Abstract
In clinical trials, once-daily oral tenofovir-based pre-exposure prophylaxis (PrEP) significantly reduced HIV-1 acquisition risk; however, this was highly dependent on medication adherence and persistence. We report clinical characteristics, PrEP usage patterns, first evidence of HIV-1, and associated risk factors among adults with commercial insurance using oral PrEP in the United States using health plan claims from the IQVIA PharMetrics® Plus database between January 1, 2015, and March 31, 2020, from individuals who newly initiated emtricitabine/tenofovir disoproxil fumarate (FTC/TDF) or FTC/tenofovir alafenamide (TAF) for daily PrEP. Overall, 25,419 individuals were included (FTC/TDF, n = 24,232; FTC/TAF, n = 1187), with generally similar characteristics reported during the 6-month baseline period across cohorts. Mean follow-up length was 504 and 77 days for FTC/TDF and FTC/TAF, respectively, corresponding with the 2019 approval of FTC/TAF for PrEP. Similarly, mean PrEP use duration was 354 and 68 days for FTC/TDF and FTC/TAF, respectively. PrEP breaks (>90-day gap) were observed in 11.1% of individuals using FTC/TDF, with a mean break duration of 249 days; 20.0% of individuals using FTC/TDF and 7.3% using FTC/TAF had ≥1 sexually transmitted infection diagnosis during follow-up. From 6 to 12 months of follow-up, mean FTC/TDF proportion of days covered (PDC; 0.74 vs. 0.67) and persistence (70.2% vs. 57.4%) decreased; real-world PDC and persistence were lower than reported in globally conducted clinical trials. First evidence of HIV-1 was infrequent among individuals using FTC/TDF (0.6%), though 60.3% had PrEP on hand when HIV-1 definition was met; high-risk sexual behavior, syphilis, and gonorrhea were the most important risk factors.
Introduction
Despite the availability of effective antiretroviral therapy (ART) regimens, HIV-1 remains a significant source of morbidity and early mortality in the ∼1.2 million people living with HIV-1 in the United States. 1 –3 Therefore, it is critical to implement both nonpharmacological (e.g., access to condoms and sterile syringes) and pharmacological HIV-1 prevention strategies. 4
Pre-exposure prophylaxis (PrEP) is a pharmacological HIV-1 prevention strategy recommended by national and international guidelines and health care societies. 5 –8 The US Food and Drug Administration (FDA) has approved both once-daily oral emtricitabine/tenofovir disoproxil fumarate (FTC/TDF; for adults: July 16, 2012; for adolescents: May 15, 2018) and FTC/tenofovir alafenamide (TAF; for adults and adolescents: October 3, 2019), and every 2-month cabotegravir long-acting (CAB-LA; for adults and adolescents: December 20, 2021) injectable for PrEP. 9 –12 Individuals who may acquire HIV-1 through sex are indicated to use FTC/TDF or CAB-LA, whereas FTC/TAF is approved to prevent sexually transmitted HIV-1, excluding exposure through receptive vaginal sex. 13 Clinical trials have shown that once-daily tenofovir-based PrEP provided significant protection against HIV-1 acquisition among various populations. 14 –18 Notably, medication adherence was a critical determinant of efficacy in these trials, 18 –21 and current guidelines emphasize the importance of adherence support when providing PrEP. 5,6 The phase 2b/3 HPTN 083 (NCT02720094) and phase 3 HPTN 084 (NCT03164564) studies demonstrated that CAB-LA led to notably lower HIV-1 incidence than once-daily oral FTC/TDF in men and transgender women who have sex with men (HPTN 083) and cisgender women (HPTN 084), supporting it as a promising strategy for improving adherence and outcomes among individuals who use PrEP. 22,23
Several studies have documented the effectiveness and usage patterns associated with FTC/TDF for PrEP in real-world US populations. In a Kaiser Permanente study, high medication adherence (mean medication possession ratio, 92%) and no HIV-1 seroconversions were observed among individuals (97.9% male; 69.6% White) using FTC/TDF over an average follow-up of 0.9 years. 24 In the prospective Demo Project, which enrolled men who have sex with men and transgender women at high risk of HIV-1 acquisition from three clinics, the HIV-1 incidence rate was low (0.43 per 100 person-years), with ≥80% of individuals showing protective FTC/TDF plasma concentrations through 48 weeks. 25 However, the ATN 113 and ATN 110 studies (NCT01769456 and NCT01772823) reported much higher rates of seroconversion (3.3–6.4 per 100 person-years) in young men (aged 15–22 years) who have sex with men, with <35% of participants exhibiting protective FTC/TDF plasma concentrations 48 weeks after initiation. 26,27 Although these studies demonstrate the effectiveness of FTC/TDF for PrEP and emphasize the critical role of adherence in maintaining protection against HIV-1 in the real world, the studies were conducted in few US clinics, which limits the generalizability of findings to other regions. 24 –27 PrEP is particularly underused in the South, where several barriers, including stigma, pre-authorization requirements from insurers, and lack of insurance, may hinder PrEP access and use. 28 –31 Additionally, there are no published studies assessing real-world effectiveness and usage patterns of FTC/TAF for PrEP, indicating a significant data gap. 32,33 Therefore, a comprehensive assessment of current unmet needs associated with oral PrEP (FTC/TDF and FTC/TAF) in a large, geographically diverse sample of the US population is warranted. The objective of this US study was to report clinical characteristics, PrEP usage patterns, first evidence of HIV-1, and associated risk factors in individuals with commercial insurance using oral PrEP.
Methods
Data source
This was a retrospective cohort study of data from the IQVIA PharMetrics® Plus (IQVIA, Raleigh, NC) database collected from January 1, 2015, to March 31, 2020. IQVIA PharMetrics Plus is a health plan claims database composed of fully adjudicated medical and pharmacy claims for >210 million enrollees since 2006. Contributors to the database are largely commercial health plans, and the database is representative of the commercially insured US national population aged <65 years and contains a longitudinal view of inpatient and outpatient services, prescription and office/outpatient administered drugs, costs, and detailed enrollment information. Data are de-identified and compliant with the Health Insurance Portability and Accountability Act; therefore, ethics committee or institutional review board approval and informed consent were not required.
Study design and population
The study population included individuals with ≥1 dispensing of FTC/TDF or FTC/TAF with ≥30 days of supply, with the date of first PrEP medication dispensing defined as the index date. Individuals who used PrEP were aged ≥18 years at the index date and had ≥6 months of continuous health plan enrollment before the index date (baseline period). Individuals with both FTC/TDF and FTC/TAF dispensed on the index date, ≥1 HIV-1 diagnosis (International Classification of Diseases, 10th Revision, Clinical Modification [ICD-10-CM]: B20, R75, Z21; ICD-9-CM: 042, V08, 795.71) during the baseline period (6 months before the index date) or on the index date, ≥1 HIV-2 diagnosis (ICD-10-CM: B97.35; ICD-9-CM: 079.53) any time during the eligibility period, ≥1 dispensing of FTC/TDF or FTC/TAF with <30 days of supply any time before or on the index date, and/or ≥1 dispensing of ART (other than FTC/TDF and FTC/TAF) during the baseline period or on the index date were excluded. For the FTC/TAF cohort only, individuals with ≥1 dispensing of FTC/TAF before its approval date for PrEP (October 3, 2019) were also excluded. Individuals who used PrEP were then classified into mutually exclusive cohorts based on their index medication (FTC/TDF or FTC/TAF).
The follow-up period (i.e., observation period) spanned from the index date until the earliest of either health plan disenrollment or end of data availability (March 31, 2020). For the evaluation of proportion of days covered (PDC) and PrEP persistence, the follow-up period was censored at first evidence of HIV-1, if applicable.
Study outcomes
Primary endpoints
Primary endpoints were the demographics and baseline clinical characteristics of individuals using PrEP and PrEP usage patterns reported overall and/or by index medication. Clinical characteristics (including sexually transmitted infection [STI] diagnoses) were evaluated during the baseline period, demographic information and year or quarter of index date were evaluated on the index date, and usage patterns and new STI diagnoses were evaluated during the follow-up period.
Usage patterns included duration of PrEP use, number of dispensings, and mean days of supply per dispensing. Proportions of people switching from FTC/TDF to FTC/TAF and vice versa were reported. PrEP breaks (defined by restarting PrEP after discontinuing for >90 days) data, including proportion of people who had ≥1, ≥3, and ≥5 PrEP breaks and mean number and duration of PrEP breaks, were evaluated. Additionally, the frequency of new STI diagnoses was reported.
Secondary endpoints
Secondary endpoints included PDC, PrEP persistence, first evidence of HIV-1, and associated risk factors.
Proportion of days covered was calculated for each individual using FTC/TDF as the total number of days with FTC/TDF on hand during the time interval of interest divided by the duration of the interval of interest and was assessed at 6 and 12 months of follow-up (censored at first evidence of HIV-1) among the subsets with ≥6 and ≥12 months of follow-up, respectively. After PDC calculation, proportions of individuals using FTC/TDF meeting certain thresholds (≥0.7 and ≥0.6) were reported. These limits were selected to conservatively allow for PrEP dosing schedules less restrictive than daily dosing.
Medication persistence among people using FTC/TDF was assessed during follow-up using Kaplan–Meier analysis, accounting for right censoring. Persistence rates were reported at 3, 6, 9, and 12 months after the index date and presented using Kaplan–Meier curves. PrEP nonpersistence was defined either as a gap of >90 days between the end of the days of supply of a dispensing and the start date of the next fill or as a gap of >90 days between the end of the days of supply of the last dispensing and the end of the observation period (censored at first evidence of HIV-1).
The primary definition of first evidence of HIV-1 was the earliest date between an ICD-9/10-CM HIV-1 diagnosis and a dispensing of non-PrEP ART, for which both a diagnosis and dispensing were required to be considered as having HIV-1. A sensitivity definition was also considered, using ≥2 dispensings of non-PrEP ART on separate days. First evidence of HIV-1 during follow-up was evaluated using Kaplan–Meier analysis and reported among all individuals using FTC/TDF. Time to first evidence of HIV-1 (both from the index date and from the latest FTC/TDF dispensing) was evaluated among individuals using FTC/TDF with evidence of HIV-1. A post hoc analysis was conducted among a larger population of individuals using FTC/TDF without requiring ≥6 months of continuous health plan enrollment before the index date. In addition, baseline demographics and clinical characteristics of people using PrEP (FTC/TDF or FTC/TAF) were considered as potential risk factors associated with the primary definition of first evidence of HIV-1.
Data and statistical analyses
All analyses were conducted using SAS® Enterprise Guide®, Version 7.15 (SAS Institute, Cary, NC). The timeline for analyses (ending on March 31, 2020) limited follow-up time for individuals using FTC/TAF (approved October 3, 2019); therefore, postindex results including FTC/TAF should be interpreted with caution.
Risk factors were evaluated and reported using odds ratios, 95% confidence intervals (CIs), and p values from a multivariable logistic regression model. Potential risk factors included baseline age, sex, region, health insurance plan type, Quan-Charlson Comorbidity Index score (derived from a published algorithm 34 ), high-risk sexual behavior, medication use, health care resource use, comorbidities, and STI diagnoses. Univariate associations were assessed for each characteristic present in ≥0.5% of the study population using logistic regression, and variables with p ≤ 0.1 were included in the subsequent variable selection process. Risk factors not associated with the outcome were discarded using stepwise selection with the Akaike information criterion. To mitigate bias associated with the stepwise selection, nonparametric bootstrap procedures with 499 replications were used to randomly resample the study population with replacement, and risk factors were evaluated in each bootstrap sample. Risk factors were chosen if they were selected in ≥50% of the bootstrap samples.
Results
Demographics and clinical characteristics
A total of 25,419 individuals using PrEP met the study inclusion criteria; 24,232 (95.3%) initiated FTC/TDF and 1187 (4.7%) initiated FTC/TAF. Mean observation duration was longer among the FTC/TDF cohort than the FTC/TAF cohort (504 vs. 77 days; Table 1), corresponding with the later FDA approval of FTC/TAF. 10 Overall, the mean age of individuals using PrEP was 35.1 years; 94.5% were male. Baseline demographics and clinical characteristics indicated individuals using PrEP were generally healthy, with a low mean Quan-Charlson Comorbidity Index score of 0.12. During the baseline period, 9.6% and 7.9% of individuals using FTC/TDF and FTC/TAF, respectively, had a diagnosis of any STI, with the most common for both cohorts being human papillomavirus, unspecified venereal disease, and syphilis. The prevalence of high-risk sexual behavior was low (1.1%); however, these diagnosis codes may be underused in administrative data.
Baseline Demographics and Clinical Characteristics of Individuals Using PrEP by Regimen and Overall
The observation (follow-up) period spanned from the index date until the earliest of end of continuous eligibility or end of data availability.
Evaluated on the index date.
Indemnity/traditional (FTC/TDF, n = 118; FTC/TAF, n = 5; overall, n = 123); consumer-directed health care (FTC/TDF, n = 25; FTC/TAF, n = 0; overall, n = 25); unknown (FTC/TDF, n = 67; FTC/TAF, n = 0; overall, n = 67).
The index date was defined as the first date of the first index medication dispensing.
Evaluated during the 6-month baseline period, excluding the index date.
Derived from a published algorithm 34 that assigns weights to categories of comorbidities. Calculated as SUM (chronic pulmonary disease, rheumatologic disease, chronic diabetes, renal disease, 2*congestive heart failure, 2*dementia, MAX [2*liver disease, 4*moderate liver disease], 2*hemiplegia, 4*AIDS, MAX [2* malignancy, 6*metastatic solid tumor]).
Identified using diagnosis codes ICD-9-CM: V69.2 and ICD-10-CM: Z72.5x.
ED, emergency department; FTC, emtricitabine; HMO, health maintenance organization; ICD-10-CM, International Classification of Diseases, 10th Revision, Clinical Modification; IQR, interquartile range; OP, outpatient; POS, point-of-service; PPO, preferred provider organization; PrEP, pre-exposure prophylaxis; SD, standard deviation; STI, sexually transmitted infection; TAF, tenofovir alafenamide; TDF, tenofovir disoproxil fumarate.
PrEP usage, adherence, and persistence
Mean duration of PrEP use was longer for the FTC/TDF cohort than the FTC/TAF cohort (354 vs. 68 days; Table 2), which reflects the aforementioned differential follow-up times between the two cohorts. The mean number of medication dispensings was 9.0 and 2.5 for FTC/TDF and FTC/TAF, respectively; 10.8% of people using FTC/TDF switched to FTC/TAF, and 0.8% using FTC/TAF switched to FTC/TDF. During the follow-up period, 11.1% of individuals using FTC/TDF had ≥1 PrEP break, with a mean duration of 249 days; 20.0% and 7.3% of individuals using FTC/TDF and FTC/TAF, respectively, had ≥1 diagnosis of any STI.
Usage Patterns Among Individuals Taking FTC/TDF and FTC/TAF
Duration was calculated as the time between the date of the first index medication dispensing and the end of the days of supply of the last dispensing.
Defined by restarting PrEP after discontinuing for >90 days.
Evaluated among individuals with ≥1 PrEP break.
FTC, emtricitabine; IQR, interquartile range; NA, not applicable; PrEP, pre-exposure prophylaxis; SD, standard deviation; STI, sexually transmitted infection; TAF, tenofovir alafenamide; TDF, tenofovir disoproxil fumarate.
Among individuals using FTC/TDF with ≥6 months (n = 18,261) and ≥12 months (n = 12,667) of follow-up, mean (standard deviation [SD]) PDC was 0.74 (0.30) and 0.67 (0.33), respectively (Fig. 1A). These values corresponded to 63.7% and 57.9% of people using FTC/TDF with PDC ≥0.7 at 6 and 12 months, respectively. Persistence was 79.4% at 3 months for FTC/TDF, 70.2% at 6 months, and declined to 57.4% at 12 months (Fig. 1B). Persistence and PDC data for individuals using FTC/TAF were not included in the analysis because of the limited follow-up time.
First evidence of HIV-1 and risk factor analysis
Using the primary definition of first evidence of HIV-1 (non-PrEP ART initiation and presence of ICD-9/10-CM HIV-1 diagnosis) in individuals with ≥6 months of continuous preindex enrollment, first evidence of HIV-1 at any time during follow-up was generally low for individuals using FTC/TDF (0.6%; Table 3). First evidence of HIV-1 was 0.3% at 3 months after PrEP initiation and increased to 0.5% at 12 months. In a post hoc analysis for which the 6-month continuous preindex enrollment requirement was removed, first evidence of HIV-1 increased numerically (3 months, 2.2%; 12 months, 2.5%). In a sensitivity analysis defining first evidence of HIV-1 by non-PrEP ART initiation only, a negligible change in first evidence of HIV-1 was observed among individuals with (3 months, 0.4%; 12 months, 0.7%) and without (3 months, 2.4%; 12 months, 2.9%) ≥6 months of continuous preindex enrollment. When all analyses were conducted post hoc, excluding individuals with first evidence of HIV-1 within 30 days after PrEP initiation, first evidence of HIV-1 numerically decreased, especially with the 6-month continuous preindex enrollment requirement removed. Using either definition of first evidence of HIV-1 in individuals with ≥6 months of continuous preindex enrollment, first evidence of HIV-1 at any time during follow-up was 0.2% for people using FTC/TAF; however, the shorter follow-up time limited this analysis.
Descriptions of First Evidence of HIV-1 While Using FTC/TDF
First evidence of HIV-1 by non-PrEP ART initiation and ICD-9/10-CM HIV-1 diagnosis was defined as the earliest date between an ICD-9/10-CM HIV-1 diagnosis and ≥1 dispensing of a non-PrEP ART, and individuals required both a diagnosis and dispensing to be considered as having HIV-1.
Evaluated among the subgroup with first evidence of HIV-1 during follow-up.
First evidence of HIV-1 by non-PrEP ART initiation was defined as ≥2 dispensings of a non-PrEP ART.
ART, antiretroviral therapy; FTC, emtricitabine; ICD-10-CM, International Classification of Diseases, 10th Revision, Clinical Modification; PrEP, pre-exposure prophylaxis; TDF, tenofovir disoproxil fumarate.
Using the primary definition of first evidence of HIV-1 in individuals with ≥6 months of continuous preindex enrollment, mean (SD) time to first evidence of HIV-1 from the index date for people using FTC/TDF was 235 (306) days (Fig. 2A and Table 3), and time from the latest PrEP dispensing was 149 (240) days. Removal of the 6-month continuous preindex enrollment requirement decreased mean (SD) time to first evidence of HIV-1 from the index date to 69 (184) days (Fig. 2B and Table 3). For people using FTC/TDF with ≥6 months of continuous preindex enrollment, 60.3% had PrEP on hand at the time the primary definition was met, and 9.2% had ≥1 PrEP break before meeting the primary definition. Removal of the 6-month continuous preindex enrollment requirement increased the proportion who had PrEP on hand to 88.9% and decreased the proportion who had ≥1 PrEP break to 3.1%. The most common first-line ART regimens dispensed for these individuals were bictegravir/FTC/TAF (24.8%) and integrase strand transfer inhibitor—based multitablet regimens (35.5%).
Proportion of individuals using PrEP after the index date by month for the first 36 months and (inset) by day within the first 30 days who met the primary definition of first evidence of HIV-1 (non-PrEP ART initiation and HIV-1 diagnosis) with
High-risk sexual behavior, syphilis, gonorrhea, and other nongonococcal urethritis were the risk factors most strongly associated with first evidence of HIV-1 (Fig. 3). Notably, individuals using PrEP enrolled in a health maintenance organization (HMO) had 2 times the odds of first evidence of HIV-1 than individuals enrolled in a preferred provider organization (PPO). Further, individuals in the Midwest, Northeast, and West had ∼40% lower odds of first evidence of HIV-1 than individuals in the South.
Odds ratio estimates (95% CI) for risk factors of first evidence of HIV-1 (non-PrEP ART initiation and HIV-1 diagnosis). Analysis included individuals using FTC/TDF or FTC/TAF. Significant estimates (p < 0.05) are indicated by bold text. ART, antiretroviral therapy; CI, confidence interval; ED, emergency department; FTC, emtricitabine; HMO, health maintenance organization; OP, outpatient; POS, point-of-service; PPO, preferred provider organization; TAF, tenofovir alafenamide; TDF, tenofovir disoproxil fumarate.
Discussion
This retrospective claims study characterized individuals using daily oral PrEP regimens, FTC/TDF or FTC/TAF, based on their demographics and clinical characteristics, usage patterns, adherence, persistence, first evidence of HIV-1, and associated risk factors. Characteristics of individuals at PrEP initiation were generally similar between FTC/TDF and FTC/TAF, and switching from FTC/TDF to FTC/TAF was common. However, breaks in PrEP use were observed in 11.1% of people using FTC/TDF and could last for several months. Although continued STI diagnoses after PrEP initiation were observed, first evidence of HIV-1 was infrequent, with higher frequency observed in individuals without continuous insurance enrollment before PrEP initiation.
Adherence to and persistence with daily oral PrEP options substantially impact effectiveness, though measuring adherence is challenging as no gold-standard measurement exists and there are disadvantages with each measure (reviewed in Berg et al.). 35 –37 Clinical trials completed after 2013 have reported generally high adherence to FTC/TDF; however, trials that rely on self-reported data may overestimate adherence levels. 37 In the DISCOVER (NCT02842086) study, 96–98% of participants reported taking either FTC/TDF or FTC/TAF >80% of the time across study visits, and median pill count adherence was 98% in both groups. 18 In contrast, dried blood spot analyses from a subset of participants revealed lower adherence, with 84–96% having tenofovir diphosphate concentrations consistent with taking four or more pills weekly (i.e., ≥57% of the time). Nevertheless, HIV-1 incidence was low in both groups (0.16–0.34 per 100 person-years). In the HPTN 083 and HPTN 084 studies, 39 of 3187 and 36 of 1586 participants using FTC/TDF acquired HIV-1 for an incidence of 1.22 and 1.85 per 100 person-years, respectively. 22,23 Although adherence was imperfect (74.2% and 41.9% of 083 and 084 participants, respectively, had tenofovir plasma concentrations consistent with daily FTC/TDF dosing), a majority (72/75) of individuals who acquired HIV-1 had suboptimal or poor adherence. 22,23,38
Despite the known benefits of PrEP, real-world adherence and persistence are typically suboptimal and further complicated by factors like cost, access, side effects, stigma, and perceived need for HIV-1 prevention. 39,40 Indeed, with regard to cost and access, a 2024 analysis of data from >58,000 people newly prescribed PrEP in the United States found that higher out-of-pocket costs were associated with an increased rate of PrEP prescription abandonment. 41 This trend is consistent with results from the present analysis, in which PDC and persistence generally declined over the 12 months after PrEP initiation, and with results from similar studies conducted in Belgium and France that indicated high rates of PrEP discontinuation in the first 6 months of use. 42,43 Additionally, overall adherence and persistence levels in the present analysis were lower than those reported in DISCOVER and HPTN 083 evaluating the same oral PrEP regimens. 18,23 Most people using PrEP with first evidence of HIV-1 in the present analysis had PrEP on hand at the time they met the definition, suggesting challenges unexplained by access; therefore, more must be done to properly screen for HIV-1 before PrEP initiation and prevent HIV-1 despite access to oral PrEP. Waning adherence and persistence among individuals using PrEP in real-world settings are important considerations for efforts aimed at improving PrEP uptake among individuals with the highest need for HIV-1 prevention, including the introduction of a long-acting PrEP option. 23,44 In HPTN 083 and HPTN 084, CAB-LA administered every 8 weeks was superior to once-daily oral FTC/TDF in preventing HIV-1 acquisition (083: hazard ratio [95% CI], 0.34 [0.18–0.62]; 084: 0.12 [0.05–0.31]), and adherence to injection visits was high (92–93% person-years covered by injections), indicating long-acting PrEP injections may be a beneficial option for individuals struggling with daily oral pill adherence. 22,23
In the current analysis, the low prevalence of the most common comorbidities captured (e.g., hypertension, obesity, diabetes) suggests that people using PrEP were in relatively good health. While research is increasing, there is still a paucity of available data on comorbidities in people using PrEP. The real-world Kaiser Permanente study of 972 individuals who initiated FTC/TDF for PrEP reported proportions of participants with hypertension (10.8%) and diabetes (3.0%) similar to this study. 24 Consistent with the current analysis, the Kaiser Permanente study and a similar study conducted in Belgium also found that STI diagnoses were common during follow-up and that HIV-1 incidence was low. 24,43 Further, a longitudinal study of 199 men who have sex with men who initiated PrEP in Amsterdam found that the proportion with anal STI diagnoses increased during the first 2 years after PrEP initiation compared with the 4 years before PrEP initiation and with a matched control cohort. 45 Indeed, syphilis and gonorrhea were determined to be significant risk factors for first evidence of HIV-1 in the current study; people using PrEP with these STI diagnoses had ∼3 times higher odds of first evidence of HIV-1. Altogether, despite the relatively good health of individuals using PrEP, the commonality of STI diagnoses suggests a continued need for HIV-1 prevention during follow-up regardless of the low rates of first evidence of HIV-1.
A pronounced increase in first evidence of HIV-1 was observed when the requirement of ≥6 months of continuous preindex enrollment was removed, suggesting that individuals with the greatest need of HIV-1 prevention may be engaging with health care services only when they perceive their likelihood of HIV-1 exposure as increased, which is supported by the short time to first evidence of HIV-1 observed in this group. Alternatively, some individuals with <6 months of continuous preindex enrollment may have had HIV-1 before the index date. Additionally, the higher odds of first evidence of HIV-1 seen with individuals using PrEP enrolled in an HMO versus PPO suggest socioeconomic factors may curb health care service engagement given that HMOs tend to be more affordable but also usually only provide coverage for in-network providers, limiting access to care. 46 This is consistent with results from a matched case–control study nested among adult men at high risk of HIV-1 acquisition in France that used data from the French National Health Data System; PrEP effectiveness was found to be significantly lower among those receiving vs not receiving complementary universal health insurance (p = 0.0096). 47 Further, when individuals with first evidence of HIV-1 within 30 days after PrEP initiation were excluded, first evidence of HIV-1 decreased and the time to first evidence of HIV-1 increased. Altogether, observing individuals with first evidence of HIV-1 quickly after starting PrEP reinforces the importance of thorough screening before initiation and comprehensive provider support early during use.
This study had limitations. Pharmacy dispensings are not a guarantee that the medication was taken as prescribed; therefore, PDC may overestimate adherence, coding inaccuracies may exist in administrative claims data, and laboratory values were not available to confirm HIV-1 diagnosis or time of acquisition. Information on postexposure prophylaxis was not directly available in the data, though very few individuals used non-PrEP ART a single time, suggesting that postexposure prophylaxis use was uncommon. This study focused on commercially insured individuals using PrEP in the United States; therefore, findings may not be generalizable to populations with no insurance or public insurance (e.g., Medicare, Medicaid). Further, this study was unable to focus on priority populations known to be disproportionately affected by HIV-1 (e.g., Black women and transgender people) 48 due to the unavailability of data. Last, because of the short time between the approval of FTC/TAF for PrEP (October 3, 2019) and the end of data availability for this study (March 31, 2020), individuals who initiated FTC/TAF had short follow-up time, limiting the interpretability of PrEP usage patterns for this population.
Using a large US claims database, this retrospective study described the real-world characteristics of people who initiated FTC/TDF or FTC/TAF for once-daily oral PrEP. Overall, individuals using PrEP were generally healthy and commonly switched from FTC/TDF to FTC/TAF. Although first evidence of HIV-1 was infrequent in this real-world cohort, adherence and persistence were lower than reported in recent clinical trials, 18,23 which could impact the long-term effectiveness of these daily oral PrEP options.
Footnotes
Acknowledgments
Editorial assistance was provided under the direction of the authors by Julia Grzymkowski, PhD, and Lauren Bragg, ELS, MedThink SciCom, and was funded by ViiV Healthcare.
Authors’ Contributions
A.O.: Conceptualization, formal analysis, funding acquisition, methodology, visualization, and writing—reviewing and editing. G.G.: Conceptualization, formal analysis, methodology, and writing—reviewing and editing. A.A.M.: Conceptualization, methodology, project administration, resources, supervision, visualization, and writing—reviewing and editing. F.L.: Conceptualization, formal analysis, methodology, project administration, and writing—reviewing and editing. S.D.M.: Data curation, formal analysis, methodology, software, and writing—reviewing and editing. A.H.: Data curation, formal analysis, methodology, software, and writing—reviewing and editing. H.S.: Conceptualization, writing—reviewing and editing. M.S.D.: Conceptualization, formal analysis, methodology, project administration, and writing—reviewing and editing.
Author Disclosure Statement
A.O., A.A.M., and H.S. are employees of
Funding Information
This study was funded by