646 fmol/punch, IQR 388–820 versus 103 fmol/punch, 10–374) were higher among people with HIVDR compared with those without, but numbers were too small for formal comparisons or meaningful breakdown by ART regimen (Table S3, https://links.lww.com/QAD/D85).

Self-reported missed doses in the past 4 days and more recently self-reported missed doses were both associated with lower urine TFV and DBS TFV-DP concentrations (Table S4, https://links.lww.com/QAD/D86 and Figure S2, https://links.lww.com/QAD/D82).

Discussion

In this cross-sectional study, we found that urine TFV and DBS TFV-DP concentrations were negatively associated with concurrent viraemia, and the association was generally stronger with dolutegravir compared with efavirenz. Furthermore, DBS TFV-DP had a better association with viraemia compared with urine TFV.

Our findings are similar to studies which have shown that DBS TFV-DP concentrations are associated with viraemia in PWH receiving TDF-based ART [9–11]. A study among 532 PWH in the United States found that higher DBS TFV-DP was associated with viral load less than 20 copies/ml. Thirty-six percent of participants were receiving integrase inhibitors, and 27% were receiving nonnucleoside reverse transcriptase inhibitors. Among people with viral load less than 20 copies/ml, the mean TFV-DP concentration was 1728 (1608–1857) fmol/punch compared with 1469 (1283–1681) fmol/punch at 20–200 copies/ml, and 633 (542–739) fmol/punch at greater than 200 copies/ml, but unlike our study, results were not presented by ART class [9]. A South African cross-sectional study among 137 people taking efavirenz used ROC curves to demonstrate that DBS TFV-DP was more strongly associated than plasma TFV with viral suppression less than 50 copies/ml [10]. Lastly, among 250 virally suppressed PWH receiving efavirenz in South Africa, baseline TFV-DP less than 400 fmol/punch was associated with increased odds of developing viraemia at least 400 copies/ml after 1 month [11]. Regarding urine TFV levels, several studies demonstrate qualitative point-of-care urine TFV levels are associated with concurrent viraemia [5–8], but no studies have assessed the relationship between quantitative TFV concentrations and viraemia.

Although both urine TFV and DBS TFV-DP concentrations were negatively associated with viraemia at least 1000 copies/ml, the association was weaker with efavirenz, likely because the high prevalence of HIVDR means viraemia persists in the presence of measurable adherence. With dolutegravir, HIVDR was rare, meaning the relationship between TFV measures and viraemia was stronger. In exploratory, post hoc analyses using ROC curves, we similarly found that the potential for urine TFV and DBS TFV-DP to predict concurrent viraemia was poor to modest with efavirenz but more acceptable with dolutegravir. Using DBS TFV-DP, a threshold of around 480 fmol/punch would have more than 80% sensitivity and specificity to predict concurrent viraemia at least 1000 copies/ml. Comparing pseudo-R2 values, models indicated superiority of fit for TFV-DP over urine TFV, suggesting that TFV-DP generally performed better. This is likely because quantitative DBS TFV-DP reflects longer term adherence, which is required to achieve viral suppression, and which is not captured by more transient, shorter term urine TFV concentrations.

Strengths of our study include the comprehensive assessment of adherence using viral load, HIVDR, self-reported adherence, urine TFV and DBS TFV-DP. To our knowledge, this is the first study to directly compare the predictive value of quantitative urine TFV concentrations against DBS TFV-DP in PWH. We used different viraemia thresholds, which reflect WHO guidelines [1]. Our study is limited by the small sample size, with participants enrolled in a clinical trial for people with recent viraemia, meaning results may not be generalizable to other populations. Although we estimate urine TFV and DBS TFV-DP thresholds to optimize sensitivity and specificity to detect viraemia at both at least 50 and at least 1000 copies/ml, these are exploratory analyses with imprecise estimates because of our small sample size. These thresholds should be evaluated in other studies, and should take into account the clinical use scenario, and whether sensitivity or specificity should be maximized [14]. Therefore, we provide ranges of thresholds and associated sensitivities and specificities in Table S2, https://links.lww.com/QAD/D84.

Our study suggests that with the global dolutegravir rollout, urine TFV or DBS TFV-DP measures of adherence may be increasingly useful to measure adherence and complement viral load testing. Qualitative point-of-care urine TFV assays have been validated [5–8], and their clinical effectiveness will need to be confirmed in clinical trials before they can be rolled out in clinical practice [15]. Although TFV-DP performed better in our study, it currently requires expensive LC-MS/MS, which limits widespread clinical use in LMICs. Development of point-of-care TFV-DP assays should be prioritized [16], alongside studies to establish if there is a reliable TFV-DP threshold for predicting concurrent viraemia among people receiving dolutegravir. People with viraemia who have a TFV-DP level higher than this threshold (suggesting good adherence) could be more likely to have HIVDR, [5,17] and so could be prioritized for HIVDR testing.

Acknowledgements

The authors would like to thank all participants in the study and acknowledge the work and support of staff at the Prince Cyril Zulu Clinic, Mafakathini Clinic, eThekwini Municipality, CAPRISA, AHRI Pharmacology Core, KRISP and the National Health Laboratory Services at Inkosi Albert Luthuli Hospital. Thank you to Ms Lara Lewis for statistical input.

Funding: this work is supported by grants from the Wellcome Trust PhD Programme for Primary Care Clinicians (216421/Z/19/Z) and the University of Oxford's Research England QR Global Challenges Research Fund (0007365). HIV drug resistance testing and drug concentration testing was funded by the National Institute for Health and Care Research (NIHR) Community Healthcare MedTech and In Vitro Diagnostics Co-operative at Oxford Health NHS Foundation Trust (MIC-2016–018); G.H., C.C.B., and P.J.T. also receive funding from this award. J.D., Academic Clinical Lecturer (CL-2022–13–005), is funded by the UK National Institute of Health and Social Care Research (NIHR). The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. For the purpose of open access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.

Author contributions: J.D., P.K.D. and N.G. conceived the study. K.G., P.M., R.L. and N.S. were responsible for laboratory testing. R.L., Y.S., P.T., C.B. and G.H. contributed to study design. J.D. analysed the data with guidance from T.F. J.D. wrote the first draft of the manuscript. All authors critically reviewed and edited the manuscript and consented to final publication.

Data sharing statement: bona fide researchers will be able to request access to anonymised trial data by contacting the corresponding author.

Conflicts of interest

There are no conflicts of interest.

References 1. World Health Organization. Consolidated guidelines on HIV prevention, testing, treatment, service delivery and monitoring: recommendations for a public health approach. 2021, World Health Organization: Geneva, Switzerland. 2. Clinton Health Access Initiative. 2021 HIV Market Report: The state of the HIV market in low- and middle-income countries. 2021, Clinto Health Access Initiative. Boston, USA. 3. Barditch-Crovo P, Deeks SG, Collier A, Safrin S, Coakley DF, Miller M, et al. Phase I/II trial of the pharmacokinetics, safety, and antiretroviral activity of tenofovir disoproxil fumarate in human immunodeficiency virus-infected adults. Antimicrob Agents Chemother 2001; 45:2733–2739. 4. Castillo-Mancilla JR, Zheng JH, Rower JE, Meditz A, Gardner EM, Predhomme J, et al. Tenofovir, emtricitabine, and tenofovir diphosphate in dried blood spots for determining recent and cumulative drug exposure. AIDS Res Hum Retroviruses 2013; 29:384–390. 5. Hermans LE, Umunnakwe CN, Lalla-Edward ST, Hebel SK, Tempelman HA, Nijhuis M, et al. Point-of-care tenofovir urine testing for the prediction of treatment failure and drug resistance during initial treatment for human immunodeficiency virus type 1 (HIV-1) infection. Clin Infect Dis 2023; 76:e553–e560. 6. van Zyl G, Jennings L, Kellermann T, Nkantsu Z, Cogill D, van Schalkwyk M, et al. Urine tenofovir-monitoring predicts HIV viremia in patients treated with high genetic-barrier regimens. AIDS 2022; 36:2057–2062. 7. Marryshow TA, Muhairwe J, Tang A, Molulela MMM, Matta R, Jordan MR. Determining the acceptability of point-of-care urine tenofovir testing and its performance in predicting HIV RNA suppression. Int J STD AIDS 2022; 33:777–783. 8. Dorward J, Lessells R, Govender K, Moodley P, Samsunder N, Sookrajh Y, et al. Diagnostic accuracy of a point-of-care urine tenofovir assay, and associations with HIV viraemia and drug resistance among people receiving dolutegravir and efavirenz-based antiretroviral therapy. J Int AIDS Soc 2023; 26:e26172. 9. Castillo-Mancilla JR, Morrow M, Coyle RP, Coleman SS, Gardner EM, Zheng JH, et al. Tenofovir diphosphate in dried blood spots is strongly associated with viral suppression in individuals with human immunodeficiency virus infections. Clin Infect Dis 2019; 68:1335–1342. 10. Phillips TK, Sinxadi P, Abrams EJ, Zerbe A, Orrell C, Hu NC, et al. A comparison of plasma efavirenz and tenofovir, dried blood spot tenofovir-diphosphate, and self-reported adherence to predict virologic suppression among South African women. J Acquir Immune Defic Syndr 2019; 81:311–318. 11. Jennings L, Robbins RN, Nguyen N, Ferraris C, Leu CS, Dolezal C, et al. ADD-ART team. Tenofovir diphosphate in dried blood spots predicts future viremia in persons with HIV taking antiretroviral therapy in South Africa. AIDS 2022; 36:933–940. 12. Dorward J, Sookrajh Y, Ngobese H, Lessells R, Sayed F, Bulo E, et al. Protocol for a randomised feasibility study of Point-Of-care HIV viral load testing to Enhance Re-suppression in South Africa: the POwER study. BMJ Open 2021; 11:e045373. 13. Harrell FE. Regression modeling strategies: with applications to linear models, logistic regression, and survival analysis. New York: Springer; 2015. 14. Altman DG, Bland JM. Diagnostic tests 3: receiver operating characteristic plots. BMJ 1994; 309:188. 15. Bardon AR, Dorward J, Sookrajh Y, Sayed F, Quame-Amaglo J, Pillay C, et al. Simplifying TREAtment and Monitoring for HIV (STREAM HIV): protocol for a randomised controlled trial of point-of-care urine tenofovir and viral load testing to improve HIV outcomes. BMJ Open 2021; 11:e050116. 16. Olanrewaju AO, Sullivan BP, Bardon AR, Lo TJ, Cressey TR, Posner JD, Drain PK. Pilot evaluation of an enzymatic assay for rapid measurement of antiretroviral drug concentrations. Virol J 2021; 18:77. 17. McCluskey SM, Govender K, Adamson J, Gandhi M, Spinelli MA, Moosa MY, et al. Point-of-care urine tenofovir testing to predict HIV drug resistance among individuals with virologic failure. AIDS 2023; 37:1109–1113.