Abstract
Stringent eligibility criteria, drug costs and antiretroviral toxicities are challenges in delivering HIV non-occupational post-exposure prophylaxis (nPEP). We reviewed patients’ nPEP eligibility and clinical outcomes at St Michael's Hospital, Toronto, Canada to identify opportunities for improvement. Of 241 patients, 59%, 36% and 6% presented for high- (receptive anal/vaginal, blood), medium- (insertive anal/vaginal) and low-risk (oral) sexual exposures, respectively, and nearly all (93%) presented within 72 hours. Of 205 patients given nPEP, 20 were known to have discontinued nPEP prematurely: three due to costs but none due to toxicities. Two HIV seroconversions occurred in patients with suspected ongoing potential exposures. Five asymptomatic syphilis diagnoses were made among 71 tested. Only 39% and 19% of nPEP patients returned to our institution for follow-up at 3–4 and six months, respectively. Our findings underscore the feasibility and importance of nPEP programmes to HIV and sexually transmitted infection control, while identifying opportunities for improvement.
Keywords
INTRODUCTION
Post-exposure prophylaxis (PEP) using antiretroviral medications (ARVs) is becoming the standard of care in industrialized countries for minimizing the risk of HIV infection after non-occupational exposures such as unprotected sexual activity and injection drug use. Guidelines for non-occupational PEP (nPEP) have been published for several years. 1 St Michael's Hospital (SMH) is an academic, tertiary care hospital in the downtown core of Toronto, Canada, and has had institution-specific protocols based on such guidelines for both occupational and non-occupational PEP since the late 1990s and 2008, respectively.
The SMH nPEP protocol is as follows. Patients presenting to the SMH Emergency Department (ED) with known or potential HIV exposures are triaged to be seen within 15 minutes and assessed using a standardized clinical pathway. Initial assessment includes baseline testing for HIV and hepatitis B and C, and depending on the level of HIV transmission risk, a five-day starter pack of ARVs may be provided. Tenofovir/emtricitabine (TDF/FTC) is the basic regimen for low-risk exposures, with lopinavir/ritonavir (LPV/r) added for medium- to high-risk exposures. Follow-up is booked within five days in the Infectious Diseases (ID) Clinic for a more detailed assessment regarding risk, nPEP regimen/continuation and related issues such as sexually transmitted infections (STIs). Patients are asked to return for follow-up at six weeks, three months and six months for further testing and counselling.
Challenges to nPEP delivery are numerous, including the stringent criteria used to identify eligible patients, high medication costs and drug tolerability. We sought to review our experience in providing this important clinical and public health intervention, and to identify opportunities to improve service provision, through a retrospective chart review of nPEP cases at SMH. Specific objectives were to summarize patient eligibility for nPEP (number testing HIV-positive at baseline, level of risk and time elapsed since exposure) and clinical outcomes (infection rates, nPEP regimen completion, adverse events and adherence with follow-up).
METHODS
Study procedures
We reviewed the charts of all patients presenting to SMH between 1 January 2008 and 31 December 2010 with a recent potential non-occupational exposure to HIV. Study personnel were trained to conduct chart reviews for this study, reviewed electronically scanned emergency room charts, laboratory results and transcribed clinic letters from the hospital electronic record, and extracted data using standardized case report forms into a Microsoft Excel database.
When calculating the times to ED presentation and to first dose, a conservative approach was taken whenever exact times were not documented, by assuming that exposures occurred at 12:00 am on the day reported, and that first doses were taken at midnight the day nPEP was provided. This strategy errs on the side of underestimating time-based nPEP eligibility. Treatment-emergent laboratory abnormalities were defined as abnormal haematology/biochemistry results not present at baseline but observed after starting nPEP, and grading was performed using the Division of AIDS Table for Grading the Severity of Adverse Events. 2
Analyses were descriptive in nature, involving measures of central tendency and dispersion for continuous variables and frequencies/proportions for categorical variables. No sample size calculation was performed because this work was primarily exploratory in nature.
RESULTS
Participant characteristics and nPEP eligibility
Participant characteristics
IQR = interquartile range; SMH = St Michael's Hospital; ED = emergency department; MSM = men who have sex with men; ARV = antiretroviral; ED = emergency department; nPEP = non-occupational post-exposure prophylaxis
*Total number = 244 as some patients were exposed to more than one source
†Includes two reactive HIV tests at baseline
‡Includes five reactive syphilis tests at baseline
The vast majority of patients arrived in the ED promptly after their exposure, allowing at least 81% to receive their first dose of ARVs within the 72-hour window for providing nPEP (timing could not be accurately determined for 15.6%). Fifteen patients (6.2%) presented beyond 72 hours, and seven (3.4%) appeared to take their first dose outside the 72-hour period. Of note, two additional patients (1.0%) took their first dose before arriving in the ED, one obtaining ARVs from a friend and the other from another hospital. Another 23 (9.5%) were deemed inappropriate for nPEP and 13 (5.4%) declined.
For each patient, the highest risk type of sexual exposure reported was recorded. The most common activities were: receptive anal intercourse (43.6%) and insertive anal intercourse (24.9%). Only 5.8% of patients presented for low-risk sexual exposures (Figure 1).
Risk levels of patients presenting to ED for nPEP at St Michael's Hospital, Toronto, Canada, 1 January 2008 to 31 December 2010
Of 244 potential exposures (3 cases involved patients simultaneously exposed to 2 different sources), 102 (41.8%) involved a known HIV-infected partner; the remainder involved partners of unknown HIV status. Only a minority of HIV-infected partners (6.1% of all exposures) were known to be on ARVs.
Of all 241 patients, 187 (77.6%), 184 (76.3%) and 184 (76.3%) were tested at baseline for HIV, hepatitis B and hepatitis C, respectively; another 18 patients had been tested at a nearby sexual health clinic before being sent to our ED. Two patients were HIV-infected at baseline; no hepatitis infections were diagnosed. Preventive interventions were provided through 37 hepatitis B and three hepatitis A vaccinations. Testing rates for other STIs were low, but revealed five syphilis infections (7.0% of those tested); all were treated.
Drug regimens
nPEP regimens given at St Michael's Hospital, Toronto, Canada, 1 January 2008 to 31 December 2010, by HIV risk level*
nPEP = non-occupational post-exposure prophylaxis
*Numbers represent numbers of patients
nPEP completion rates and adverse effects
Of the 205 patients given nPEP, 71 (34.6%) were known to have completed the full 28-day course, as determined by physician documentation of patient self-report at the time of their next follow-up visit. Another 20 (9.8%) were known to have stopped their medications early, after a median of five days. Reasons for discontinuing nPEP early included patient preference (n = 5), cost (n = 3), low risk as determined by SMH staff (n = 3), the source patient testing negative for HIV (n = 4) and misinterpretation of instructions (n = 4). The final patient who stopped early did so upon testing positive for HIV at baseline. The only other patient who tested positive for HIV at baseline transitioned directly onto therapy. Rates of nPEP completion were unknown for 114 (55.6%) patients given nPEP because many patients did not follow up at SMH.
Although information on drug tolerability and toxicity was documented inconsistently on patient charts, no patients were noted to discontinue nPEP due to side-effects. Among the 55 patients with reported adverse effects, the most common was diarrhoea (n = 20), followed by non-specific gastrointestinal upset (n = 14) and nausea (n = 13).
Among the 205 patients given nPEP, grade 1 abnormalities in electrolytes, liver transaminases and bilirubin were noted in 40, 19 and 20 individuals, respectively, at baseline, and abnormalities beyond grade 1 were noted in three patients (grade 2 elevations in alanine transaminase [ALT] and/or aspartate transaminase), but did not affect clinical decision-making. Treatment-emergent laboratory abnormalities were noted in only 7.9% of the nPEP patients, but testing was performed inconsistently, in only 89%, 19% and 7% of patients at baseline, two and four weeks, respectively. Follow-up lab testing information was only available for patients who returned to SMH for follow-up visits.
Twenty patients switched from the nPEP regimen originally provided in the ED, including 13 patients switching from three to two drugs by dropping the LPV/r component, and seven switching from two to three drugs (adding LPV/r in 5 patients, and switching to zidovudine/lamivudine/LPV/r or abacavir/lamivudine/efavirenz in 2 others). Reasons for switching included HIV risk re-assessment and potential drug interactions with concomitant medications.
Drug coverage information was available for only 45 (18.7%) patients, who were covered under methods such as their employment drug plan, compassionate supplies of ARVs, the Ontario Drug Benefit or Interim Federal Health Programs.
Follow-up
Although 130 nPEP patients attended the ID clinic within five days of presenting to ED as per our institutional protocol and another 46 (22.4%) were referred elsewhere for follow-up (e.g. family physicians), 29 (14.1%) patients attended the ID clinic after the five-day window and 36 (17.6%) patients failed to attend their scheduled clinic visit altogether. Significant proportions of subsequent visits were also missed, leading to suboptimal adherence with the HIV follow-up testing schedule.
Two patients who were initially HIV seronegative tested positive for HIV when returning for their three- and six-month visits. Since data relating to their ongoing sexual exposures was incomplete, these infections could not be definitively established as nPEP failures.
DISCUSSION
As a major local provider of nPEP, we sought to review our programme's recent experience and to identify obstacles to nPEP provision and opportunities for improvement. An encouraging finding was that 92.6% of patients presented within the requisite 72-hour window, and that 94.2% reported high- or medium-risk exposures. Consistent with the experience of other centres, 3 these figures show that even in the absence of an active community awareness campaign, patient requests for nPEP are highly appropriate.
Although adverse events are reported to be the main cause of nPEP discontinuation, 3 with reported nPEP-related side-effect rates as high as 93%, 4 it was also encouraging that ARV tolerability among those who followed-up at our institution was good, with no known toxicity-related regimen discontinuations and few significant laboratory abnormalities (recognizing limitations related to incomplete data). This paucity of laboratory abnormalities has prompted us to now limit routine follow-up blood work for nPEP patients to a complete blood count, serum creatinine and ALT only. Toxicity and tolerability are major reasons for the emerging interest in alternative regimens that substitute lopinavir/ritonavir with newer agents. For instance, a recent report describes favourable rates of side-effects with raltegravir, tenofovir and emtricitabine, 5 and clinical trials evaluating maraviroc- or raltegravir-based regimens are underway. Because our study may not have captured all treatment-related toxicities that occurred, consideration of such regimens may be appropriate in future if improvements in toxicity can be balanced against increases in costs.
Several challenges in providing nPEP were also apparent in this review. For instance, 31.1% of patients reported home addresses at least 15 km away from the hospital, raising concerns that perceived or actual nPEP availability may not be adequately distributed geographically, even in a major metropolitan centre. Indeed, 10.8% of patients were referred to our institution after initially presenting to another facility. In another report, we have recently documented highly variable levels of knowledge and familiarity with nPEP among EDs and sexual health clinics across the Province of Ontario. 6 These phenomena suggest possible inequities in access, and may introduce problematic delays in initiating nPEP, underscoring the need for greater education and standardization of nPEP services as in other jurisdictions. 7
Drug costs represent another important barrier to nPEP provision as medications are not universally covered in Canada outside of hospitalization, although drug coverage and cost could not be adequately evaluated in this retrospective study. Three patients discontinued nPEP due to medication costs. Anecdotally, our team has had to work with patients to employ creative strategies incorporating out-of-pocket, private insurer and compassionate drug supplies to facilitate access to nPEP, but this approach is not sustainable. Preliminary data from frontline providers in Ontario suggests support for increased public funding for this HIV prevention strategy. 6
This chart review also identified quality gaps where improvement will require addressing both provider and patient factors. For example, 24 patients received the two-drug regimen starter pack despite having high-risk exposures, while six received the three-drug regimen for low-risk exposures (Table 2). Plans to address these discrepancies involve inclusion of clearer risk assessment tools on standard order sets and education sessions for frontline providers, which have proven beneficial in other settings. 7 Notably, although decision modelling suggests that the potential for premature PEP discontinuation due to toxicity from the third agent justifies the use of two-drug PEP regimens in many situations, it has also been suggested that better tolerated regimens could even support using a universal three-drug regimen which would also eliminate problems with regimen selection.5,8 However, such a strategy would likely increase unnecessary drug exposure at considerable cost.
Another quality gap was that HIV testing at baseline was incomplete (77.6%), although some patients were tested at another facility immediately prior to presentation. Two patients were HIV-infected at baseline and rapidly linked to HIV care at our institution, emphasizing the importance of such testing. Increased availability of point-of-care HIV tests might improve this situation. STI testing was also low, such that standardized STI testing has since been formally incorporated into our institution's nPEP protocol. Moreover, patient adherence with the follow-up diagnostic testing schedule was poor, although our estimated rates are artificially low because records could not be reviewed for patients opting to follow-up with their family physicians. Abandonment of follow-up is a common challenge in nPEP delivery; similarly, low three-month follow-up rates of 22–46% have been observed in other industrialized settings.7,9,10
These observations, and our finding that 7.0% of those tested at baseline had asymptomatic syphilis, raise concerns about potential undiagnosed HIV and other important infections in this population, despite demonstrated willingness to seek medical attention. Lost opportunities for testing, diagnosis, preventive interventions such as vaccination and linkage to care are serious problems, potentially fuelling ongoing propagation of HIV and STI epidemics, and novel proactive strategies to improve patient follow-up are needed. A seminal prospective study from San Francisco employing proactive follow-up strategies and patient-to-patient counselling boosted nPEP regimen completion to 78% and patient follow-up completion to 75%, albeit in the context of a clinical trial; 11 other strategies that warrant study might include telephone reminders and mobile phone text messaging. 7
Our study is limited by challenges inherent to the retrospective nature of chart review. In particular, missing data due to many patients choosing to follow-up elsewhere, and variation in provider practice, limited our ability to accurately ascertain regimen completion and toxicity rates as well as some clinical outcomes. Present efforts to standardize our clinical data collection will better equip us to quantify these important measures of programme success in the future.
Sexual activity remains the most common route of transmission for HIV infection in our setting as it is worldwide, and PEP for such non-occupational exposures is an important clinical and public health intervention that warrants continued implementation research, quality improvement initiatives and public investment.
Footnotes
ACKNOWLEDGEMENTS
The authors gratefully acknowledge Ann-Marie Doherty for her work in identifying the patients relevant to our study. DHST received research fellowship funding from the Ontario HIV Treatment Network and Canadian Institutes of Health Research during the conduct of this work.