We determined that tenofovir diphosphate in dried blood spots, which is a measure of adherence and cumulative drug exposure, was strongly associated with viral suppression (<20 copies/mL) in persons living with human immunodeficiency virus taking tenofovir -based regimens.
Keywords: adherence, dried blood spots, tenofovir diphosphate, antiretroviral therapy, pharmacokinetics
Abstract
Background
Although tenofovir diphosphate (TFV-DP) in dried blood spots (DBS) is a predictor of adherence and pre-exposure prophylaxis efficacy, its utility in human immunodeficiency virus (HIV) treatment remains unknown.
Methods
DBS for TFV-DP were collected up to 3 times over 48 weeks in persons living with HIV (PLWH) who were receiving TFV disoproxil fumarate (TDF)-based therapy. Log-transformed baseline TFV-DP was compared using t-tests or analyses of variance; generalized estimating equations were used to estimate the adjusted odds ratio (aOR) of viral suppression (<20 copies/mL) based on the TFV-DP concentration at the study visit.
Results
We analyzed 1199 DBS from 532 participants (76 female; 101 Black, 101 Hispanic). Among the virologically-suppressed participants at baseline (n = 347), TFV-DP was lower in Blacks (geometric mean 1453, 95% confidence interval [CI] 1291–1635) vs Whites (1793, 95% CI 1678–1916; P = .002) and Hispanics (1760, 95% CI 1563–1982; P = .025); in non-boosted (1610, 95% CI 1505–1723) vs. boosted (1888, 95% CI 1749–2037; P = .002) regimens; and in non-nucleoside reverse transcription inhibitor–based (1563, 95% CI 1432–1707) vs. boosted protease inhibitor–based (1890, 95% CI 1704–2095; P = .006) and multiclass-based (1927, 95% CI 1650–2252; P = .022) regimens. The aOR of virologic suppression, after adjusting for age, gender, race, body mass index, estimated glomerular filtration rate, CD4+ T-cell count, antiretroviral drug class and duration of therapy, was 73.5 (95% CI 25.7–210.5; P < .0001) for a TFV-DP concentration ≥1850 fmol/punch compared to <350 fmol/punch.
Conclusions
TFV-DP in DBS is strongly associated with virologic suppression in PLWH on TDF-based therapy and is associated with certain participant characteristics. Further research is required to evaluate this drug adherence and exposure measure in clinical practice.
Clinical Trials Registration
The main driver of antiretroviral therapy (ART) exposure and human immunodeficiency virus (HIV)-related outcomes is medication adherence [1–3]. Optimal adherence is critical to achieve and sustain viral suppression in people living with HIV (PLWH) [4, 5]. Despite this, we currently lack a gold standard measure of ART adherence and exposure that can be used in clinical practice [6, 7]. In particular, measures that distinguish between viremia due to suboptimal adherence vs inadequate viral response (eg, persistent viremia due to drug resistance or low cumulative drug exposure despite high adherence) are currently unavailable.
Tenofovir diphosphate (TFV-DP), arising from tenofovir disoproxil fumarate (TDF)- and tenofovir alafenamide (TAF)-based regimens [8, 9], is the phosphorylated anabolite of tenofovir (TFV) and exerts its pharmacodynamic effect in HIV-infected cells. This anabolite is also abundant in red blood cells [10, 11] and dried blood spots (DBS) [11–13], where it exhibits unique pharmacology, with a long half-life of 17 days and a 25-fold accumulation from the first dose to steady state [11]. These distinctive pharmacologic characteristics have been leveraged to develop an adherence gradient in HIV-negative persons [11, 14] that is predictive of HIV pre-exposure prophylaxis (PrEP) efficacy [15–17] and correlates with other adherence measures in HIV [18, 19]. However, this biomarker as an indicator of viral suppression and drug exposure in PLWH on ART has not been evaluated. This is a significant gap in knowledge given the >20 million PLWH receiving ART worldwide [20] (many of whom are on TFV-based regimens in the United States [>85%] [21] and around the world [22]) and the proportion in whom viral suppression has not been achieved [23].
We therefore investigated the association of TFV-DP with viral suppression in a clinical cohort of PLWH being treated with TDF-based ART.
METHODS
Study Design and Participants
Study participants were recruited at the University of Colorado Hospital Infectious Diseases Group Practice in Aurora, Colorado. Inclusion criteria required that participants were 18 years or older, taking a TDF-based regimen (for any duration of time), and planned to have blood drawn for routine HIV viral load (VL) analysis at every study visit as part of their routine medical care. There were no restrictions on the co-existence of comorbidities for enrollment. Prospective enrollment of participants occurred on a first-come and first-served basis during their clinic visit. After obtaining informed consent, 4–6 mL of whole blood were collected in 1 ethylenediaminetetraacetic acid tube from peripheral venipuncture during the participant’s clinical blood draw. A maximum of 3 study visits (ie, 3 blood samples) were obtained within a 48-week period of time. Participants were compensated $10.00 for every study visit, for a maximum of $30.00. Study visits were dependent on participants’ scheduled follow-up with their provider, but were required to be at least 2 weeks apart to allow for 1 half-life of TFV-DP in DBS [11]. Study enrollment was initiated in June 2014, and follow-up for the last enrolled participant concluded in July 2017. Throughout this period, the rate of viral suppression within the University of Colorado Hospital Infectious Diseases Group Practice ranged from 84–90%, and TDF continued to be the most-prescribed nucleoside analogue, as the clinic’s prescribing practices did not automatically switch to TAF after its 2015 approval. The study was approved by the Colorado Multiple Institutional Review Board (COMIRB #13–2104) and registered with clinicaltrials.gov (NCT02012621).
Our primary outcome, HIV VL, was evaluated at all study visits to inform patient care, and DBS samples were collected to coincide with these available measurements. Through August 2016, DBS at each study visit were consecutively assayed, regardless of HIV VL, as originally planned. At this point, logistical and practical constraints limited our analytical capacity and a reduced DBS assay strategy was implemented. For the remainder of the study, DBS assays were performed for all study visits for participants with a detectable HIV VL measurement at 1 or more visits, while DBS assays were discontinued for participants with suppressed HIV VL at all observed visits. This outcome-dependent sampling strategy is similar to a case (viremic) control (suppressed) study design extended to a longitudinal setting. Under outcome-dependent sampling, statistical power is maintained by enriching the sample for the reasonably rare outcome (<20% of study visits) [24].
Quantification of Tenofovir Diphosphate in Dried Blood Spots
For the DBS preparation, 25 mcl of whole blood were pipetted 5 times into a Whatman 903 Protein Saver card, as previously described [11, 12]. These cards were allowed to dry at room temperature for at least 2 hrs (and up to overnight), after which they were stored in plastic bags with humidity indicators at -80oC until analysis [11, 12, 14]. TFV-DP was quantified from a 3 mm punch using liquid chromatography/tandem mass spectrometry, as previously validated [12].
Self-reported Adherence Assessment
At each study visit, participants were asked about their 3-month adherence to their current ART regimen using a visual analog scale ranging from 0% to 100% adherence [25, 26]. Briefly, study personnel asked the participant to mark on a horizontal line that ranged from 0% to 100%, with predefined intervals every 10%, their best estimate of adherence in that period. Since study visits were dependent on scheduling for clinical care, overlap on the time period assessing self-reported adherence was possible.
HIV Viral Load Analysis
Quantitative HIV-1 VL analysis was performed using the Roche cobas 6800 HIV test (linear range 20 to 107 copies/mL) at the University of Colorado Hospital clinical laboratory, which is certified under the Clinical Laboratory Improvement Amendment of 1988.
Statistical Analysis
TFV-DP concentrations were log-transformed for analysis to address right skew, and results were back-transformed to the geometric mean for ease of interpretation in units of drug concentration (fmol/punch) [27]. Baseline TFV-DP concentrations were compared using t-tests or analyses of variance, or their non-parametric counterparts, as appropriate. In order to analyze all available visit data, drug concentrations below the limit of quantification were imputed to 12.5 fmol/punch, which is between 0 and the lower limit of quantification of the assay [12]. TFV-DP concentrations were classified into 5 categories based on previous studies in HIV-uninfected individuals [14]. While the adherence:concentration benchmarks in PLWH have not been determined, this classification was considered to be a reasonable starting point for this analysis. To accommodate repeated measures over time, generalized estimating equations with a logit link were used to estimate the odds ratios (OR) of viral suppression (<20 copies/mL), comparing each TFV-DP category to the reference category of <350 fmol/punch. An adjusted OR (aOR) was obtained by including covariates for age, gender, race, body mass index (BMI), estimated glomerular filtration rate (using Modification of Diet in Renal Disease equation), CD4+ T-cell count, ART class, and duration of ART in the model. Adjustment variables were selected a priori based on previous findings on the pharmacology of TFV in plasma [28, 29] and TFV-DP in peripheral blood mononuclear cells (PBMCs) [30] and DBS [14]. Current ART therapy duration was also categorized, with the highest category being >6 months, when TFV-DP in DBS would be at a steady state [11, 14]. A separate, generalized estimating equation model of viral suppression considered the association of suppression with the 3-month, self-reported adherence categories: <28.5% (reference category), 28.5% to <50%, 50% to 84%, 85% to 99%, and 100%, aiming to mirror the TFV-DP categories and also based on previous studies in HIV-uninfected volunteers. These models of viral suppression, utilizing either TFV-DP in DBS or 3-month self-reported adherence, were compared using the quasilikelihood under the independence model criterion [31].
To address the outcome-dependent sampling and confirm the strength of our findings, we performed a sensitivity analysis limiting data to those collected during the initial phase of the study (ie, before August 2016, when the DBS assay strategy was modified).
RESULTS
Study Population
A total of 807 participants were enrolled, 619 (77%) of whom were suppressed to <20 copies/mL at enrollment, contributing 1939 person-visits and 1936 DBS samples, with 444 completing all 3 visits (275 suppressed at all visits), 244 completing 2 visits (189 suppressed at both visits), and 119 completing only 1 visit (98 suppressed). The demographic characteristics of the participants for whom DBS were analyzed for drug concentrations (n = 532) are presented in Table 1, and comparisons with the participants for whom drug concentrations in DBS were not quantified (n = 272) are presented in Supplementary Table 1, without major differences observed between these 2 groups.
Table 1.
Demographic Characteristics of Enrolled Participants Included in the Analysis
| Characteristic | Participants Included in Analysis (n = 532) | Participants With DBS at Enrollment Visit (n = 521) | |
|---|---|---|---|
| No. (%) or Median (IQR) | |||
| No. (%) or Median (IQR) | Viremic (n = 174) | Suppressed (n = 347) | |
| Age | 46 (36, 52) | 43 (33, 51) | 46 (38, 53) |
| Gender | |||
| Female | 76 (14%) | 18 (10%) | 56 (16%) |
| Male | 456 (86%) | 156 (90%) | 291 (84%) |
| Race/ethnicity | |||
| Black | 101 (19%) | 35 (20%) | 64 (18%) |
| White | 305 (57%) | 95 (55%) | 203 (59%) |
| Hispanic | 101 (19%) | 36 (21%) | 63 (18%) |
| Other | 25 (5%) | 8 (5%) | 17 (5%) |
| Body mass index (Kg/m2) | |||
| <18.5 | 21 (4%) | 7 (4%) | 14 (4%) |
| 18.5–25 | 225 (42%) | 80 (46%) | 138 (40%) |
| 25–30 | 175 (33%) | 54 (31%) | 119 (34%) |
| >30 | 109 (21%) | 33 (19%) | 74 (21%) |
| eGFR (mL/min) | 87 (74, 102) | 90 (76, 105) | 86 (73, 100) |
| CD4+ T-cell count (cells/mm3) | |||
| <200 | 58 (11%) | 43 (25%) | 15 (4%) |
| 200–350 | 80 (15%) | 29 (17%) | 49 (14%) |
| 350–500 | 78 (15%) | 26 (15%) | 49 (14%) |
| >500 | 316 (59%) | 76 (44%) | 234 (67%) |
| HIV viral load (copies/mL) | 132 (43, 699) | 132 (42, 660) | - |
| Time on current ART (months) | |||
| <1 | 21 (4%) | 17 (10%) | 4 (1%) |
| 1–3 | 45 (8%) | 22 (13%) | 22 (6%) |
| 3–6 | 30 (6%) | 15 (9%) | 15 (4%) |
| >6 | 436 (82%) | 120 (69%) | 306 (88%) |
| Type of ART | |||
| NNRTI-based | 141 (27%) | 22 (13%) | 116 (33%) |
| INSTI-based | 191 (36%) | 77 (44%) | 110 (32%) |
| b/PI-based | 133 (25%) | 47 (27%) | 84 (24%) |
| Multiclass | 67 (13%) | 28 (16%) | 37 (11%) |
| Pharmacologic booster | |||
| No | 259 (49%) | 60 (34%) | 193 (56%) |
| Yes | 273 (51%) | 114 (66%) | 154 (44%) |
| 3-month self-reported adherence (%)a | 98 (90, 100) | 90 (80, 100) | 99 (90, 100) |
Abbreviations: ART, antiretroviral therapy; b/PI, boosted protease inhibitor; DBS, dried blood spots; eGFR, estimated glomerular filtration rate; HIV, human immunodeficiency virus; INSTI: integrase strand-transfer inhibitor; IQR, inter-quartile range; NNRTI, non-nucleoside reverse transcriptase inhibitor.
aData on self-reported adherence in the preceding 3 months were available for n = 482 participants.
Among the 532 participants for whom DBS were analyzed (n = 1199 person-visits), the median time between visits was 16 (range 2, 48) weeks. Across person-visits, 839 (70%) samples had an HIV VL <20 copies/mL, 213 (18%) had 20–200, and 147 (12%) had >200 copies/mL. The HIV VL in viremic participants upon enrollment is shown in Table 1, and ranged from 20 to 331000 copies/mL.
Tenofovir Diphosphate Concentrations in Dried Blood Spots
TFV-DP was quantified in DBS from 1199 person-visits (derived from 532 participants) and included in the drug concentration analysis; 11 samples (<1%) were below the limit of quantification. TFV-DP concentrations were available for 521 participants at enrollment (7 samples had no paired HIV VL and 4 samples had missing drug concentrations). The baseline demographic characteristics of these 521 participants are shown in Table 1. The geometric mean concentration of TFV-DP in all participants at enrollment was 1450 fmol/punch (95% confidence interval [CI] 1357–1548), with higher overall concentrations in the suppressed (1728, 95% CI 1601–1865) vs. viremic participants (1021, 95% CI 917–1138; P < .0001).
The TFV-DP concentrations, according to participant characteristics and viral suppression status at enrollment, are shown in Table 2. In an unadjusted analysis limited to the suppressed group only, TFV-DP concentrations were significantly lower in Blacks (1453, 95% CI 1291–1635) vs. Whites (1793, 95% CI 1678–1916; P = .002) and Hispanics (1760, 95% CI 1563–1982; P = .025). In this same group, participants with BMIs <18.5 Kg/m2 had higher TFV-DP concentrations (2404, 95% CI 1875–3082) when compared to participants with a BMI 25–30 Kg/m2 (1718, 95% CI 1577–1870; P = .012) or >30 Kg/m2 (1416, 95% CI 1271–1578; P = .0002). Furthermore, suppressed participants taking non-nucleoside reverse transcription inhibitor (NNRTI)-based regimens also had lower concentrations of TFV-DP (1563, 95% CI 1432–1707) when compared to participants taking a boosted protease inhibitor (1890, 95% CI 1704–2095; P = .006) or a multiclass regimen (1927, 95% CI 1650–2252; P = .022). Lastly, use of a pharmacologic booster among virologically-suppressed participants (1888, 95% CI 1749–2037) was associated with higher TFV-DP concentrations compared to no booster use (1610, 95% CI 1505–1723; P = .002).
Table 2.
TFV-DP Concentrations in Dried Blood Spots at the Time of Enrollment
| Characteristic | TFV-DP (fmol/punch) GM (95% CI) | ||||
|---|---|---|---|---|---|
| All Participants (n = 521) | Viremic (n = 174) | P-Value | Suppressed (n = 347) | P-Value | |
| Gender | |||||
| Male | 1436 (1338, 1542) | 1045 (885, 1233) | REF | 1704 (1611, 1802) | REF |
| Female | 1533 (1287, 1826) | 842 (516, 1373) | .41 | 1860 (1637, 2112) | .22 |
| Race/ethnicity | |||||
| Black | 1253 (1077, 1457) | 955 (671, 1359) | REF | 1453 (1291, 1635) | REF |
| White | 1508 (1382,1645) | 1041 (841, 1290) | .680 | 1793 (1678, 1916) | .002 |
| Hispanic | 1479 (1272, 1720) | 1092 (771,1546) | .600 | 1760 (1563, 1982) | .025 |
| Other | 1493 (1106, 2016) | 808 (386, 1691) | .690 | 1993 (1585, 2506) | .016 |
| Body mass index (Kg/m2) | |||||
| <18.5 | 1750 (1260, 2430) | 928 (421, 2043) | REF | 2404 (1875, 3082) | REF |
| 18.5–25 | 1465 (1324, 1623) | 968 (767, 1223) | .920 | 1863 (1722, 2017) | .056 |
| 25–30 | 1494 (1333, 1675) | 1099 (827, 1460) | .690 | 1718 (1577, 1870) | .012 |
| >30 | 1293 (1118, 1495) | 1054 (732, 1516) | .770 | 1416 (1271, 1578) | .0002 |
| CD4+ T-cell count (cells/mm3) | |||||
| <200 | 988 (814, 1200) | 809 (590, 1109) | REF | 1754 (1369, 2246) | REF |
| 200–350 | 1402 (1185, 1657) | 1036 (705, 1521) | .328 | 1676 (1462, 1922) | .750 |
| 350–500 | 1381 (1164, 1638) | 952 (634, 1428) | .534 | 1682 (1467, 1929) | .770 |
| >500 | 1589 (1461, 1729) | 1188 (937, 1506) | .056 | 1747 (1641, 1860) | .980 |
| HIV viral load (copies/mL) | |||||
| <20 | 1728 (1608, 1857) | … | … | 1728 (1608, 1857) | … |
| 20–200 | 1469 (1283, 1681) | 1469 (1283, 1681) | … | … | … |
| >200 | 633 (542, 739) | 633 (542, 739) | … | … | … |
| Time on current ART (months) | |||||
| <1 | 691 (501 954) | 660 (402, 1084) | REF | 842 (525, 1351) | REF |
| 1–3 | 1523 (1219, 1902) | 1389 (898, 2148) | .028 | 1669 (1365, 2042) | .009 |
| 3–6 | 1754 (1340, 2297) | 1567 (924, 2658) | .020 | 1964 (1539, 2507) | .002 |
| >6 | 1476 (1374, 1585) | 974 (808, 1174) | .150 | 1738 (1646, 1834) | .003 |
| Type of ART | |||||
| NNRTI-based | 1428 (1256, 1623) | 885 (567, 1380) | REF | 1563 (1432, 1707) | REF |
| INSTI-based | 1433 (1283, 1599) | 1095 (863, 1389) | .400 | 1729 (1580, 1892) | .117 |
| b/PI-based | 1518 (1330, 1731) | 1026 (757, 1391) | .590 | 1890 (1704, 2095) | .006 |
| Multiclass | 1413 (1172, 1704) | 938 (632, 1391) | .850 | 1927 (1650, 2252) | .022 |
| Pharmacologic booster | |||||
| No | 1431 (1302, 1573) | 979 (748, 1280) | REF | 1610 (1505, 1723) | REF |
| Yes | 1468 (1339, 1609) | 1045, (860, 1269) | .700 | 1888 (1749, 2037) | .002 |
| 3-month self-reported adherence (%) | |||||
| 100 | 1840 (1665, 2035) | 1766 (1312, 2375) | REF | 1863 (1726, 2010) | REF |
| 85 to 99 | 1519 (1377, 1676) | 1144 (877, 1491) | .032 | 1686 (1560, 1821) | .071 |
| 50 to 84 | 802 (673, 955) | 456 (326, 639) | <.0001 | 1488 (1256, 1763) | .018 |
| 28 to <50 | 545 (356, 833) | 475 (247, 916) | .0004 | 1002 (514, 1954) | .071 |
| <28.5a | 680 (336, 1375) | 561 (180, 1749) | .056 | 1207 (470, 3102) | … |
The italicized values are meant to show statistical significance.
Abbreviations: ART, antiretroviral therapy; b/PI, boosted protease inhibitor; CI, confidence interval; GM, geometric mean; HIV, human immunodeficiency virus; INSTI, integrase strand-transfer inhibitor; NNRTI, non-nucleoside reverse transcriptase inhibitor; REF, reference; TFV-DP, tenofovir diphosphate.
aOnly 1 participant who was suppressed reported <28.5% adherence.
In both viremic and suppressed participants, a longer current ART duration upon enrollment was associated with higher TFV-DP concentrations (Table 2). Similarly, drug concentrations increased significantly with higher 3-month, self-reported adherence in both groups, as shown in Table 2. TFV-DP concentrations in viremic and suppressed participants who reported 100% adherence were 1766 (95% CI 1313–2375) and 1863 (95% CI 1726–2010) fmol/punch, respectively.
After adjusting for age, race, gender, BMI, CD4+ T-cell count, estimated glomerular filtration rate, ART class, duration of therapy, and self-reported adherence in the virologically-suppressed group, male participants had 13% (95% CI -24–-0.8; P = .037) lower TFV-DP concentrations compared to women (Supplementary Table 2). Similarly, Whites and Hispanics had 15% (95% CI 0.4–31; P = .044) and 20% (95% CI 2–41; P = .024) higher TFV-DP concentrations vs Blacks, respectively. Participants with BMIs >30 Kg/m2 had 47% (95% CI 31–59; P < .0001) lower drug concentrations compared to those with BMIs <18.5 Kg/m2 (Supplementary Table 2).
Self-reported Adherence
A total of 482 participants (337 suppressed, 145 viremic) had data available for 3-month, self-reported adherence at enrollment, with a median adherence of 99% (interquartile range [IQR] 90–100) and 90% (IQR 80–100) adherence in suppressed and viremic participants, respectively (Table 1). Median adherence was similar in women (100%, IQR 90–100) vs men (98%, IQR 90–100; P = .38) and in Blacks (95%, IQR 90–100) compared to Whites (98%, IQR 90–100; P = .18) and Hispanics (98%, IQR 90–100; P = .65). Adherence in participants taking an NNRTI-based regimen (100%, IQR 94–100) was higher compared to participants taking a boosted protease inhibitor–based (96%, IQR 85–100; P = .01), an integrase-based (97%, IQR 90–100; P = .07), or a multiclass-based (95%, IQR 87–100; P = .006) regimen. Participants taking a pharmacologic booster (95%, IQR 85–100) had lower adherence compared to those who were not (99%, IQR 93–100; P < .0001).
Association of Tenofovir Diphosphate in Dried Blood Spots With Viral Suppression
The distribution of all person-visits by TFV-DP concentration category and 3-month self-reported adherence are shown in Tables 3 and 4, respectively. The aORs of viral suppression, based on the TFV-DP concentration categories, are shown in Table 3 and Figure 1. The highest estimated aOR for suppression was 73.5 (95% CI 25.7–210.5; P < .0001) for TFV-DP concentrations ≥1850 fmol/punch vs <350 fmol/punch. The estimated aOR of viral suppression, based on self-reported adherence, are shown in Table 4 and Supplementary Figure 1, estimated as 8.5 (95% CI 2.3–30.9; P = .0012) for 100% vs. <28.5% adherence and 8.6 (95% CI 2.3–31.2; P = .011) for 85–99% vs. <28.5% adherence. The quasilikelihood under the independence model criterion for the model including TFV-DP in DBS was substantially lower vs. the model including 3-month self-reported adherence (1149 vs. 1233), indicating the TFV-DP model fit the data better.
Table 3.
Odds Ratio of Viral Suppression by TFV-DP in Dried Blood Spots (n = 1199 Person-Visits)
| TFV-DP (fmol/punch) | Person-Visits, n (%) | Median (IQR) HIV VL, copies/ mL | OR Suppression (95% CI) | P-Value | aOR Suppressiona (95% CI) | P-Value |
|---|---|---|---|---|---|---|
| <350 | 47 (4%) | 17540 (3600, 51300) | 1 | REF | 1 | REF |
| 351 to 699 | 63 (5%) | 312 (70, 7640) | 5.7 (2.3, 14.4) | .0002 | 8.9 (2.6, 30.4) | .0005 |
| 700 to 1249 | 217 (18%) | 142 (38, 1268) | 18.1 (8.4, 39.0) | <.0001 | 32.8 (11.6, 93.1) | <.0001 |
| ≥1250 to 1849 | 357 (30%) | 51 (28, 146) | 31.2 (14.5, 67.5) | <.0001 | 49.2, (17.3, 139.8) | <.0001 |
| ≥1850 | 515 (43%) | 55 (33, 219) | 40.3 (18.7, 87.1) | <.0001 | 73.5 (25.7, 210.5) | <.0001 |
The italicized values are meant to show statistical significance.
Abbreviations: aOR, adjusted odds ratio; ART, antiretroviral therapy; CI, confidence interval; HIV VL, human immunodeficiency virus viral load; IQR, interquartile range; OR, odds ratio; TFV-DP, tenofovir diphosphate;
aAdjusted for age, gender, race, body mass index, estimated glomerular filtration rate, CD4+ T-cell count, ART class, and duration of ART.
Table 4.
OR of Viral Suppression by Self-reported Adherence in the Preceding 3 Months (n = 1152 Person-Visits)
| 3-Month Self- report (%) | Person-Visits, n (%) | Median (IQR) HIV VL (copies/mL) | OR Suppression (95% CI) | P-Value | aOR Suppressiona (95% CI) | P-Value |
|---|---|---|---|---|---|---|
| <28.5 | 10 (1%) | 49700 (3210, 57900) | 1 | REF | 1 | REF |
| 28 to <50 | 18 (2%) | 474 (120, 27775) | 1.5 (0.4, 5.4) | .55 | 1.6 (0.3, 8.9) | .57 |
| 50 to 84 | 137 (12%) | 624 (85, 24150) | 3.4 (1.3, 8.8) | .0135 | 3.0 (0.8, 11.4) | .11 |
| 85% to 99 | 456 (40%) | 64 (36, 340) | 8.9 (3.5, 22.5) | <.0001 | 8.6 (2.3, 31.2) | .0011 |
| 100 | 531 (46%) | 53 (30, 298) | 10.3 (4.1, 26.1) | <.0001 | 8.5 (2.3, 30.9) | .0012 |
The italicized values are meant to show statistical significance.
Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; HIV VL, human immunodeficiency virus viral load; IQR, interquartile range; OR, odds ratio.
aAdjusted for age, gender, race, body mass index, estimated glomerular filtration rate, CD4+ T-cell count, ART class and duration of ART.
Figure 1.
The aOR of HIV VL <20 copies/mL by concentration of TFV-DP in dried blood spots (N = 1199 person visits). Abbreviations: aOR, adjusted odds ratio; ART, antiretroviral therapy; CI, confidence interval; HIV VL, human immunodeficiency virus virus load; IQR, interquartile range; TFV-DP, tenofovir diphosphate. *Adjusted for age, gender, race, body mass index, estimated glomerular filtration rate, CD4+ T-cell count, ART class, and duration of ART.
Sensitivity Analysis
In the subset of participants for whom DBS were consecutively analyzed prior to the modification of the assay strategy, the aOR for viral suppression was 81.3 (95% CI 17.9–368.6; P < .0001) for a TFV-DP concentration of ≥1850 fmol/punch, compared to <350 fmol/punch (Supplementary Table 3).
DISCUSSION
In this study, we found that TFV-DP in DBS was strongly associated with virologic suppression to <20 copies/mL, and that this association was stronger than self-reported adherence. This association increased with higher TFV-DP concentrations, and remained significant after covariate adjustment, including ART class and therapy duration. In comparison to the extensive body of literature that has demonstrated the utility of TFV-DP in DBS as a predictor of PrEP efficacy [15–17], this study provides new additional insights about its association with viral suppression and potential applications in HIV infection.
Our findings suggest PLWH may have higher TFV-DP concentrations in DBS than HIV-uninfected volunteers [11, 14] and that TFV-DP concentrations in DBS are likely to be influenced by individual characteristics unique to this population, such as the use of concomitant medications or additional comorbidities. For example, while TFV-DP concentrations were lower in virologically-suppressed Black PLWH compared to other races, they were still higher than previous estimates in HIV-uninfected Blacks [14], despite similar self-reported adherence. This suggests that ART adherence may not be the main driver of these differences; they could be influenced by drug-drug interactions, chronic inflammation [32], and/or unique pharmacogenetics/biology in PLWH. Similarly, BMI was also associated with TFV-DP concentrations in participants with viral suppression, even after adjusting for race and gender, suggesting that TFV and TFV-DP exposure could be influenced by body habitus. This is consistent with previous observations, where low BMI was associated with high plasma TFV [28, 29, 33]. We also observed associations between TFV-DP exposure and ART regimen. For instance, virally-suppressed participants taking an NNRTI-based regimen (mainly efavirenz) had an average TFV-DP concentration of 1563 fmol/punch vs. 1888 fmol/punch for suppressed participants taking a regimen including a pharmacologic booster. These findings are consistent with previous observations, where high levels of TFV in plasma [28–30, 34–38], TFV-DP in PBMCs [30, 39] and TFV in hair [40] were observed in patients taking pharmacologic boosters. Among the possible explanations are that ritonavir and cobicistat may increase TDF (and consequently TFV) uptake into the systemic circulation by altering the activity of P-glycoprotein (ie, via reduction of TDF efflux) and by inhibiting esterase cleavage of TDF at the luminal gut level [41] or that pharmacologic boosters may slow the renal clearance of TFV due to an inhibition of efflux transporters, such as the multidrug resistance proteins 4 and 2 in the renal proximal tubule [38, 42]. Collectively, these findings suggest TFV-DP in DBS may have a different pharmacokinetic profile in PLWH compared to HIV-uninfected adults, and that the known adherence benchmarks in the HIV-uninfected population are unlikely to apply directly to PLWH. A directly-observed dosing study in PLWH would help establish expected TFV-DP benchmarks in this population and help parse sources of TFV-DP variability.
Along with the association of TFV-DP with viral suppression in PLWH, potential applications in clinical practice should be discussed. These include the possibility of utilizing TFV-DP in DBS as a tool to inform ART adherence and exposure beyond HIV VL, in particular if this biomarker is developed as a point-of-care test. While HIV VL has been traditionally used as a surrogate for ART adherence, it cannot provide information about drug exposure or pharmacologic forgiveness. For example, HIV VL cannot detect low drug exposure in suppressed individuals, which could trigger an early intervention to improve adherence and prevent viral rebound. In this context, TFV-DP in DBS could serve as a true predictor of viral rebound, in particular if a low concentration is associated with future viremia (despite viral suppression at the time of analysis). Conversely, since HIV VL cannot identify high cumulative drug exposure in the presence of viremia, a high TFV-DP concentration in a patient with a persistently-elevated HIV VL could suggest drug resistance and trigger early HIV genotyping. Furthermore, given its unique pharmacology, TFV-DP in DBS could also be used as a tool to monitor and/or predict drug-related toxicity in clinical practice, as it is plausible that high accumulation could precede overt clinical toxicity or drug discontinuation (as seen in high-risk individuals taking TDF for PrEP) [43, 44]. Future research is needed to evaluate these potential applications.
Our study offers several strengths, including a large sample size within a prospective clinical cohort reflective of a routine clinical practice. This provides for generalizability of our findings and sets the framework for the use of TFV-DP as a measure of cumulative adherence and exposure in clinical care. Additional strengths include the stronger association of this objective adherence biomarker in comparison with self-reported ART adherence and the range of potential applications in clinical practice. Among the limitations are that our study was observational, with adherence based on self-report and virologic suppression. In addition, our outcome-dependent sampling strategy could have strengthened our association, although our sensitivity analysis, limited to person-visits that were consecutively obtained, provided a similar, although larger, aOR. Lastly, the study was restricted to participants on TDF-based regimens. However, DBS from participants switching to TAF-based regimens during the course of the study were collected and will be analyzed in the future.
In summary, we demonstrated that TFV-DP in DBS is strongly associated with viral suppression in PLWH. In addition, TFV-DP exposure in our population was higher than that previously observed in HIV-negative volunteers, and was associated with several patient characteristics, such as race, BMI, ART class, and the use of a pharmacologic booster. Further research is needed to better understand how these individual factors contribute to the pharmacology of TFV-DP in patients on ART, including those on TDF- and TAF-based regimens. Finally, the utility of this adherence biomarker as a tool to monitor ART adherence in clinical practice should be evaluated.
Supplementary Data
Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.
Notes
Author contributions. J. R. C.-M. led the conception and study design; obtained the funding and regulatory approvals; led all aspects regarding study monitoring, logistics, data and sample collection, and result interpretation; wrote the first manuscript draft; and performed all the edits for all the subsequent drafts. M. M. performed data and statistical analysis and interpretation, generated figures and tables, and made substantial edits and critical revisions of the manuscript. R. P. C. and S. S. C. gathered participant consent and performed data and sample collection, data management, and data analysis and interpretation, and made substantial edits to and critical revisions of the manuscript. E. M. G. participated in the study design, assisted with the adherence data interpretation, and performed manuscript editing and critical revisions. J.-H. Z., L. E., and L. R. B. led the sample processing, pharmacologic analysis, and data validation for the drug concentrations and made substantial edits and critical revisions of the manuscript. J. J. K. participated in the study design, adherence, and pharmacologic data interpretation and performed manuscript editing and critical revisions. S. M. contributed to the study design and conceptualization; performed the sample size calculation, data management, statistical analysis, and interpretation; generated figures and tables; and made substantial edits and critical revisions of the manuscript. P. L. A. co-led the study conception and design, assisted with obtaining the funding, supported the study monitoring and logistics, directed and supported all aspects of the pharmacologic and drug concentration analysis, collaborated with data interpretation, and made substantial edits and critical revisions of the original manuscript and all its subsequent versions.
Acknowledgments. The authors thank the study participants and the personnel at the Colorado Antiviral Pharmacology Laboratory for their invaluable assistance and support of this study. They thank the director of the University of Colorado Hospital Human Immunodeficiency Virus program (Steven Johnson, MD) and the medical assistants (Nancy Olague, Brittany Limon, Ariel Cates, Maureen Sullivan and Missy Sorrell) at the University of Colorado Hospital-Infectious Disease Group Practice for their invaluable contributions and support of this study.
Financial support. This work was supported by the National Institutes of Health (grant numbers K23 AI104315 and R21 AI124859 to J. R. C.-M. and R01 AI122298 to P. L. A.).
Potential conflicts of interest. P. L. A. and J. J. K. received research funding from Gilead Sciences, paid to their institution. All other authors report no potential conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
Presented in part: Conference on Retroviruses and Opportunistic Infections, Boston, MA, 5–8 March 2018. Abstract number 25.
References
- 1. Maggiolo F, Airoldi M, Kleinloog HD, et al. Effect of adherence to HAART on virologic outcome and on the selection of resistance-conferring mutations in NNRTI- or PI-treated patients. HIV Clin Trials 2007; 8:282–92. [DOI] [PubMed] [Google Scholar]
- 2. Mannheimer S, Friedland G, Matts J, Child C, Chesney M. The consistency of adherence to antiretroviral therapy predicts biologic outcomes for human immunodeficiency virus-infected persons in clinical trials. Clin Infect Dis 2002; 34:1115–21. [DOI] [PubMed] [Google Scholar]
- 3. Gardner EM, Burman WJ, Steiner JF, Anderson PL, Bangsberg DR. Antiretroviral medication adherence and the development of class-specific antiretroviral resistance. AIDS 2009; 23:1035–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Gardner EM, Burman WJ, Maravi ME, Davidson AJ. Durability of adherence to antiretroviral therapy on initial and subsequent regimens. AIDS Patient Care STDS 2006; 20:628–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Nachega JB, Marconi VC, van Zyl GU, et al. HIV treatment adherence, drug resistance, virologic failure: evolving concepts. Infect Disord Drug Targets 2011; 11:167–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Castillo-Mancilla JR, Haberer JE. Adherence measurements in HIV: new advancements in pharmacologic methods and real-time monitoring. Current HIV/AIDS Reports 2018; 15:1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Garrison LE, Haberer JE. Technological methods to measure adherence to antiretroviral therapy and preexposure prophylaxis. Curr Opin HIV AIDS 2017; 12:467–74. [DOI] [PubMed] [Google Scholar]
- 8. King T, Bushman L, Kiser J, et al. Liquid chromatography-tandem mass spectrometric determination of tenofovir-diphosphate in human peripheral blood mononuclear cells. J Chromatogr B Analyt Technol Biomed Life Sci 2006; 843:147–56. [DOI] [PubMed] [Google Scholar]
- 9. Ray AS, Fordyce MW, Hitchcock MJ. Tenofovir alafenamide: A novel prodrug of tenofovir for the treatment of Human Immunodeficiency Virus. Antiviral Res 2016; 125:63–70. [DOI] [PubMed] [Google Scholar]
- 10. Durand-Gasselin L, Da Silva D, Benech H, Pruvost A, Grassi J. Evidence and possible consequences of the phosphorylation of nucleoside reverse transcriptase inhibitors in human red blood cells. Antimicrob Agents Chemother 2007; 51:2105–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Castillo-Mancilla JR, Zheng JH, Rower JE, et al. Tenofovir, emtricitabine, and tenofovir diphosphate in dried blood spots for determining recent and cumulative drug exposure. AIDS Res Hum Retroviruses 2013; 29:384–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Zheng JH, Rower C, McAllister K, et al. Application of an intracellular assay for determination of tenofovir-diphosphate and emtricitabine-triphosphate from erythrocytes using dried blood spots. J Pharm Biomed Anal 2016; 122:16–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Castillo-Mancilla J, Coyle RP, Zheng JH, et al. Tenofovir diphosphate arising from TAF is quantifiable in dried blood spots [Abstract 405]. In: Program and abstracts of Conference on Retroviruses and Opportunistic Infection 2017 (Seattle, WA). [Google Scholar]
- 14. Anderson PL, Liu AY, Castillo-Mancilla JR, et al. Intracellular tenofovir-diphosphate and emtricitabine-triphosphate in dried blood spots following directly observed therapy. Antimicrob Agents Chemother 2018; 62:e01710–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Grant RM, Anderson PL, McMahan V, et al. ; iPrEx study team. Uptake of pre-exposure prophylaxis, sexual practices, and HIV incidence in men and transgender women who have sex with men: a cohort study. Lancet Infect Dis 2014; 14:820–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Liu AY, Cohen SE, Vittinghoff E, et al. Preexposure prophylaxis for HIV infection integrated with municipal- and community-based sexual health services. JAMA Intern Med 2016; 176:75–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Hosek SG, Rudy B, Landovitz R, et al. ; Adolescent Trials Network for HIVAIDS Interventions. An HIV preexposure prophylaxis demonstration project and safety study for young MSM. J Acquir Immune Defic Syndr 2017; 74:21–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Castillo-Mancilla JR, Searls K, Caraway P, et al. Short communication: Tenofovir diphosphate in dried blood spots as an objective measure of adherence in HIV-infected women. AIDS Res Hum Retroviruses 2015; 31:428–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Warne P, Robbins R, Anderson P, et al. Utility of dried blood spot-derived ARV biomarkers as an objective measure of treatment adherence in South Africa [Abstract 241]. In: Program and abstracts of International Association of Physicians in AIDS Care Adherence 2015 (Miami, FL). [Google Scholar]
- 20. World Health Organization/United Nations Programme on HIV/AIDS. Fact sheet-Global HIV statistics Available at: http://wwwunaidsorg/sites/default/files/media_asset/UNAIDS_FactSheet_enpdf Accessed 1 July 2018.
- 21. Eaton EF, Tamhane A, Davy-Mendez T, et al. Trends in antiretroviral therapy prescription, durability and modification: new drugs, more changes, but less failure. AIDS 2018; 32:347–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Brennan AT, Bor J, Davies M-A, et al. Drug side effects and retention on HIV treatment: a regression discontinuity study of tenofovir implementation in South Africa and Zambia. Am J Epidemiol 2018; 187:1990-2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Gonsalves GS, Paltiel AD, Cleary PD, et al. A flow-based model of the HIV care continuum in the United States. J Acquir Immune Defic Syndr 2017; 75:548–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Schildcrout JS, Schisterman EF, Mercaldo ND, Rathouz PJ, Heagerty PJ. Extending the case-control design to longitudinal data: stratified sampling based on repeated binary outcomes. Epidemiology 2018; 29:67–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Giordano TP, Guzman D, Clark R, Charlebois ED, Bangsberg DR. Measuring adherence to antiretroviral therapy in a diverse population using a visual analogue scale. HIV Clinical Trials 2004; 5:74–9. [DOI] [PubMed] [Google Scholar]
- 26. Amico KR, Fisher WA, Cornman DH, et al. Visual analog scale of ART adherence: association with 3-day self-report and adherence barriers. J Acquir Immune Defic Syndr 2006; 42:455–9. [DOI] [PubMed] [Google Scholar]
- 27. Motulsky H. Intuitive biostatistics: a nonmathematical guide to statistical thinking. Oxford, UK: Oxford University Press, 2014. [Google Scholar]
- 28. Cattaneo D, Minisci D, Baldelli S, et al. Effect of cobicistat on tenofovir disoproxil fumarate (TDF): what is true for TAF may also be true for TDF. J Acquir Immune Defic Syndr 2018; 77:86–92. [DOI] [PubMed] [Google Scholar]
- 29. Baxi SM, Greenblatt RM, Bacchetti P, et al. Common clinical conditions—age, low BMI, ritonavir use, mild renal impairment—affect tenofovir pharmacokinetics in a large cohort of HIV-infected women. AIDS 2014; 28:59–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Pruvost A, Negredo E, Théodoro F, et al. Pilot pharmacokinetic study of human immunodeficiency virus-infected patients receiving tenofovir disoproxil fumarate (TDF): investigation of systemic and intracellular interactions between TDF and abacavir, lamivudine, or lopinavir-ritonavir. Antimicrob Agents Chemother 2009; 53:1937–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Cui J, de Klerk N, Abramson M, et al. Fractional polynomials and model selection in generalized estimating equations analysis, with an application to a longitudinal epidemiologic study in Australia. Am J Epidemiol 2009; 169:113–21. [DOI] [PubMed] [Google Scholar]
- 32. Seifert SM, Castillo-Mancilla JR, Erlandson KM, Anderson PL. Inflammation and pharmacokinetics: potential implications for HIV-infection. Expert Opin Drug Metab Toxicol 2017; 13:641–50. [DOI] [PubMed] [Google Scholar]
- 33. Gervasoni C, Meraviglia P, Landonio S, et al. Low body weight in females is a risk factor for increased tenofovir exposure and drug-related adverse events. PLoS One 2013; 8:e80242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Agarwala S, Eley T, Villegas C, et al. Pharmacokinetic interaction between tenofovir and atazanavir coadministered with ritonavir in healthy subjects. In: Program and abstracts of the Sixth International Workshop on Clinical Pharmacology of HIV Therapy 2005. [Google Scholar]
- 35. Calcagno A, Gonzalez de Requena D, Simiele M, et al. Tenofovir plasma concentrations according to companion drugs: a cross-sectional study of HIV-positive patients with normal renal function. Antimicrob Agents Chemother 2013; 57:1840–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Kearney BP, Mathias A, Mittan A, Sayre J, Ebrahimi R, Cheng AK. Pharmacokinetics and safety of tenofovir disoproxil fumarate on coadministration with lopinavir/ritonavir. J Acquir Immune Defic Syndr 2006; 43:278–83. [DOI] [PubMed] [Google Scholar]
- 37. Hoetelmans RM, Mariën K, De Pauw M, et al. Pharmacokinetic interaction between TMC114/ritonavir and tenofovir disoproxil fumarate in healthy volunteers. Br J Clin Pharmacol 2007; 64:655–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Kiser JJ, Carten ML, Aquilante CL, et al. The effect of lopinavir/ritonavir on the renal clearance of tenofovir in HIV-infected patients. Clin Pharmacol Ther 2008; 83:265–72. [DOI] [PubMed] [Google Scholar]
- 39. Lahiri CD, Tao S, Jiang Y, et al. Impact of protease inhibitors on intracellular concentration of tenofovir-diphosphate among HIV-1 infected patients. AIDS 2015; 29:1113–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Gandhi M, Gandhi RT, Stefanescu A, et al. ; A5321 Team. Cumulative antiretroviral exposure measured in hair is not associated with measures of HIV persistence or inflammation among individuals on suppressive ART. J Infect Dis 2018; 218:234–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Tong L, Phan TK, Robinson KL, et al. Effects of human immunodeficiency virus protease inhibitors on the intestinal absorption of tenofovir disoproxil fumarate in vitro. Antimicrob Agents Chemother 2007; 51:3498–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Ray AS, Cihlar T, Robinson KL, et al. Mechanism of active renal tubular efflux of tenofovir. Antimicrob Agents Chemother 2006; 50:3297–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Tang EC, Vittinghoff E, Anderson PL, et al. Changes in kidney function associated with daily tenofovir disoproxil fumarate/emtricitabine for HIV preexposure prophylaxis use in the United States demonstration project. J Acquir Immune Defic Syndr 2018; 77:193–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Gandhi M, Murnane PM, Bacchetti P, et al. Hair levels of preexposure prophylaxis drugs measure adherence and are associated with renal decline among men/transwomen. AIDS 2017; 31:2245–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.