Defining success with HIV pre-exposure prophylaxis

Introduction To date, six clinical trials have explored the efficacy of pre-exposure prophylaxis (PrEP) against HIV infection using oral tenofovir and/or emtricitabine/tenofovir 1–6. Efficacy ranged from 0 to 75% 7, the variation of which can largely be explained by differences in PrEP adherence 8. Efficacy data directly correlate with the objective adherence measures in those studies, and efficacy estimates were between 90 and 100% when adherence was consistently high 2,9,10. PrEP is clearly efficacious when taken as prescribed. Pre-exposure prophylaxis adherence is, however, complex and should be understood within the context of variable risk for HIV infection and use of other HIV-prevention methods. Different levels of adherence may be needed in different populations to achieve HIV prevention, and the optimal methods for achieving the necessary adherence for both individual and public health benefits are unknown. In moving beyond clinical trials toward demonstration projects (i.e. studies of PrEP delivery methods) and broader implementation in clinical care, guidance for PrEP use must consider these questions to determine if PrEP-based HIV-prevention programs are successful at both the individual and public health level. In this article, we propose a new paradigm for understanding and measuring PrEP adherence, termed prevention-effective adherence, which incorporates dynamic HIV-acquisition risk behaviors and the use of HIV-alternative prevention strategies. We focus on sexual transmission of HIV, as the role of PrEP in preventing HIV transmission via injection drug use is less well understood. We discuss the need for daily PrEP use only during periods of risk for HIV exposure, describe key issues for measuring and understanding relevant behaviors, review lessons from another health-prevention field (i.e. family planning), and provide guidance for prevention-effective PrEP use. Moreover, we challenge emerging calls for sustained, near-perfect PrEP adherence, regardless of risk exposure, and offer a more practical and public health-focused vision for this prevention intervention. Understanding pre-exposure prophylaxis adherence Sustained lifelong high adherence: the paradigm for antiretroviral therapy The adherence message for antiretroviral therapy (ART) of HIV infection is simple: viral suppression and improved health require indefinite, consistently high adherence following ART initiation 11. Whereas early regimens required at least 95% adherence 12, contemporary regimens reliably suppress viral replication at somewhat lower adherence levels (e.g. 80%) 13,14, depending on adherence patterns and prior duration of viral suppression 15,16. Nevertheless, lifelong sustained adherence is required to stop disease progression. While scheduled intermittent dosing of ART (e.g. planned drug holidays) was suggested to ease the adherence burden, studies of prescribed intermittent ART and unplanned missed doses showed increased rates of viral rebound and worse clinical outcomes 17,18. Sustained high adherence for the duration of pre-exposure prophylaxis clinical trials High levels of sustained adherence are also required for the evaluation of efficacy and safety of potential PrEP agents. Thus, in placebo-controlled, randomized clinical trials, one could argue for selecting participants who are motivated to adhere, as well as providing intense adherence support to maintain consistently high adherence regardless of dynamic risk behaviors. This scenario, however, does not generalize to open-label use and uptake of PrEP that is now known to be effective. Prevention-effective adherence: a novel paradigm Several differences between ART and PrEP (as used outside of clinical trials) are noted in Table 1, and the type of adherence needed for each setting is presented in Fig. 1. The message for PrEP adherence is not simply 100% adherence for life or the duration of a research study. Because an individual's risk for HIV acquisition changes over time and alternative prevention strategies may be used, the indication for PrEP also changes over time. PrEP use, for example, may not be indicated if sexual activity is restricted to a monogamous relationship with a known HIV-negative partner without other risk exposures, or if another effective HIV-prevention tool (e.g. condoms) is consistently used. Perfect PrEP adherence in the absence of risk confers cost, side effects, and toxicity, but no benefit. Consistently high adherence during periods of exposure, however, is critical when PrEP is being used for effective HIV prevention.Table 1: Differences between antiretroviral therapy and pre-exposure prophylaxis as used outside of clinical trials.Fig. 1: Adherence success by adherence paradigm.The paradigm of consistently high adherence for all individuals applies to ART and PrEP in clinical trials (panel a). The prevention-effective paradigm applies to PrEP in demonstration projects and wider implementation (panel b). ART, antiretroviral therapy; PrEP, pre-exposure prophylaxis.Prevention-effective PrEP adherence means the use of PrEP only during periods of risk exposure such that it leads to effective protection against HIV acquisition. Failure to understand this concept may result in missed opportunities for HIV prevention. For example, individuals who struggle to use PrEP at certain times in their lives (e.g. young women in high HIV-prevalence areas) may be able to use it at others if they are guided through the process of understanding risk and choice of effective prevention options. Moreover, this concept can lead to efficient PrEP use that will limit adverse events and costs, and potentially widen availability in settings of resource scarcity. The paradigm presented here is applicable to both PrEP demonstration project and wider implementation. Unlike in clinical trials, target participants in demonstration projects and implementation are those who are interested in PrEP when they are at risk for HIV exposure, but actual PrEP uptake and time-on-PrEP will vary considerably as needs change over time. Guidelines suggest PrEP should be used in populations at high risk for HIV acquisition 19,20. PrEP is unlikely to be a preferred or cost-effective HIV-prevention tool for individuals with rare or unpredictable HIV exposure 21, for whom other prevention options may be more appropriate e.g. postexposure prophylaxis (PEP) for victims of rape, or condom use for individuals with infrequent sexual encounters who can use condoms effectively. Measuring prevention-effective combination adherence Prevention-effective adherence requires measurement of HIV-risk exposure, use of PrEP, and use of other effective HIV-prevention tools. These factors are then combined in the calculation of adherence. Risk exposure The first step in measuring prevention-effective adherence is the assessment of risk for HIV exposure. Measurement of risk exposure is challenging and largely limited to self-report of behavior or perceived risk, which may be inaccurate due to recall and social desirability biases 22. Accuracy, however, may increase via short recall periods 23,24 and can be measured concurrently with medication adherence (e.g. 24-h recall by SMS surveys) 25. Biomarkers of sexual intercourse (e.g. prostate-specific antigen in vaginal secretions) can provide objective evidence of sex; however, they are difficult to obtain and impractical for clinical practice 26,27. They also do not comment on male sexual activity. Surrogate markers (e.g. incidence of other sexually transmitted infections and pregnancy) may provide some indication of potential HIV exposure, but are generally informative at the population, not individual, level. Risk assessments should aim to identify periods of sustained risk, for which PrEP might be a preferred HIV-prevention option. Ongoing studies (i.e. IPERGAY, HPTN067, and HPTN 069) involve extensive data collection on sexual behavior and risk for HIV exposure, and may provide further insights into optimal measurement strategies. Use of pre-exposure prophylaxis The next step in assessing prevention-effective adherence is to quantify PrEP use. No gold standard exists for measuring medication-taking behavior; all have strengths and weaknesses (see Table 2) 28,29. Because all adherence measures are imperfect, a context-appropriate set of more than one measure is needed to understand adherence behavior. Importantly, self-report did not correlate with drug levels in several PrEP clinical trials and was largely uninformative due to over-reporting of adherence 4,6. As noted above for self-reported risk behavior, self-reported adherence may be improved through SMS surveys of short recall periods and/or questions that have that been validated with objective PrEP adherence measures. The accuracy of self-report may also improve in the absence of consequences for reporting socially undesired behavior (e.g. burdensome intervention components when condomless sex or nonadherence is reported). Importantly, reporting of nonadherence is more likely to be accurate than reporting of adherence 30. Pharmacy refill records provide an objective approach that has performed well in predicting viral suppression in ART programs 31 and may be feasible for routine clinical use with PrEP. While cost limits electronic adherence monitoring to research studies, it is the only method that provides dose-by-dose measurement for assessing adherence patterns, which are critical for understanding prevention-effective adherence 32. Electronic monitoring in at least a subset of participants in demonstration projects should therefore be considered. Wireless electronic monitoring is also available and can provide data in real time, reduce the risk of data loss, and decrease the human resources needed for data collection in geographically dispersed settings 33. Drug levels were critical in providing objective documentation of drug ingestion in the PrEP clinical trials; however, they are also impractical for routine use. Stored samples for selective testing offer a less expensive approach that may be appropriate for demonstration projects.Table 2: Strengths and weaknesses of adherence measurement tools.Demonstration projects that involve multiple PrEP adherence measurements will help determine the best measures for prevention-effective PrEP adherence, which can then be compared to assess the most informative, accurate, and affordable measures for use in the less resourced context of implementation. For example, electronic monitoring or drug levels may be used to determine which types of self-report questions best identify individuals with adherence challenges. Prevention-effective adherence also requires measurement of PrEP initiation and discontinuation. An objective approach is to determine when medication was picked up via pharmacy refill records. The timing of discontinuation, however, is difficult to know, as individuals may initiate and discontinue PrEP multiple times during a single refill period. Electronic monitors can provide more precise patterns of PrEP initiation and discontinuation. Self-report may be considered as a complementary tool to further define either of these objective measurements. For example, individuals can be asked whether periods of nonadherence were intentional or unintentional. The above-noted limitations of self-reported data, however, should be kept in mind. A key distinction should be drawn between periodic and intermittent PrEP use. ‘Periodic use’ involves voluntary starting and stopping of daily PrEP use depending on HIV-prevention needs and choices (e.g. while trying to conceive a child within a serodiscordant couple 34). Periodic use is based on the once-daily dosing recommendation, and periods of time can be considered as on or off PrEP for as long as PrEP is part of an overall HIV-prevention approach. In contrast, ‘intermittent PrEP’ refers to a prescription of less-than-daily PrEP, as was used in the International AIDS Vaccine Initiative (IAVI) pilot randomized controlled trial 35,36 and is currently being evaluated in IPERGAY and HPTN067. Preliminary data from IPERGAY are promising, although a full understanding of the data is pending 37. Importantly, adherence to intermittent doses was lower than to fixed doses in the IAVI trial. Use of other HIV-prevention tools While measuring use of the other HIV-prevention tools is complex and beyond the scope of this article, it is an important concept in defining prevention-effective adherence. Briefly, condom use and behavior modification (e.g. partner reduction, knowledge of HIV status) typically rely on self-report. Medical male circumcision does not require ongoing adherence measurement; however, a 50% reduction in risk of HIV acquisition is arguably insufficient to recommend it as a sole prevention strategy. ART in the HIV-infected partner of a stable serodiscordant couple can prevent secondary transmission, and adherence may be measured as described above for PrEP, as well as with HIV RNA levels. Other forms of PrEP beyond tablets (e.g. vaginal rings, gels, depo injections) may become available and individuals may choose different formulations at different times. Ongoing assessments of when and how various HIV prevention tools are used are therefore needed. Execution and persistence of prevention-effective pre-exposure prophylaxis adherence The above three factors are all required to calculate two types of adherence: execution and persistence. First, execution of prevention-effective PrEP adherence refers to adherence during the time an individual relies upon PrEP for protection against HIV acquisition (Fig. 2). It depends on periods of HIV risk as defined by behavior and/or use of other effective prevention tools, as well as initiation/discontinuation of PrEP. It is not simply lifelong as with ART or for the duration of a clinical trial. Execution of prevention-effective adherence should include the doses needed to achieve effective drug concentrations during PrEP initiation (discussed below), but should not include periods without risk exposure.Fig. 2: Prevention-effective execution and persistence of adherence.Execution refers to adherence during the time that an individual is at risk for HIV acquisition and is intending to rely upon it for protection. Persistence describes the duration of PrEP use during periods of HIV risk and should refer the absolute time of use. As an example, for individuals taking PrEP during periods of HIV risk, the striped shape represents 100% executed adherence for the first 3 months of PrEP use, followed by 50% adherence for the next 3 months of use, with a persistence of 100%. The dotted shape represents 80% executed adherence with a persistence of 3 months. PrEP, pre-exposure prophylaxis.For example, a sex worker may engage in unprotected sex for 6 months in an urban setting, but then return home to a rural area for the following 6 months where she does not engage in any sex. If she took PrEP daily for all but the final 2 weeks of the 6 months spent doing sex work, her executed adherence would be 92%. If she stopped sex work a month early, her executed adherence would be 100%. Conversely, if she continued daily PrEP for the full year without sex work in the final six months, her executed prevention-effective adherence would be 200% (thus, a potential misallocation of resources and unnecessary risk for side effects). Both PrEP use and the number of expected dosing events should be censored when HIV-transmission risk approaches zero. Persistence of prevention-effective PrEP adherence describes the duration of PrEP use before an individual stops it (either temporarily or permanently). End of use should be defined by self-reported intentional cessation or a clear break in use (e.g. 28 days of non-use), as consistent with the concept of periodic PrEP use. Missing a few doses because of temporary lapses still constitutes persistence, albeit with imperfectly executed adherence. With ART, persistence is desired for life, so no denominator is needed. With prevention-effective PrEP adherence, persistence should be considered over the duration of risk based on behavior and/or use alternative HIV-prevention tools. Importantly, the length of the denominator should be noted to reflect the extent of risk. For example, an individual using PrEP for the first 5 of 6 months with HIV risk behavior and no other effective prevention tools has a persistence of 83% over 6 months. Another individual using PrEP for the first 10 of 12 months with risk and no alternative effective prevention tools also has a persistence of 83% over 12 months. Although the percentage of persistence is the same, the overall risk for the latter individual is higher, given the potential for exposure during 1 month more than the former individual. This time-dependent nature of risk is important in considering the overall effectiveness of a given level of persistence. Additionally, it is important to note that someone may have high persistence and still not achieve prevention-effective adherence, if the execution of adherence during that time is inadequate. HIV testing The effectiveness aspect of prevention-effective adherence requires HIV testing. Current guidelines recommend testing prior to initiation of PrEP and periodic testing during use although the optimal for periodic testing is not Both and approaches are being explored in demonstration projects the of evidence on the optimal for adherence to HIV testing at a given should not PrEP use. testing prior to of PrEP a of is critical to drug due to ongoing use during of risk could result in exposure without PrEP use, such testing important when prevention-effective adherence. of pre-exposure prophylaxis adherence by of infection PrEP adherence in the context of risk exposure, two factors must be considered to if a given PrEP adherence will provide the drug and drug level at the time of exposure. These factors with the of exposure and are critical for how PrEP is drug Efficacy in most oral PrEP trials has been with concentrations of tenofovir estimates that daily oral doses will result in drug concentrations in data are available to when concentrations are in vaginal and but data generally concentrations in providing a for days of PrEP use prior to achieving protection. The of missed doses on efficacy may have more when has not been (i.e. is likely less for missed that PrEP use may be these factors must be considered PrEP use, not at Drug level at the time of exposure To provide the adherence must result in a drug level at the time of risk exposure that is to prevent HIV In studies based on two doses on may provide doses and doses efficacy tenofovir concentrations are in compared to vaginal vaginal may require more oral dosing to maintain drug levels in a Additionally, the precise timing as to when risk from an exposure has is because it is how long HIV to be from the guidance can be drawn from for which 28 days of continued use exposure PrEP from in that early replication is by PrEP from ongoing use, and 28 days of postexposure dosing without further risk may be challenging for some 28 days of continued dosing time for alternative HIV-prevention tools. research in this area is needed. in achieving prevention-effective adherence for individual and public health benefits The paradigm of prevention-effective adherence will be new for as well as and In efficacy trials, individuals are typically given one or more HIV-prevention tools and are expected to consistently to all prevention strategies regardless of need or In demonstration projects and a of prevention options is more likely to reflect how individuals will achieve HIV prevention. guidance is needed in the of HIV-prevention options and from family Although and HIV infection are clearly different both via sex and multiple strategies for lessons from family may therefore be for PrEP have that women options A may use oral for some time, but stop when she no has a sexual may choose depo or some other method if her partner does not to reliably use several she may to condoms as an effective alternative for with can and change their choice of to their and A approach could be used for HIV prevention, and measurement of adherence to each method will then determine if the individual has prevention-effective adherence. or are available for PrEP, options will still for some at certain times in their the still HIV infections will still likely the availability of PrEP and other prevention tools. Nevertheless, more HIV will be with these tools than without as family programs in real settings have unplanned in of risk assessment and use of research on understanding and prevention-effective adherence is therefore needed to for prevention-effective pre-exposure prophylaxis adherence To understand prevention-effective PrEP adherence, individuals will need guidance in prevention options. include their and behaviors, as well as the timing of potential adherence for PrEP is in Table for PrEP use and adherence should be through other such as and other and may need the needs of each for PrEP to prevention-effective daily PrEP may be as a while an HIV-infected partner and on ART, as is being in the use of may suggest the need for the more of PrEP. If depo of PrEP become daily PrEP may be a to if side Adherence to will require at infrequent events (e.g. depo which has been to be challenging for Both execution and persistence of adherence can help individuals determine if PrEP will be an effective prevention tool for who between PrEP and for example, are unlikely to achieve prevention-effective adherence with PrEP and should consider on condoms or another strategy. Adherence should also the for daily doses to achieve levels of as well as risk for infection an exposure, as prevention-effective adherence to define success for HIV-prevention programs should not on consistently high PrEP adherence for life, as is the standard for ART programs and clinical trials the adherence success may be defined as providing to multiple prevention tools for the most at risk for HIV with high adherence to at least one tool at any given time. This approach may be as prevention To extent is the To extent does PrEP increase that and at the using PrEP were not the individuals who were using condoms at PrEP a prevention In adherence is critical for PrEP than PrEP as a daily taken however, it should be through the of prevention-effective adherence, which into multiple prevention strategies and dynamic risk for HIV acquisition. Measurement of PrEP adherence must include an individual's choice of tools, the timing of those and the of risk behaviors. Although further research is needed on the best measures for both PrEP use and HIV risk, the concept of prevention-effective adherence should be considered as programs PrEP for HIV prevention. of PrEP does not it may be a and efficient choice for a given individual. HIV-prevention programs may be able to from other prevention such as family so they can the and support needed to individuals through a of HIV prevention options. The would to the and and for a of the to discuss key issues in PrEP adherence in and of and the for AIDS of to the of the in this the first of the which all other of are no of