The Impact of Treatment as Prevention on the HIV Epidemic in British Columbia, Canada

Abstract

Purpose of Review

This study proposes to describe the impact of a publicly funded Treatment as Prevention (TasP) strategy in British Columbia (BC), Canada, in decreasing the individual and public health impact of the HIV/AIDS Epidemic.

Recent Findings

In BC, TasP has been associated with a steady decline in HIV-related morbidity and mortality. At the same time, a demographic transition was observed among people living with HIV (PLWH), with the majority of those on antiretroviral treatment (ART) now ≥50 years of age, living with at least one comorbidity, and dying from age-associated comorbidities. We also documented a progressive increase in the proportion of viral load suppression as a result of ART expansion. While the pre-ART CD4 T cell count has increased steadily in recent years, there is still a large proportion of PLWH being diagnosed in later stages of HIV infection. New HIV diagnoses have been rapidly declining, however to a lesser extent among men who have sex with men (MSM), and BC is currently experiencing an increase in infectious syphilis cases in this population. These facts reinforce the effectiveness of TasP in decreasing HIV transmission, but at the same time, it highlights the need for further innovation to enhance the control of HIV and syphilis among MSM.

Summary

This study supports the development of new approaches that address existing gaps in the TasP strategy in BC, and the future health needs of PLWH.

Introduction

Antiretroviral treatment (ART) dramatically decreases HIV/AIDS-related morbidity and mortality among persons living with HIV (PLWH), as well as the risk of onward HIV transmission [1•-4]. To maximize the benefit of ART, two bottlenecks need to be effectively and simultaneously addressed. The first is expanding HIV testing and the second is facilitating ART access. Together, these form the strategy commonly referred to as Treatment as Prevention (TasP) or Test and Treat (T&T) [2, 5]. Addressing these bottlenecks is enabled by the current HIV treatment guidelines, adopted in 2015, which recommend starting ART at HIV diagnosis, irrespective of CD4 cell count [6, 7]. Indeed, the most recent global ART roll out target proposed by the United Nations Joint AIDS Program (UNAIDS) and approved by the United Nations (UN) in 2015, has at its foundation TasP as its strategy. The UN 90-90-90 Target proposes that by 2020 at least 90% of PLWH throughout the world should be diagnosed and aware of their HIV status; at least 90% of those diagnosed be on ART; and at least 90% of those on ART be virologically suppressed, which translates into a 73% virologic suppression coverage among all PLWH [8]. It has been proposed that the successful implementation of the UN 90-90-90 Target would bring the “End of AIDS as a Pandemic” by 2030 [8, 9].

Several countries, including Botswana, Cambodia, Denmark, Iceland, Singapore, Sweden, the United Kingdom, as well as the cities of Amsterdam, New York City, Melbourne, Paris, and San Francisco have reportedly met the UN 90-90-90 Target ahead of schedule [10-12•]. However, it is unclear to what extent meeting this target has lowered HIV/AIDS-related morbidity and mortality, and reduced HIV transmission in these jurisdictions. Furthermore, HIV testing, ART uptake, and related outcomes remain sub-optimal in many jurisdictions around the globe [10, 13]. As a result, in July 2019, UNAIDS called for greater urgency in meeting the UN 90-90-90 Target [14]. Recently, clinical trials have shown that optimizing the impact of TasP in high prevalence settings can be challenging, even when high uptake of testing and ART coverage have been documented [15-17]. Clearly, as pointed out by the accompanying editorial [18], further efforts will be needed to achieve the “End of AIDS as a Pandemic”.

The province of British Columbia (BC), Canada, has a fully centralized publicly funded ART program operated through the BC Centre for Excellence in HIV/AIDS (BC-CfE). The BC-CfE Drug Treatment Program (DTP) has enabled a steady expansion of ART coverage since 1996 [19]. A province-wide expansion of ART in 2010 had the explicit goal of accelerating control of the HIV/AIDS epidemic in the province [20]. Thus, the objective of this population-based study was to characterize the impact of continuous expansion of ART coverage on HIV/AIDS-related morbidity and mortality and transmission, as well as to characterize the progress of BC towards the UN 90-90-90 Target.

Methods

Main Data Sources

This study employed two different sources of data. We utilized longitudinal data from all PLWH ever registered in BC-CfE’s DTP [20]. The BC-CfE has provided centralized publicly funded ART since 1992, and specialized HIV laboratory monitoring to all BC residents living with HIV since 1996. In BC, ART is prescribed by primary care providers and specialized physicians who are guided by the BC-CfE’s HIV therapeutic guidelines, and can access clinical training at the BC-CfE for the diagnosis, and management of PLWH [21]. Data captured in the DTP include socio-demographic (e.g., sex, age, ethnicity, geographic location of residence), clinical (e.g., CD4 cell count, plasma HIV viral load, AIDS-defining illness, mortality) and treatment variables (e.g., antiretroviral regimen information, date of ART initiation, treatment interruptions, ART adherence).

The DTP registry database has been linked to various provincial administrative databases to form the second data source, the Seek and Treat for Optimal Prevention of HIV/AIDS (STOP HIV/AIDS) population-based cohort. The STOP HIV/AIDS cohort complements the DTP data by containing administrative health information on all diagnosed PLWH regardless of whether they are accessing ART in BC [20, 22-27]. This linkage is done annually, in a de-identified manner, with two- to three-year data lag between this cohort end of follow-up date and the DTP last entry date. In both datasets, participants were followed until the last contact date with the provincial healthcare system (e.g., DTP entry, physician visits, hospitalizations, laboratory test), the date of death, study end or the date in which they moved out of BC. The databases included in the STOP HIV/AIDS cohort, along with their corresponding data capture are comprehensively detailed in the Supplemental Material.

Study Outcomes

The following sessions describe the different indicators that were investigated in this study. It is important to mention that the period of data coverage for each indicator will vary due to data availability.

HIV and Infectious Syphilis Surveillance Indicators

Population-level data on testing episodes from 2009 to 2017, HIV and infectious syphilis diagnosis from 1996 to 2017, were obtained through the BC Centre for Disease Control (BCCDC), which houses HIV surveillance and screening data (>95% of all HIV screening tests in BC) as well as all HIV confirmatory tests in BC from its Public Health Laboratory [28]. This laboratory also performs virtually all syphilis tests done in the province. Annual surveillance data regarding new HIV diagnosis, stratified by HIV exposure group (men who have sex with men (MSM) and people who have ever injected drugs (PWID)) were extracted from STOP HIV/AIDS quarterly monitoring reports and complemented with BCCDC annual reports, and BCCDC Reportable Diseases Data Dashboard [26, 29, 30]. Data for active cases of infectious syphilis were extracted from the BCCDC annual Sexually Transmitted Infections reports and the above mentioned dashboard [31].

AIDS and HIV-Related Morbidity and Mortality

AIDS-defining illness (ADI) case reports were obtained from the BC-CfE and supplemented with clinical records from St. Paul’s Hospital admissions, the BC Cancer and BC Vital Statistics Agencies, from 1985 to 2017. All these agencies use passive reporting systems to collect data for ADI cases [32]. AIDS-defining illnesses were categorized according to established definitions shown in Table 1. The BC Vital Statistics Agency provides HIV- and AIDS-related mortality information for all DTP participants on a monthly basis, and this data was supplemented by ADI cases obtained from the BC Cancer Agency, which oversees provision of all cancer-related care in the province and provides data that are annually linked to the DTP. Note that ADI cases have been updated since our previous publication, but still remain incomplete due to the delays in reporting [4]. To further complement this data, we classified deaths as HIV/AIDS-related if an individual had at least one ADI in the three months before the date of death [33], or if the cause of death contained at least one HIV-related diagnosis code in any field. The codes for HIV-related death were based on the International Classification of Disease (ICD) codes versions 9 and 10 (see Supplemental Material). Note that due to data constraints, we focused on the deaths reported between 1987 and 2017, and we were not able to ascertain whether the cause of death was HIV/AIDS-related prior to 1993.

Table 1.

List of AIDS-defining illnesses used in this study

AIDS-defining illnesses	Abbreviation
Candidiasis of esophageal, bronchi, trachea, or lungs	CANDIDA
Invasive cervical cancer	CERVCA
Cytomegalovirus disease (site not specified)	CMV
Coccidiomycosis disseminated or extrapulmonary	COCCID
Cryptococcosis extrapulmonary	CRYPTOCO
Cryptosporidiosis chronic intestinal	CRYPTOSP
HIV encephalopathy	DEMENT
Histoplasmosis disseminated or extrapulmonary	HISTO
Herpes simplex infection, chronic mucocutaneous	HSV
lsosporiasis chronic intestinal	ISOSPOR
Kaposi's sarcoma	KS
Lymphoma primary of brain	LYMPHOMA
Mycobacterium Avium Complex (MAC) or Mycobacterium kansasii disseminated or extrapulmonary	MAC
Tuberculosis AIDS-defining unspecified	MTB
Mycobacterium other or unspecified species disseminated or extrapulmonary	MYCO
Lymphoma, non-Hodgkins, AIDS defining unspecified	NHL
Pneumocystis jiroveci/carinii pneumonia	PCP
Progressive multifocal leukoencephalopathy	PML
Recurrent bacterial pneumonia	REC PNEUM
Recurrent salmonella septicemia	SALM SEP
Toxoplasmosis of brain	TOXO
Wasting syndrome due to HIV	WASTING

Age-Associated Morbidity

Base on the DTP data, from 1996 until 2017, we estimated the annual distribution of participants by different age groups (<30, 30-39, 40-49, 50-59 and ≥60 years) to assess the demographic shift in our population. Additionally, we estimated the number of age-related comorbidities our participants developed based on the STOP HIV/AIDS cohort data, from 1996 until 2016. These comorbidities were identified through the physician billing and hospital discharge databases using case-finding algorithms based on both ICD-9 and 10 versions (see Supplemental Material). The age-related comorbidities included cardiovascular, kidney and liver diseases, chronic obstructive pulmonary disease (COPD), diabetes mellitus and hypertension [34-36]. We did not include non-AIDS related cancers, as the linkage between the STOP HIV/AIDS cohort and BC Cancer Agency cancer registry remains incomplete.

Treatment Monitoring

Treatment monitoring indicators were obtained using the DTP data from 1996 to 2017. For this study, we obtained the highest viral load for each participant per year. Since the limit of quantification of viral load assays has changed over time, we used two cut-offs 1000 copies/mL (a public health endpoint) and 200 copies/mL (a clinical endpoint). Additionally, to estimate how late participants initiate ART in the course of their illness, we obtained CD4 T cell count data at ART initiation among treatment naïve participants. Note that all viral load tests in BC are performed at the St Paul’s Hospital virology laboratory and are uploaded automatically in the DTP database. CD4 T cell counts, however are performed in different laboratories across the province, so we estimate that the DTP captures approximately 85% of all CD4 tests done during the study period.

UN 90-90-90 Target

BC's performance towards fulfilling the UN 90-90-90 Target was estimated between 2010 and 2016 from the STOP HIV/AIDS database. In order to estimate the denominator for the first "90" (the proportion of PLWH in BC who have been diagnosed), we obtained province-wide HIV prevalence estimates from a previous published manuscript [37], and defined successful HIV diagnosis as the first instance of one of the following: a confirmed HIV-positive test, a detectable viral load, an HIV-related physician visit or hospitalization, a reported AIDS-defining illness or the dispensation of ART during the calendar year [37, 38]. The definition for the second "90" (the proportion of PLWH diagnosed and on ART) was modified from the definition used in our previous publications, as the latter was too conservative and not aligned with the more liberal definitions used in other settings [37-40]. As a compromise between the conservative and liberal definitions, we defined the second "90" as participants who received ART for at least a month in the calendar year. The definition for the third "90" (the proportion of PLWH on ART and virologically suppressed) was defined as participants whose last viral load within the calendar year was below 200 copies/mL [39, 40]. However, defining suppression on the basis of a single viral load measurement in the calendar year is extremely liberal. Since the objective of ART is to achieve long-term viral suppression, we therefore added a fourth indicator to the UN 90-90-90 Target. This indicator, which we termed "Remained suppressed", considered all viral loads in a given calendar year to define suppression [37]: specifically, among those whose last viral load was <200 copies/mL, it quantified the proportion of participants who were able to maintain a suppressed viral load during the entire calendar year. Finally, and in contrast to our previous publications [38], we censored participants at death or at lost to follow-up (defined as ≥18 months without any record of healthcare utilization).

Statistical Analysis

As this was a descriptive study, categorical variables were compared using Fisher's exact test, and continuous variables were compared using the Kruskal-Wallis test [41]. Trends were modeled using generalized additive models to account for the non-linear temporal trends in some of the outcomes [42]. These analyses were performed using the statistical software SAS (version 9.4) or R© version 3.5.0 mgcv library (functions gam, summary, gam.check, and residuals.gam). All p-values reported are two-sided, and significance was set at the 5% level.

Ethics Approval

This study was approved by the University of British Columbia ethics review committee at the St Paul’s Hospital, Providence Health Care site (H18-02208 and H05-50123).

Results

Between 1996 and 2017, ART was progressively expanded in BC in four consecutive phases: 1996-2000; 2001-2004; 2005-2010; and 2011-2017 (overall percent increase 819%; p-value <0.001). Since 2010, HIV testing rates have more than doubled from 262 in 2009 to 611 in 2017 per 1,000,000 population (p-value <0.001). Since 1996 the number of new HIV cases diagnosed per year has decreased from 702 in 1996 to 182 in 2017 (overall percent decrease −74%; p-value <0.001), with PWID experiencing a greater decline than MSM (percent decrease −95% versus −34% respectively). Of note, during the same time frame, new infectious syphilis cases per year have increasing, and more recently, we observed nearly a five-fold increase from 154 cases in 2010 to 682 cases in 2017 (p-value <0.001).

HIV-Related Morbidity and Mortality

Figure 2(A) displays the ADI incident cases in BC between 1985 and 2017. ADI incidence in BC peaked in 1994 with 696 cases, decreasing steadily thereafter to 53 cases in 2017 (−92%; p-value <0.001). ADI incidence has been consistently higher among males, with the male/female ratio reaching its peak in 1991 (ratio 49:1) and remaining constant since 2013 (ratio 3:1). There were also marked age differences over time; at the peak of the AIDS epidemic, we observed that ADI incidence was higher among those 30-39 years (>50% of all cases), and more recently, ADI incidence has been rapidly increasing among those ≥50 years (since 2016 almost 60% of all cases).

Figure 2.

Distribution of the number of AIDS-defining illnesses (A) and death cases due to HIV/AIDS and other causes (B) in British Columbia, Canada, by calendar year.

The spectrum of ADIs (see Table 1 for abbreviations) has also evolved over time, with PCP, KS, MAC and CANDIDA accounting for >50% cases before 1996, and REC PNEUM, PCP, WASTING and CANDIDA accounting for >50% cases in the last five years. During the study period there were 5780 deaths from all causes, of which 3255 (56%) were AIDS-related (Figure 2(B)). In 1994, the mortality rate reached its peak (19 per 100 person-years), and since then it has decreased by 76% by 2017. Since 2004 however, non-AIDS related deaths have steadily increased, going from 35% of all cases (1 per 100 person-years) in 2004 to 78% (4 per 100 person-years) in 2017. The most common causes of death are age-related comorbidities and more recently, as of 2015, opioid-related overdoses.

Age-Associated Morbidity

The age of DTP participants has increased from a median of 38 (25^th-75^th percentile: 33-44) years in 1996 to 51 (43-58) years in 2017 (Figure 3). Using the STOP HIV/AIDS dataset, we observed that the prevalence of age-associated comorbidities (excluding mental health) has increased since 1996, with 16% of PLWH aged 50 or more now living with at least two comorbidities, and this proportion increases to 24% among PLWH 60 years and older (Figure 4). In our population, hypertension (12%) represented the highest burden of illness in 2016 (Figure 5).

Figure 3.

Distribution of the age of participants of the Drug Treatment Program in British Columbia, Canada, by calendar year.

Figure 4.

Distribution of the number of age-associated comorbid conditions by age in British Columbia, Canada, in 2016.

Figure 5.

Distribution of age-associated comorbid conditions in British Columbia, Canada, in 2016.

Treatment Monitoring

Figure 6 shows the pre-ART CD4 T cell count of DTP participants from 1996 until 2017. As shown, the median pre-ART CD4 T cell count increased from 260 (120-380) cells/μL in 1996 to 380 (180-590) cells/μL in 2017. In the last five years, the pre-ART CD4 T cell count has been quite constant, and although we found no difference in pre-ART CD4 by gender (males: 430 (250-620) cells/μL vs. females 400 (220-620) cells/μL; p-value 0.2854), we observed that those younger than 40 years had the highest CD4 levels (<30 years: 490 (325-655) cells/μL; 30-39 years: 450 (270-630) cells/μL; 40-49 years: 380 (190-597) cells/μL; ≥50 years: 380 (185-570) cells/μL; p-value <0.0001).

Figure 6.

Distribution of pre-ART CD4 T cell count in British Columbia, Canada, by calendar year of ART initiation

To characterize the population-level distribution of viral load, we adopted a conservative approach by recording the highest viral load for every participant in a specific calendar year (Figure 7). The trend in viral load over time is quite striking. Before 2006, the proportion of participants with viral load ≥1000 copies/mL was nearly 50%, and this proportion decreased dramatically to 10% in 2017 (p-value <0.0001). Note that in our most recent year, the proportion of participants with a viral load <200 copies/mL was estimated at 88%.

Figure 7.

Distribution of the highest participant’s viral load in British Columbia, Canada, by calendar year.

UN 90-90-90 Target

From 2010 to 2016, the estimated number living with HIV (diagnosed and undiagnosed) in BC decreased slightly from 10,688 to 10,662 cases. In this same period, the total number of diagnosed PLWH increased from 7,945 to 8,989 (13% relative increase; p<0.0001), the total number of participants on ART increased from 5,968 to 7,624 (28% increase; p<0.0001), the total number of those virologically suppressed (by the last viral load in the year) increased from 5,355 to 7,123 (33% increase; p<0.0001), and the total number of participants who remained suppressed during the full year increased from 4557 to 6550 (44% increase; p<0.0001) (Figure 8).

Figure 8.

Progress in the UN 90-90-90 Target in British Columbia, by calendar year.

As a final analysis, we wished to report these in terms of BC's performance in meeting the UN 90-90-90 Target. Overall in BC, between 2010 and 2016, the % Diagnosed increased from 74% to 84% (13% increase; p<0.0001), the %On ART (among diagnosed PLWH) increased from 75% to 85% (13% increase; p<0.0001), the %Suppressed (among those on ART) increased from 90% to 93% (4% increase; p<0.0001), and the %Remained Suppressed (among those Suppressed) increased from 85% to 92% (8% increase; p<0.0001). That is, in 2016, BC's performance towards the UN 90-90-90 Target was 84-85-93. Altogether, this translates into 67% of all PLWH virologically suppressed in 2016 (a 33% increase from 2010, when only 50% of PLWH were virologically suppressed; p<0.0001).

Discussion

Our study shows that in BC, expanded access to HIV testing and publicly funded ART, within the context of a province wide TasP strategy, has been associated with steady decline in ADIs and HIV/AIDS-related mortality, as well as in new HIV diagnoses. We also documented a progressive increase in the proportion of PLWH with undetectable viral load as a result of ART use. Regarding the UN 90-90-90 Target, in 2016, BC achieved 84-85-93 compared with 74-75-90 in 2010, demonstrating a rapid and steady progress towards the UN 90-90-90 Target. Nevertheless, there is still room for improvement. While the median CD4 T cell count at the time of starting ART has increased steadily in recent years, it is still below 500 cells/μL, suggesting that a large proportion of PLWH is still being diagnosed in later stages of HIV infection (i.e., Stages 2 or 3). In some indicators, we demonstrated that there still exists heterogeneity by gender and age, likely attributable to socio-demographic (e.g., income, geography) and health-related (e.g., substance use, mental health, comorbidities) factors. Therefore, further initiatives are needed to identify modifiable risk factors that can be targeted to narrow these gaps and improve access to services and health outcomes among the most vulnerable in our population.

The number of new HIV diagnoses among PWID declined substantially from 1996 to 2008 and remained at a very low level, which is probably a result of a favorable synergistic interaction between harm-reduction programs and ART in BC [43]. Note that BC has been a leader in the expansion of harm-reduction and addiction treatment programs in Canada for the last two decades, despite the recent rise in opioid-related overdose deaths. These programs have included needle and syringe supply programs, opioid substitution therapy programs, supervised injection, and more recently, overdose prevention sites and drug checking sites [44, 45]. Somewhat in contrast, new HIV diagnoses among MSM have shown a slower decline in recent years, and it is notable that BC is currently experiencing a major expansion in the number of cases of infectious syphilis in this population [26]. This reinforces the effectiveness of TasP in decreasing transmission of HIV among MSM, but at the same time, it highlights the need for further innovation to enhance the control of HIV and syphilis (as well as other sexually transmitted infections) in this population. On this note, in January of 2018, the BC-CfE launched a publicly funded pre-exposure prophylaxis (PrEP) program for BC residents at high-risk of acquiring HIV, including but not restricted to MSM. Given the timing of this study the use of PrEP is unlikely to have contributed to these findings.

The data also revealed a demographic transition among PLWH in BC, with the majority of PLWH on ART now ≥50 years of age and living with at least one comorbidity. In keeping with this demographic shift, we found that the majority of deaths among PLWH are now due to age-associated comorbidities. Whether and to what degree HIV infection accelerates the incidence of some of these comorbid conditions is yet to be determined, but a significant portion of these comorbidities may be related to the high prevalence of hepatitis C, smoking and alcohol use in this population; all of which would benefit from intensified prevention and treatment initiatives. Nevertheless, the combined effects of a declining number of new HIV diagnoses per year, and the increase in non-AIDS-related mortality, and possibly in/out migration, have led to a plateau in the number of PLWH on ART: for the first time since modern ART was launched in 1996 the number of PLWH on ART has stabilized.

There are some limitations associated with our study. As this study was based on the analysis of both ecological and administrative data, caution is needed when interpreting the results and causality cannot be inferred. Second, data on ADIs may be incomplete as this data relies on a passive system of data collection; as such we tried to mitigate underreporting by performing multiple data linkages. Third, data on comorbidities relied on case-finding algorithms based on healthcare administrative databases, which can be prone to coding errors. Last, due to the limited availability of data on key social determinants of health (e.g., income level, ethnicity, behavioural factors), we were not able to conduct further analyses to fully characterize the contextual effects of these factors on our indicators.

Conclusion

In conclusion, in an environment of publicly funded expanded access to HIV testing, care and monitoring, within a province that has adopted the TasP strategy, HIV/AIDS related morbidity, mortality and HIV transmission have decreased to the lowest levels seen in decades. However, these gains have not benefited all population sub-groups equally, and we have also observed a marked increase in the prevalence of age-associated comorbidities among PLWH, especially among those 50 years of age or older. Therefore, until a cure for HIV is found, there remains an urgent need for new research on policy and program interventions to help PLWH age in as healthy a manner as possible [46, 47]. The evidence provided by this study can help support the development of interventions that address existing gaps in care for PLWH, and to inform new approaches that best address their future health needs.

Figure 1.

Surveillance indicators for HIV and infectious syphilis in British Columbia, Canada, by calendar year.