Abstract
BACKGROUND
Mortality rates among persons with HIV (PWH) have fallen since 1996 following the widespread availability of effective antiretroviral therapy (ART). Patterns of cause-specific mortality are evolving as the population of PWH ages. We aimed to investigate longitudinal trends in cause-specific mortality rates among PWH starting ART in Europe and North America.
METHODS
We used data on PWH aged ≥16 years old when starting ART between 1996 and 2020 from 17 European and North American HIV cohorts contributing data to the Antiretroviral Therapy Cohort Collaboration. Causes of death were classified by both a clinician and an algorithm if ICD9/10 data were available, or independently by two clinicians. Disagreements were resolved through panel discussion. We used Poisson models to compare cause-specific mortality rates during calendar periods 1996–99, 2000–03, 2004–07, 2008–11, 2012–15 and 2016–20, adjusted for time-updated age, CD4 count, and whether ART-naïve at the start of each period.
FINDINGS
Among 189,301 PWH, 16,832 (8.9%) died. Causes of death were classified for 13180 (78%) deaths: the most common causes were AIDS (4203 deaths; 25%), non-AIDS non-hepatitis malignancy (2311; 14%) and cardiovascular (1403; 8%). The proportion of deaths due to AIDS declined from 49% during 1996–9 to 16% during 2016–20. Rates of all-cause mortality per 1000 person-years decreased from 16.8 (95%CI: 15.4–18.4) during 1996–99 to 7.9 (7.6–8.2) during 2016–20. Rates of all-cause mortality declined with time: the average adjusted mortality rate ratio [aMRR] per calendar period was 0.85 (95%CI 0.84–0.86). Rates of cause-specific mortality also declined: the most pronounced reduction was for AIDS-related mortality (average aMRR per period 0.81; 95%CI 0.79–0.86). There were also reductions in rates of cardiovascular-related, liver-related, non-AIDS infection-related, non-AIDS-non-hepatocellular carcinoma malignancy-related, and suicide/accident-related mortality (average aMRRs per period 0.83 (0.79–0.87), 0.88 (0.84–0.93), 0.91 (0.86–0.96), 0.94 (0.90–0.97), and 0.89 (0.82–0.95), respectively). Mortality rates among people who acquired HIV through injecting drug use increased among women and decreased less for men than other groups.
INTERPRETATION
There have been reductions over time in rates of most major causes of death, particularly AIDS-related deaths, among PWH on ART. However, such reductions were not seen for all subgroups. Interventions targeted at high-risk groups, substance use, and comorbidities may further increase life expectancy in PWH towards that in the general population.
FUNDING
US National Institute on Alcohol Abuse and Alcoholism.
INTRODUCTION
Before the introduction of combination antiretroviral therapy (ART) in 1996, people with HIV (PWH) experienced very high mortality rates, mostly due to AIDS(1). Since then, the vast majority of PWH in high-income countries are now on ART(2). The successful suppression of viral replication with ART has led to substantial reductions in the risk of AIDS and death and corresponding dramatic increases in life expectancy(3). With reduced AIDS-related mortality PWH are ageing and experiencing and increased mortality due to other, age-related causes such as cardiovascular disease and cancer(4, 5). Substance use and comorbidities such as hepatitis C virus (HCV) are more common among PWH than in the general population, leading to increased rates of mortality related to these causes(6–8).
Between 2012 and 2015, following the INSIGHT START Study(9), international treatment guidelines evolved to recommend that all PWH should receive ART regardless of their disease stage and CD4 count(10). Accordingly, PWH are starting ART earlier from diagnosis. Additionally, more potent and well tolerated ART regimens, particularly second-generation integrase inhibitors, have continued to become available and are now widely used(11), leading to ongoing reductions in AIDS incidence and mortality(12). Policy makers need current data to understand the changing burden of cause-specific mortality among PWH, both overall and in subgroups defined by age, sex, and mode of acquisition of HIV.
The Antiretroviral Therapy Cohort Collaboration (ART-CC)(13) has classified causes of death among PWH living in Europe and North America since 1996. We investigated longitudinal trends in rates of cause-specific mortality among adult PWH who started ART in Europe and North America between 1996 and 2020.
METHODS
Study design and population
Through the Antiretroviral Therapy Cohort Collaboration (ART-CC)(13), data were combined from 17 European and North American HIV cohort studies for which causes had been classified for over 70% of deaths. The included cohorts are listed in supplementary table 1 (appendix page 1). Ethics committees or institutional review boards approved the individual cohorts, which used standardised data collection methods, and regularly followed-up patients. Cohorts gathered information on mortality through linkage with vital statistics agencies and hospitals or physician report, and the active follow-up of participants. Eligible participants were ≥16 years old when starting combination ART (three drugs or more) and had no prior exposure to mono or dual ART regimens. Included participants had a CD4 count and HIV-1 RNA viral load measurement within a window of one month before and one week after starting ART.
Procedures
We adapted the Cause of Death (CoDe) project protocol(14) to classify causes of death information into a single cause in the HICDEP format (https://hicdep.org/Wiki/v/10/pt/4/Table/104/FieldID/1321). Information on cause of death was recorded either as International Classification of Diseases, Ninth Revision (ICD-9) or Tenth Revision (ICD-10), or free text. If ICD-9 or ICD-10 codes were available, causes of death were classified by a clinician and a computer algorithm(14). When ICD-9 or ICD-10 codes were not available, two clinicians independently classified each death. Disagreements between clinicians and/or computer-assigned codes were resolved via panel discussion.
Clinicians classified deaths using summary tables of patient data including ICD-9/ICD-10 codes or free text for cause of death, patient characteristics at ART initiation (age, sex, transmission risk group, AIDS-defining conditions, and HCV antibody status), use of ART prior to death, AIDS-defining conditions after starting ART, latest CD4 cell count (within 6 months of death). Appendix page 2 contains the rules used: these define causes of death based on particular combinations of information, or lack of information. Deaths were coded as AIDS-related if there was a serious AIDS defining condition prior to death and/or a low CD4 count (<100/μL) within a year of death (18 months if off treatment), and a diagnosis compatible with AIDS as cause of death(8, 14). All other deaths, including those of unknown cause, were considered non-AIDS related. Liver-related deaths for which there was no evidence of hepatitis B or C, and those including mention of alcohol, were coded as liver failure. Liver-related deaths in PWH with a history of hepatitis B or C were coded as due to chronic viral hepatitis B or C, respectively. Some cohorts performed the clinician coding “in-house”, whilst for others this was done by a team of clinicians from ART-CC cohorts. Some cohorts used a mix of in-house and ART-CC coding. A second computer algorithm mimicked this process for the cohorts with the largest numbers of deaths requiring assignment of causes. This built on the first algorithm, but additionally included added coding based on the above rules for coding causes of death (Appendix page 4).
Causes of death were grouped into the following categories based on HICDEP codes: AIDS including AIDS infections and malignancies [HICDEP category 01 and sub-categories]; Non-AIDS infection including infections other than AIDS infections [02]; Non-AIDS, non-hepatitis malignancy – all malignancies except for hepatocellular carcinoma (HCC) [04]; Liver – chronic viral hepatitis, liver failure, and HCC [03, 14, 04.20]; Cardiovascular/heart – acute myocardial infarction, stroke, and heart or vascular [08, 09, 24]; Respiratory – chronic obstructive lung disease and other respiratory diseases [13, 25]; Substance use – active substance use including acute intoxication [19]; violent death – suicide, accident, or other violent death [16, 17]; central nervous system (CNS) – CNS disease including Parkinson’s and Alzheimer’s [23]; unknown/unclassifiable – unclassifiable causes and unknown [91, 92]; Other – All others, including ‘other causes’ [90].
Statistical analyses
Data were split into calendar periods of follow-up (1996–99, 2000–03, 2004–07, 2008–11, 2012–15, and 2016–20). We derived participants’ characteristics at the time of starting ART and at the beginning of each subsequent period. The last follow-up date was April 2020. People were censored at the first of a cohort-specific administrative censoring date, date of death, or date of loss-to-follow-up. Loss-to-follow-up was defined as a period of at least a year between the person’s last visit and the cohort-specific administrative censoring date, with the date of loss to follow-up defined as 6 months after this last visit date. When a participant was lost-to follow-up but then found to have died via linkage to death registry, their follow-up was instead censored at their date of death. We performed a sensitivity analysis excluding deaths that happened after loss to follow-up. As ART should be continued once started, we did not consider ART discontinuation in the analyses.
We used Poisson regression to estimate cause-specific mortality rate ratios (MRRs) for each follow-up calendar period compared with the 2000–03 period, as well as the average MRR per-period assuming that changes across periods were log-linear. The 2000–03 period was chosen as the comparator as all cohorts had data for this period, whereas two cohorts had limited data for the 1996–99 period. The Poisson models accounted for each individual’s exposure time (in days) during each calendar year period of follow-up. These periods began on 1st January 1996, 2000, 2004, 2008, 2012, and 2016.
Estimated MRRs were adjusted for time-updated age (16–29, 30–39, 40–49, 50–59, ≥60 years), time-updated CD4 count (0–199, 200–349, 350–499, ≥500 cells/mm3, or missing), whether the person was ART-naïve at the start of each period, and cohort. For time-updating CD4 counts, a window from 365 days prior to the start of that follow-up period, to 30 days afterwards was used. In the categorised time-updated CD4 count variable, a separate missing data category was included so that people with missing data were not excluded. We estimated MRRs overall and stratified into subgroups based on time-updated age (16–39, 40–59, ≥60 years), time-updated CD4 count (0–199, 200–349, 350–499, ≥500 cells/mm3, and missing), whether ART-naïve at the start of each period (to capture the period at the start of ART when mortality rates are highest), time-updated prior AIDS-defining event status (no, yes; defined here: https://hicdep.org/Wiki/v/10/pt/4/Table/94/FieldID/1206), time-updated HCV antibody status (negative, positive), cohort region (Europe, North America), and self-reported sex/HIV acquisition risk group (men who have sex with men [MSM], men with injecting drug use [IDU], women with IDU, men having sex with women, women having sex with men, and either sex with risk group other/unknown [PWH in this final category are mostly from the Veterans Aging Study Cohort]). Men and women with unknown/other mode of HIV acquisition were combined into one subgroup for the cause-specific mortality analyses.
Due to data sharing rules regarding cause-of-death information, analyses for one of the cohorts were done separately based on deaths classified using the second computer algorithm. Mortality rate ratios and their confidence intervals were produced separately for this cohort (in Yale, US) and for the other cohorts (in Bristol, UK). To combine them, their natural logs were taken and then combined using inverse-variance weighted meta-analysis (in Bristol, UK) using the metan command in Stata v17.1 on the log-transformed rate ratios and their standard errors. Estimated MRRs for 1996–99 vs 2000–03 exclude this cohort, for which data were unavailable for the 1996–99 period. All other estimated MRRs are based on all cohorts.
In a sensitivity analysis, we removed the two cohorts for which the second computer algorithm was used instead of clinician input.
An additional, post-hoc analysis was performed to investigate longitudinal changes in rates of causes of death among people who acquired HIV through IDU, separately among those living in Europe and North America. We estimated per-period MRRs for each cause of death adjusted for time-updated age, CD4 count, whether the person was ART-naïve at the start of each period and cohort.
All analyses were performed using Stata version 16.1.
Role of the funding source
The funders had no role in the collection, analysis or interpretation of data, report writing, or the decision to submit this study for publication.
RESULTS
There were 16,832 deaths among 189,301 PWH followed for 1,519,200 person-years. The crude all-cause mortality rate was 11.1 (95% CI: 10.9–11.3) per 1000 person-years, declining from 16.8 (15.4–18.4) in 1996–99, rising to 17.1 (16.4–17.8) in 2000–03 and declining to 7.9 (7.6–8.2) in 2016–20. The age at death increased from median 39 (Interquartile range [IQR]: 34–49) years during 1996–99 to 56 (48–65) years during 2016–20.
The number of PWH contributing data increased from 20,448 during 1996–99 to 145,996 during 2016–20 (Table 1). The median age at starting ART increased from 35 (IQR: 31–42) years during 1996–99 to 38 (30–48) years during 2016–20. Participants’ median age increased from 37 (32–43) on 01/01/2000 to 43 (36–50) on 01/01/2008 and 47 (38–54) on 01/01/2016. The percentage of PWH with CD4 cell counts of 0–199 cells/mm dropped from 42.6% at the start of the 1996–99 period to 7.6% at the start of 2016–20. Of the 189,301 included PWH, 44,348 (23%) were women, 36,311 (19%) had AIDS when starting ART, and 14,967 (8%) had tested positive for HCV antibodies before starting ART. Missing data is also indicated in table 1. Time-updated CD4 count at the start of a new calendar year period was missing for some PWH, but this was not for a large percentage of PWH (7.5% for 2016–20). Most missing data was regarding HCV antibody testing, but this improved over time from 50.4% in the earlier follow-up period, to 11.1% in 2016–20.
TABLE 1:
Characteristics (% with characteristic) of persons with HIV overall and at the start of follow up within each period*.
| Variable | Number of deaths | N alive during each calendar period | |||||
|---|---|---|---|---|---|---|---|
| 1996–99 | 2000–03 | 2004–07 | 2008–11 | 2012–15 | 2016–20 | ||
| Total | 16832 | 20448 | 47124 | 74386 | 108946 | 139851 | 145996 |
| Age (years) | |||||||
| 16–39 | 5728 (34.1%) | 13969 (68.3%) | 28242 (60.0%) | 37463 (50.4%) | 46792 (43.0%) | 54166 (38.7%) | 48140 (33.0%) |
| 40–59 | 8724 (51.8%) | 5874 (28.3%) | 17073 (36.2%) | 33049 (44.4%) | 54489 (50.0%) | 72899 (52.1%) | 79989 (54.8%) |
| ≥60 | 2380 (14.1%) | 605 (3.0%) | 1809 (3.8%) | 3874 (5.2%) | 7665 (7.0%) | 12786 (9.1%) | 17867 (12.2%) |
| Started ART during | |||||||
| A previous period | NA | NA (0.0%) | 18956 (40.2%) | 42138 (56.7%) | 66508 (61.1%) | 96478 (69.0%) | 123370 (84.5%) |
| Current period | 16832 (100.0%) | 20448 (100.0%) | 28168 (59.8%) | 32248 (43.4%) | 42438 (39.0%) | 43373 (31.0%) | 22626 (15.5%) |
| Sex/HIV mode of acquisition | |||||||
| Men who have sex with men | 3410 (20.3%) | 7052 (34.5%) | 13598 (28.9%) | 22303 (30.0%) | 38831 (35.6%) | 57036 (40.8%) | 62353 (42.7%) |
| Male, injecting drug use | 2637 (15.7%) | 3428 (16.8%) | 5902 (12.5%) | 7290 (9.8%) | 8014 (7.4%) | 8111 (5.8%) | 7148 (4.9%) |
| Female, injecting drug use | 690 (4.1%) | 956 (4.7%) | 1673 (3.6%) | 2057 (2.8%) | 2243 (2.1%) | 2241 (1.6%) | 1930 (1.3%) |
| Male, heterosexual contact | 2809 (16.7%) | 3684 (18.0%) | 8685 (18.4%) | 13816 (18.6%) | 19665 (18.1%) | 24377 (17.4%) | 25278 (17.3%) |
| Female, heterosexual contact | 1578 (9.4%) | 3561 (17.4%) | 9516 (20.2%) | 16560 (22.3%) | 23254 (21.3%) | 27500 (19.7%) | 28744 (19.7%) |
| Male, other/unknown acquisition | 5350 (31.8%) | 1307 (6.4%) | 6559 (13.9%) | 10481 (14.1%) | 14545 (13.4%) | 17826 (12.8%) | 17658 (12.1%) |
| Female, other/unknown acquisition | 358 (2.1%) | 460 (2.3%) | 1191 (2.5%) | 1879 (2.5%) | 2394 (2.2%) | 2760 (2.0%) | 2885 (2.0%) |
| CD4 cell count (cells/mm 3 ) | |||||||
| 0–199 | 9882 (58.7%) | 8708 (42.6%) | 18093 (38.4%) | 21106 (28.4%) | 19892 (18.3%) | 17084 (12.2%) | 11084 (7.6%) |
| 200–349 | 4012 (23.8%) | 5089 (24.9%) | 11985 (25.4%) | 20729 (27.9%) | 27868 (25.6%) | 22697 (16.2%) | 15970 (10.9%) |
| 350–499 | 1743 (10.4%) | 3696 (18.1%) | 6869 (14.6%) | 11907 (16.0%) | 23744 (21.8%) | 30865 (22.1%) | 24083 (16.5%) |
| ≥500 | 1195 (7.1%) | 2955 (14.5%) | 8954 (19.0%) | 16546 (22.2%) | 30713 (28.2%) | 60483 (43.3%) | 83935 (57.5%) |
| Missing | NA (0.0%) | NA (0.0%) | 1223 (2.6%) | 4098 (5.5%) | 6729 (6.2%) | 8722 (6.2%) | 10924 (7.5%) |
| Prior AIDS-defining event | |||||||
| No | 10300 (61.2%) | 15494 (75.8%) | 34446 (73.1%) | 53835 (72.4%) | 81317 (74.6%) | 107762 (77.1%) | 113253 (77.6%) |
| Yes | 6532 (38.8%) | 4954 (24.2%) | 12678 (26.9%) | 20551 (27.6%) | 27629 (25.4%) | 32089 (23.0%) | 32743 (22.4%) |
| Hepatitis C virus (antibodies) | |||||||
| Negative | 8702 (51.5%) | 7350 (35.9%) | 20841 (44.2%) | 40803 (54.9%) | 70099 (64.3%) | 102630 (73.4%) | 114898 (78.7%) |
| Positive | 3514 (20.8%) | 2798 (13.7%) | 6628 (14.1%) | 10615 (14.3%) | 13607 (12.5%) | 15332 (11.0%) | 14828 (10.2%) |
| Missing | 4616 (27.4%) | 10300 (50.4%) | 19655 (41.7%) | 22968 (30.9%) | 25240 (23.2%) | 21889 (15.7%) | 16270 (11.1%) |
| Ethnicity | |||||||
| White | 10512 (62.5%) | 15713 (76.8%) | 29090 (61.7%) | 42829 (57.6%) | 62492 (57.4%) | 81166 (58.0%) | 84879 (58.1%) |
| Black | 3917 (23.3%) | 2246 (11.0%) | 7682 (16.3%) | 14094 (19.0%) | 20145 (18.5%) | 25245 (18.1%) | 27130 (18.6%) |
| Hispanic | 445 (2.6%) | 231 (1.1%) | 607 (1.3%) | 1321 (1.8%) | 2490 (2.3%) | 3788 (2.7%) | 4573 (3.1%) |
| Other | 531 (3.2%) | 703 (3.4%) | 1605 (3.4%) | 2541 (3.4%) | 3733 (3.4%) | 5075 (3.6%) | 5645 (3.9%) |
| Unknown | 1427 (8.5%) | 1555 (7.6%) | 8140 (17.3%) | 13601 (18.3%) | 20086 (18.4%) | 24577 (17.6%) | 23769 (16.3%) |
Each person can contribute data to multiple follow-up periods. Characteristics taken at earliest point each person contributed data to each period. For example, 2000–03 includes characteristics at 01/01/2000 for PWH who started antiretroviral therapy (ART) before that date, and at date of starting ART for those who started ART on 01/01/2001 to 31/12/2003.
The cause could not be classified for 3652 (21.7%) of the 16832 deaths, a percentage that remained quite consistent across the calendar year periods. The most common causes of death were AIDS (4203), non-AIDS non-hepatitis malignancy (2311) and cardiovascular (1403). Figure 1 shows changing proportions of deaths due to each cause over the calendar year periods. There was a large reduction in the proportion of AIDS-related mortality, from 49% during 1996–9 to 16% during 2016–20, with increases in the proportion of mortality due to cancers from 5% during 1996–99 to 19% during 2016–20 (Figure 1). There was a reduction in the proportion of AIDS-related deaths over calendar time for all age groups (Figure 2): the proportion of AIDS-related deaths was lowest for the oldest age group who, correspondingly, had the highest proportions of cardiovascular/heart-related and cancer-related mortality.
FIGURE 1:
Percentages of each cause of death category among PWH who died, by calendar year period of death.
*CNS: Central nervous system.
FIGURE 2:
Percentage of categorised causes of death among persons with HIV who died, by calendar year period of death, stratified by age group at death: a) 16–39; b) 40–59; c) ≥60 years
*CNS: Central nervous system.
Rates of all-cause mortality declined over time (average adjusted mortality rate ratio [aMRR] per period 0.85 [95% CI: 0.84–0.86] per period, figure 3). Supplementary table 2 (appendix page 5) shows estimated average aMRRs per period and for each period compared with 2000–03. Rates of each cause of death declined on average over time, except for central nervous system (average aMRR per period 1.03 [0.91–1.18]) and respiratory (1.04 [0.94–1.14]) mortality (figure 2). The largest average per period reductions were for AIDS-related mortality (0.81 [0.79–0.83]) and cardiovascular/heart-related mortality (0.83 [0.79–0.87]).
FIGURE 3:
Adjusted* cause-specific mortality rate ratios [MRRs] per period (1996–99, 2000–03, 2004–07, 2008–11, 2012–15, 2016–2020 ⱡ), with 95% confidence intervals.
*Adjusted for CD4 category and age group at the start of the period, whether they were ART-naïve when starting the period, and cohort.
CNS: Central Nervous System (Containing Parkinson’s and Alzheimer’s).
Non-AIDS non-hep cancer: Non-AIDS, non-hepatocellular carcinoma malignancies
ⱡ Calendar year period is modelled as a continuous variable; the MRRs can be interpreted as a per-period decrease.
Rates of all-cause mortality declined more steeply over calendar time in PWH aged 16–39 years (average aMRR per period 0.82 [95%CI: 0.80–0.85]) than in PWH aged 40–59 (0.85 [0.84–0.87]) or ≥60 (0.90 [0.88–0.93]) (Table 2). Similar patterns were seen for AIDS-related and cardiovascular/heart-related mortality. For liver mortality and for non-AIDS, non-hepatitis malignancy mortality, the steepest average declines were in the youngest age-group, with little evidence of a decline among PWH aged ≥60 years. There were larger reductions in all-cause mortality over calendar time among PWH who had started ART in the current period (average aMRR per period 0.81 [0.79–0.83]) than among those who had started ART during a prior period (0.88 [0.86–0.89]). This pattern was also seen for most categories of cause-specific mortality.
TABLE 2:
Adjusted* cause-specific mortality rate ratios [MRRs] with 95% CIs per period (1996–99, 2000–03, 2004–07, 2008–11, 2012–15, 2016–20 Ω) stratified by population subgroups.
| Subgroup | Overall (16832 deaths) |
AIDS (4203 deaths) |
Cardiovascular/heart† (1403 deaths) |
Central nervous systemⱡ (184 deaths) |
Liver, including hepatocellular carcinoma (1165 deaths) |
Non-AIDS infection (1043 deaths) |
Non-AIDS, non-hepatitis malignancy (2311 deaths) |
Other (1481 deaths) |
Respiratory (362 deaths) |
Substance use (380 deaths) |
Suicide/accident (648 deaths) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| All | 0.85 (0.84, 0.86) | 0.81 (0.79, 0.83) | 0.83 (0.79, 0.87) | 1.03 (0.91, 1.18) | 0.88 (0.84, 0.93) | 0.91 (0.86, 0.96) | 0.94 (0.90, 0.97) | 0.86 (0.82, 0.90) | 1.04 (0.94, 1.14) | 0.94 (0.86, 1.02) | 0.89 (0.82, 0.95) |
| Age (years)⊕ | |||||||||||
| 16–39 | 0.82 (0.80, 0.85) | 0.77 (0.73, 0.81) | 0.79 (0.70, 0.89) | 1.07 (0.83, 1.38) | 0.82 (0.74, 0.91) | 0.93 (0.85, 1.03) | 0.86 (0.78, 0.95) | 0.90 (0.84, 0.98) | 1.22 (0.98, 1.51) | 0.91 (0.79, 1.04) | 0.88 (0.78, 0.98) |
| 40–59 | 0.85 (0.84, 0.87) | 0.83 (0.80, 0.86) | 0.82 (0.77, 0.88) | 0.98 (0.81, 1.18) | 0.88 (0.83, 0.94) | 0.89 (0.83, 0.96) | 0.93 (0.89, 0.98) | 0.83 (0.78, 0.89) | 1.07 (0.93, 1.23) | 0.97 (0.86, 1.09) | 0.90 (0.82, 1.00) |
| ≥60 | 0.90 (0.88, 0.93) | 0.83 (0.77, 0.89) | 0.87 (0.80, 0.94) | 1.16 (0.88, 1.53) | 1.11 (0.93, 1.32) | 0.94 (0.83, 1.06) | 1.02 (0.95, 1.10) | 0.91 (0.81, 1.02) | 0.91 (0.76, 1.08) | 0.80 (0.53, 1.20) | 0.83 (0.66, 1.03) |
| Started ART during | |||||||||||
| A previous period | 0.88 (0.86, 0.89) | 0.81 (0.78, 0.84) | 0.87 (0.82, 0.92) | 1.15 (0.97, 1.35) | 0.94 (0.88, 1.00) | 0.93 (0.87, 1.00) | 0.95 (0.91, 0.99) | 0.90 (0.85, 0.95) | 1.03 (0.92, 1.15) | 0.95 (0.85, 1.06) | 0.89 (0.82, 0.97) |
| Current period | 0.81 (0.79, 0.83) | 0.81 (0.78, 0.84) | 0.76 (0.69, 0.82) | 0.85 (0.68, 1.06) | 0.76 (0.69, 0.83) | 0.87 (0.79, 0.95) | 0.91 (0.85, 0.98) | 0.76 (0.70, 0.82) | 1.05 (0.85, 1.30) | 0.90 (0.75, 1.08) | 0.87 (0.77, 0.97) |
| Sex/HIV acquisition group | |||||||||||
| MSM | 0.87 (0.85, 0.89) | 0.82 (0.78, 0.86) | 0.73 (0.66, 0.81) | 0.99 (0.76, 1.29) | 1.04 (0.89, 1.21) | 0.95 (0.83, 1.08) | 0.91 (0.85, 0.98) | 0.86 (0.77, 0.95) | 0.92 (0.76, 1.11) | 1.06 (0.83, 1.35) | 0.86 (0.76, 0.98) |
| Male, IDU | 0.96 (0.93, 0.99) | 0.81 (0.74, 0.88) | 0.94 (0.81, 1.10) | 1.05 (0.70, 1.56) | 0.98 (0.90, 1.07) | 0.97 (0.86, 1.11) | 1.05 (0.93, 1.18) | 1.06 (0.97, 1.15) | 1.16 (0.91, 1.48) | 0.98 (0.83, 1.15) | 1.08 (0.89, 1.30) |
| Female, IDU | 1.07 (1.00, 1.14) | 0.88 (0.75, 1.04) | 1.10 (0.79, 1.53) | 0.87 (0.37, 2.03) | 1.12 (0.92, 1.37) | 1.20 (0.96, 1.50) | 1.17 (0.92, 1.48) | 1.28 (1.06, 1.55) | 1.58 (1.00, 2.48) | 1.10 (0.85, 1.41) | 1.19 (0.80, 1.78) |
| Male, het contact | 0.91 (0.88, 0.93) | 0.85 (0.81, 0.91) | 0.91 (0.82, 1.02) | 1.39 (1.02, 1.89) | 1.04 (0.90, 1.21) | 0.92 (0.81, 1.05) | 0.93 (0.86, 1.01) | 0.84 (0.76, 0.93) | 1.09 (0.87, 1.37) | 0.92 (0.69, 1.23) | 0.87 (0.73, 1.03) |
| Female, het contact | 0.94 (0.90, 0.98) | 0.93 (0.86, 1.00) | 0.92 (0.79, 1.08) | 1.25 (0.89, 1.76) | 0.86 (0.71, 1.05) | 0.97 (0.82, 1.15) | 1.01 (0.90, 1.13) | 0.87 (0.75, 0.99) | 1.40 (0.90, 2.16) | 0.67 (0.36, 1.25) | 0.74 (0.56, 0.99) |
| Either sex, unknown | 0.81 (0.79, 0.82) | 0.75 (0.72, 0.79) | 0.84 (0.78, 0.91) | 0.88 (0.66, 1.18) | 0.84 (0.75, 0.93) | 0.90 (0.81, 1.00) | 0.95 (0.88, 1.03) | 0.81 (0.73, 0.90) | 1.02 (0.84, 1.24) | 1.05 (0.88, 1.25) | 0.96 (0.84, 1.09) |
| CD4 (cells/mm 3 ) | |||||||||||
| 0–199 | 0.85 (0.83, 0.87) | 0.83 (0.80, 0.85) | 0.78 (0.71, 0.85) | 1.01 (0.82, 1.25) | 0.89 (0.82, 0.97) | 0.91 (0.84, 0.99) | 1.01 (0.95, 1.09) | 0.79 (0.73, 0.86) | 1.19 (1.01, 1.40) | 1.09 (0.92, 1.30) | 0.92 (0.79, 1.06) |
| 200–349 | 0.87 (0.84, 0.89) | 0.75 (0.70, 0.81) | 0.91 (0.81, 1.01) | 1.06 (0.79, 1.44) | 0.91 (0.82, 1.02) | 0.91 (0.80, 1.03) | 0.97 (0.90, 1.04) | 0.83 (0.75, 0.92) | 1.06 (0.84, 1.35) | 0.96 (0.79, 1.18) | 0.89 (0.76, 1.04) |
| 350–499 | 0.83 (0.80, 0.86) | 0.70 (0.63, 0.79) | 0.81 (0.72, 0.90) | 1.00 (0.65, 1.55) | 0.75 (0.66, 0.87) | 0.95 (0.81, 1.13) | 0.88 (0.80, 0.96) | 0.86 (0.76, 0.97) | 0.94 (0.72, 1.23) | 0.78 (0.64, 0.96) | 0.76 (0.64, 0.89) |
| ≥500 | 0.81 (0.79, 0.84) | 0.70 (0.62, 0.78) | 0.80 (0.73, 0.88) | 1.03 (0.77, 1.37) | 0.80 (0.71, 0.91) | 0.82 (0.71, 0.94) | 0.86 (0.79, 0.92) | 0.83 (0.75, 0.92) | 0.75 (0.60, 0.95) | 0.77 (0.65, 0.90) | 0.90 (0.80, 1.02) |
| Prior AIDS-defining event | |||||||||||
| No | 0.83 (0.82, 0.84) | 0.72 (0.69, 0.76) | 0.79 (0.74, 0.84) | 1.00 (0.83, 1.20) | 0.85 (0.80, 0.91) | 0.91 (0.85, 0.98) | 0.91 (0.87, 0.95) | 0.85 (0.80, 0.90) | 0.95 (0.82, 1.09) | 0.87 (0.77, 0.97) | 0.88 (0.81, 0.95) |
| Yes | 0.89 (0.88, 0.91) | 0.88 (0.85, 0.91) | 0.89 (0.82, 0.96) | 1.09 (0.90, 1.32) | 0.94 (0.87, 1.02) | 0.93 (0.86, 1.01) | 0.98 (0.93, 1.04) | 0.90 (0.83, 0.96) | 1.15 (1.00, 1.32) | 1.06 (0.92, 1.23) | 0.92 (0.81, 1.05) |
| HCV (Antibodies) | |||||||||||
| Negative | 0.90 (0.88, 0.91) | 0.85 (0.82, 0.88) | 0.86 (0.80, 0.92) | 1.24 (1.02, 1.51) | 0.99 (0.89, 1.11) | 0.97 (0.90, 1.06) | 1.00 (0.95, 1.05) | 0.82 (0.76, 0.88) | 1.05 (0.91, 1.21) | 1.20 (1.01, 1.43) | 0.99 (0.90, 1.09) |
| Positive | 0.93 (0.90, 0.95) | 0.83 (0.77, 0.89) | 0.90 (0.80, 1.01) | 0.98 (0.66, 1.46) | 0.96 (0.89, 1.03) | 1.02 (0.92, 1.14) | 0.98 (0.90, 1.08) | 0.97 (0.90, 1.05) | 1.26 (1.03, 1.54) | 0.93 (0.82, 1.05) | 1.08 (0.93, 1.27) |
| Region | |||||||||||
| Europe | 0.87 (0.86, 0.89) | 0.83 (0.81, 0.85) | 0.83 (0.79, 0.88) | 1.11 (0.96, 1.29) | 0.91 (0.86, 0.97) | 0.93 (0.87, 0.98) | 0.94 (0.90, 0.98) | 0.87 (0.83, 0.91) | 1.04 (0.93, 1.16) | 0.86 (0.78, 0.96) | 0.89 (0.82, 0.96) |
| North America | 0.79 (0.77, 0.81) | 0.74 (0.71, 0.78) | 0.82 (0.75, 0.90) | 0.76 (0.57, 1.01) | 0.76 (0.68, 0.85) | 0.84 (0.75, 0.95) | 0.90 (0.82, 0.98) | 0.73 (0.64, 0.84) | 1.03 (0.84, 1.26) | 1.14 (0.97, 1.34) | 0.88 (0.78, 1.00) |
MSM: men who have sex with men. IDU: injecting drug use. Het contact: Heterosexual contact. HCV: Hepatitis C virus.
Adjusted for CD4 count and age group time-updated at the start of the period, whether they were ART-naïve when starting the period, and cohort.
Including stroke. ⱡ Including Parkinson’s and Alzheimer’s.
Calendar year period is modelled as a continuous variable; the MRRs can be interpreted as a per-period decrease.
Rates of all-cause mortality declined over calendar time for MSM, and for both men and women who acquired HIV through heterosexual sex, with a smaller decline among men who acquired HIV through IDU (average aMRR per period: 0.96 [95%CI: 0.93–0.99]), and there was an increase for women who acquired HIV through IDU (1.07 [1.00–1.14]). For both men and women who acquired HIV through IDU, there were increases over calendar time in rates of suicide/accident-related and respiratory-related mortality. For MSM, there was a large decrease over calendar time in rates of cardiovascular/heart-related mortality (average aMRR per period 0.73 [0.66–0.80]).
Declines in all-cause and cause-specific mortality were steeper for PWH with higher CD4 counts, without prior AIDS-defining events, and without HCV, compared to PWH with lower CD4 counts, with prior AIDS-defining events, and with HCV, respectively. PWH with the highest CD4 counts, experienced larger average declines in rates of all-cause mortality than those with lower CD4 counts. For many categories of cause-specific mortality, there were also larger average declines in mortality rates among PWH with higher CD4 counts compared to those with lower CD4 counts. Declines in rates of all-cause mortality over calendar time were larger among PWH with no prior AIDS-defining events than for those with prior AIDS-defining events, with similar patterns observed for most causes of death. Rates of all-cause mortality declined over calendar time in PWH without HCV, but there were less steep declines among PWH with HCV.
Declines in rates of all-cause mortality rates were less substantial for PWH living in Europe than for PWH living in North America. For most categories of cause-specific mortality, declines in rates of mortality over calendar time were greater in North America than in Europe. However, rates of substance use-related mortality declined on average among PWH in Europe but increased among PWH in North America.
In a sensitivity analysis omitting the two cohorts coded by the second algorithm (containing around half of the deaths), the main differences were that there were reductions in rates of central nervous system-related mortality over calendar time (average aMRR per period 0.75 [0.59–0.95]) and substance use-related mortality (0.88 [0.79–0.98]), which were not seen when including all cohorts. These differences may possibly have been due to one of these cohorts contributing little data in the 1996–99 period (supplementary table 3; appendix page 6). Results of the sensitivity analysis excluding deaths that happened after loss to follow-up is shown in supplementary table 4 (appendix page 7). The average aMRRs per period were similar to those in the analysis including these deaths, except for “Other” causes-of-death where the aMRR was 0.87 (0.83–0.91) excluding the deaths compared with 0.81 (0.77–0.86) including them.
Supplementary table 5 (appendix page 8) shows the results for a post-hoc analysis comparing MRRs for cause-specific mortality among PWH who acquired HIV through IDU living in Europe with those living in North America. For most causes of death there were wide confidence intervals, so, it was difficult to ascertain differences between PWH in Europe and North America.
DISCUSSION
There have been reductions in rates of all-cause mortality and of most major causes of death among PWH on ART between 1996 and 2020. For the overall population of PWH, reductions in mortality rates over calendar time were seen for each cause of mortality except for central nervous system, respiratory, and substance use deaths. The biggest reductions were for AIDS-related and cardiovascular/heart-related mortality. The reductions in mortality rates over calendar time differed between subgroups of PWH. Declines in heart/cardiovascular mortality were most marked among MSM, and may reflect improvements in post-cardiovascular disease care in the same periods among the general population, or perhaps the reduced toxicity of ART regimens used the studied time periods. For men who acquired HIV through IDU, the decreases in mortality rates were less than for other groups, while mortality rates increased in women who acquired HIV through IDU. For most causes of death, reductions in mortality rates were larger in North America than in Europe. However, there were reductions in substance use-related mortality over calendar time among PWH living in Europe and increases in North America.
Other literature has investigated changes in cause-specific mortality over time among PWH on ART. A Spanish population-based cohort of PWH, including those not on ART, found that all-cause mortality decreased from 1999 to 2018, with the decrease driven by AIDS-related mortality, whilst non-AIDS mortality remained stable(15). The largest decrease in mortality was among PWH who had had AIDS-defining events. HIV-related mortality decreased among PWH without a prior AIDS-defining event, but the number of such deaths was low(15). A study among PWH on ART in British Columbia, Canada, observed reductions in all-cause, AIDS-related, cardiovascular, liver-related, and suicide-related mortality over calendar time from 2001 and 2012(16), whilst the HIV Atlanta VA Cohort Study found declines in all-cause, AIDS-related, and non-AIDS-related mortality rates during the ART era(17). A previous ART-CC study examined changes from 1996 to 2015 in cause-specific mortality by ART start year, which is comparable to the changes in the present study among PWH who were ART-naive in each follow-up period(3). That study also observed reductions over time in AIDS-related mortality, non-AIDS infection, non-AIDS cancers, cardiovascular, suicide/accident-related, and other mortality(3). Our finding of an increase in substance use-related mortality in North America and a decrease in Europe perhaps reflects an increase in opioid use in the USA(18), where opioid-related mortality is over 10 times higher than in the European Union(19). Unlike in the USA(20), in some European countries there have been reductions in the prevalence of injecting drug use(21), although not necessarily opiate use(19). Additionally, opiate substitution therapy coverage is higher in Western European countries than in North America, which may contribute to this difference(22). Our finding that rates of respiratory-related mortality increased among people with HCV is likely due to the high prevalence of HCV among people who acquired HIV through IDU, among whom there was also an increase in rates of respiratory-related mortality.
The main strengths of this study are the large sample size, geographical diversity, representativeness of the included PWH, and the availability of data on cause-specific mortality, classified according to a common protocol(14). However, causes of death were classified retrospectively and without complete patient histories, so there is likely to have been misclassification compared with classifications based on full medical history. Autopsy, which is becoming less common over time(23), remains the gold standard for classifying causes of death, and clinical classifications may not correlate well with those from autopsy reports(24, 25). For some PWH there was information on a number of comorbidities and/or potential causes of death, which could lead to the death being classified as of unknown cause or unclassifiable. It is possible that changes in how causes of death are captured on death certificates have changed over time. In particular, HIV was previously more likely to be mentioned as a cause on a death certificate because the person who died had HIV, even if it was irrelevant to the death. The CoDe processes and rules were set-up to minimise the impact of this, by accounting for recent CD4 counts and AIDS diagnoses rather than relying solely on death certificate information. For our models of cause-specific mortality, we adjusted for time updated CD4 counts and age to capture major changes in HIV prognostic markers for mortality. Data on some established risk factors for chronic diseases, particularly smoking and alcohol consumption(26), were not available for all patients and therefore we were unable to estimate their contributions to cause-specific mortality. Other potentially important patient characteristics, including socio-economic factors such as education, poverty, and homelessness, were not routinely collected across the cohorts, so we were unable to include them in our analyses. Data on HIV acquisition risk group were available for all but one cohort, while data for one cohort were unavailable in 1996–99. The cohorts are from various countries with different health systems, population characteristics, and migration patterns, so pooling such data may obscure within-country patterns, although we conducted analyses stratified by region. Loss-to-follow-up was high in some cohorts using our definition of a gap of at least a year between the person’s last visit and the cohort-specific administrative censoring date, particularly for hospital-based cohorts where a large percentage of their patients subsequently moved to general-practitioner-based care (supplementary table 1). However, we censored at the date of death when a participant was lost-to-follow-up but information on death was later available through linkage to registries, so that in such situations loss to follow-up did not affect ascertainment of deaths. Results of sensitivity analyses excluding deaths after loss to follow-up were similar. Most of our findings are likely generalisable to other high-income countries such as Australia and Japan, although some particular trends may be more context-specific, for example, regarding substance use deaths.
The trends in cause-specific mortality captured here should assist policy makers in targeting improvements in the care of PWH towards conditions that are amenable to interventions and have the biggest influence on mortality. The per-period reductions in cause-specific mortality for PWH on ART are likely partially due to changes in treatment guidelines meaning that people start ART sooner after diagnosis(27), as well as more effective and less toxic regimens becoming available(11), and better care in general for PWH(28). Whilst mortality among PWH on ART has decreased(3), there is still higher mortality among PWH compared to the general population, due both to the consequences of HIV infection and to a higher prevalence of comorbidities and risk behaviours among PWH(15, 29, 30). Whilst there were reductions in rates of non-AIDS related mortality, such as cancer and cardiovascular disease, non-AIDS deaths make up an increasingly large proportion of mortality among PWH. Expanding access to prevention, screening, and treatment of these conditions is required to close the comorbidity prevalence gap between adults with and without HIV and funders should recognise this. Progress in reducing cause-specific mortality has not been evenly spread across subgroups of PWH, often with the most marginalised populations experiencing the least benefits. The least evidence of reductions in mortality for PWH, was for people who acquired HIV through IDU, particularly women. People with histories of substance use conditions have higher rates of homelessness and other comorbidities(31, 32) and often experience additional barriers and stigma whilst attempting to access care(33). This indicates that targeted interventions, such as addressing social determinants of health and bringing comorbidity care to needle and syringe dispensing locations, are required for people who acquired HIV through IDU. That there was a decrease in substance abuse-related mortality in Europe, but an increase in North America, which saw a large increase in opioid use during this period(18), shows that context-specific analyses are also required to understand the epidemiology of mortality among PWH in different settings and that the most impactful interventions will likely vary by location.
Supplementary Material
Research in context.
Evidence before this study
We searched PubMed for English-language studies published up to 27th February 2023 that had estimated trends in cause-specific mortality among persons with HIV (PWH) on antiretroviral therapy (ART) in Europe or North America. Our searches used the terms “((cause-specific mortality) OR (causes of mortality)) AND (HIV) AND (trend)”. Studies of HIV cohorts in Northern Spain, British Columbia (Canada), Atlanta (USA), and previous cross-country analyses by the Antiretroviral Therapy Cohort Collaboration have found that rates of all-cause, AIDS-related, and non-AIDS-related mortality have declined over time among PWH on ART. However, previous analyses have also found that prognosis for all-cause mortality tends to vary substantially across different subgroups of PWH.
Added value of this study
Our study, using data up to 2020 from multiple countries in Europe and North America, is the largest and most detailed investigation to date of rates of cause-specific mortality among PWH on ART. We analysed data, combined from 18 HIV cohort studies, on 189,301 PWH and classified causes for 13,180 of 16,832 deaths. Based on this large dataset we investigated trends in cause-specific mortality both overall and within 21 subgroups of PWH defined by demographic and clinical characteristics. Rates of most categories of cause-specific mortality declined between 1996 and 2020: the largest reductions were in rates of AIDS-related and cardiovascular/heart-related mortality. Rates of all-cause mortality declined over calendar time for MSM, and for both men and women who acquired HIV through heterosexual sex but did not decline in women who acquired HIV through injection drug-use. In these groups, rates of substance-use, suicide/accident-related, and respiratory-related mortality increased over time.
Implications of all the available evidence
Improvements in ART and HIV care have led to reductions over time in rates of most major causes of death among PWH on ART, especially AIDS-related deaths. Unequal reductions in mortality among different populations of PWH indicate that interventions should be targeted at high-risk groups, substance use, and comorbidities to further increase life expectancy among PWH.
Acknowledgements:
We would like to thank our funders (US NIAAA) as well as all the patients and the clinical teams associated with the participating cohort studies.
Funding:
The ART-CC is funded by the US National Institute on Alcohol Abuse and Alcoholism (U01-AA026209). JACS is funded by National Institute for Health Research Senior Investigator award NF-SI-0611-10168. AT is funded by the Wellcome Trust under a Sir Henry Wellcome Postdoctoral Fellowship (222770/Z/21/Z).
Funding for the individual ART-CC cohorts included in this analysis was from Alberta Health, Gilead, ANRS (France REcherche Nord&Sud Sida-hiv Hépatites), the French Ministry of Health, the Austrian Agency for Health and Food Safety (AGES), Stichting HIV Monitoring, the Dutch Ministry of Health, Welfare and Sport through the Centre for Infectious Disease Control of the National Institute for Public Health and the Environment, the TP-HIV by the German Centre for Infection Research (DZIF) (NCT02149004), the Instituto de Salud Carlos III through the Red Temática de Investigación Cooperativa en Sida (RD06/006, RD12/0017/0018 and RD16/0002/0006) as part of the Plan Nacional I + D + i and co-financed by ISCIII-Subdirección General de Evaluación and the Fondo Europeo de Desarrollo Regional (FEDER), ViiV Healthcare, Preben og Anna Simonsens Fond, ANRS-Maladies infectieuses émergentes, Institut National de la Santé et de la Recherche Médicale (INSERM), BMS, Janssen, MSD, the US National Institute on Alcohol Abuse and Alcoholism (U01-AA026230), the Spanish Ministry of Health, the Swiss National Science Foundation (grant 33CS30_134277), CFAR Network of Integrated Clinical Systems (1R24 AI067039-1, P30-AI-027757), the US Department of Veterans Affairs, the US National Institute on Alcohol Abuse and Alcoholism (U01-AA026224, U01-AA026209, U24-AA020794), the VHA Office of Research and Development, US National Institute of Allergy and Infectious Diseases (Tennessee Center for AIDS Research: P30 AI110527).
Transparency declarations:
AI has received financial compensation for lectures, educational activities, consultancy work, as well as funds for research, from Gilead Sciences, Janssen-Cilag, Merck Sharp & Dohme, and ViiV Healthcare. VP has received honoraria in the last 3 years from ad hoc membership of national HIV advisory boards, Merck, Gilead, and ViiV. RZ has not received honoraria in the last 3 years. PR, through his institution, has received scientific grant support for investigator-initiated studies from Gilead Sciences, Janssen Pharmaceuticals Inc, Merck & Co and ViiV Healthcare, and has served on scientific advisory boards for Gilead Sciences, ViiV Healthcare, and Merck & Co, honoraria for which were all paid to his institution. JB reports honoraria for advice or public speaking from GILEAD, GSK, JANSSEN, MSD, and ViiV Healthcare; and grants from GILEAD, MSD, and ViiV Healthcare. MJG has received honoraria in the last 3 years from ad hoc membership of national HIV advisory boards, Merck, Gilead, and ViiV. HC has received research grant funding from ViiV, NIH, and AHRQ paid to their institution, and sits on the NIH Office of AIDS Research Advisory Council. CW reports honoraria for advice or public speaking from Abbott, GILEAD, JANSSEN, MSD, Pfizer, and ViiV Healthcare. KK, TRS, SMI, SG, SA, RT, MH, MJS, JACS, AT, JLG, KM, and AdM report no conflicts of interest.
Data sharing statement:
Due to the data sharing agreements between individual cohorts and ART-CC, the data collected for this study cannot be shared. Data are owned by the individual cohorts and those wishing to access these data should contact the individual cohorts, details of which are given in the appendix.
References
- 1.Krentz HB, Kliewer G, Gill MJ. Changing mortality rates and causes of death for HIV-infected individuals living in Southern Alberta, Canada from 1984 to 2003. HIV Med. 2005;6(2):99–106. [DOI] [PubMed] [Google Scholar]
- 2.UNAIDS. AIDSinfo 2023. [Available from: https://aidsinfo.unaids.org/.
- 3.Trickey A, May MT, Vehreschild JJ, Obel N, Gill MJ, Crane HM, et al. Survival of HIV-positive patients starting antiretroviral therapy between 1996 and 2013: a collaborative analysis of cohort studies. Lancet Hiv. 2017;4(8):E349–E56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Althoff KN, Stewart CN, Humes E, Zhang JB, Gerace L, Boyd CM, et al. The shifting age distribution of people with HIV using antiretroviral therapy in the United States. Aids. 2022;36(3):459–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Marty L, Diawara Y, Rachas A, Grabar S, Costagliola D, Supervie V. Projection of age of individuals living with HIV and time since ART initiation in 2030: estimates for France. J Int Aids Soc. 2022;25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Justice AC. HIV and aging: time for a new paradigm. Curr HIV/AIDS Rep. 2010;7(2):69–76. [DOI] [PubMed] [Google Scholar]
- 7.Alejos B, Hernando V, Iribarren J, Gonzalez-Garcia J, Hernando A, Santos J, et al. Overall and cause-specific excess mortality in HIV-positive persons compared with the general population: Role of HCV coinfection. Medicine. 2016;95(36). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ingle SM, May MT, Gill MJ, Mugavero MJ, Lewden C, Abgrall S, et al. Impact of Risk Factors for Specific Causes of Death in the First and Subsequent Years of Antiretroviral Therapy Among HIV-Infected Patients. Clin Infect Dis. 2014;59(2):287–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lundgren D, Babiker AG, Gordin F, Emery S, Sharma S, Avihingsanon AC, et al. Initiation of Antiretroviral Therapy in Early Asymptomatic HIV Infection. New Engl J Med. 2015;373(9):795–807. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eholie SP, Badje A, Kouame GM, N’takpe JB, Moh R, Danel C, et al. Antiretroviral treatment regardless of CD4 count: the universal answer to a contextual question. Aids Res Ther. 2016;13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Vitoria M, Rangaraj A, Ford N, Doherty M. Current and future priorities for the development of optimal HIV drugs. Curr Opin Hiv Aids. 2019;14(2):143–9. [DOI] [PubMed] [Google Scholar]
- 12.Brooks KM, Sherman EM, Egelund EF, Brotherton A, Durham S, Badowski ME, et al. Integrase Inhibitors: After 10 Years of Experience, Is the Best Yet to Come? Pharmacotherapy. 2019;39(5):576–98. [DOI] [PubMed] [Google Scholar]
- 13.May MT, Ingle SM, Costagliola D, Justice AC, de Wolf F, Cavassini M, et al. Cohort Profile: Antiretroviral Therapy Cohort Collaboration (ART-CC). Int J Epidemiol. 2014;43(3):691–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kowalska JD, Friis-Moller N, Kirk O, Bannister W, Mocroft A, Sabin C, et al. The Coding Causes of Death in HIV (CoDe) Project Initial Results and Evaluation of Methodology. Epidemiology. 2011;22(4):516–23. [DOI] [PubMed] [Google Scholar]
- 15.Fontela C, Aguinaga A, Moreno-Iribas C, Reparaz J, Rivero M, Gracia M, et al. Trends and causes of mortality in a population-based cohort of HIV-infected adults in Spain: comparison with the general population. Sci Rep-Uk. 2020;10(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Cheung CC, Ding E, Sereda P, Yip B, Lourenco L, Barrios R, et al. Reductions in all-cause and cause-specific mortality among HIV-infected individuals receiving antiretroviral therapy in British Columbia, Canada: 2001–2012. Hiv Medicine. 2016;17(9):694–701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Vyas KJ, Marconi VC, Moanna A, Rimland D, Guest JL. Trends in Cause-Specific Mortality Among Veterans With HIV: A 35-Year (1982–2016) Analysis of the HIV Atlanta VA Cohort Study. Jaids-J Acq Imm Def. 2023;92(1):17–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Singh GK, Kim IE, Girmay M, Perry C, Daus GP, Vedamuthu IP, et al. Opioid Epidemic in the United States: Empirical Trends, and A Literature Review of Social Determinants and Epidemiological, Pain Management, and Treatment Patterns. Int J MCH AIDS. 2019;8(2):89–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kalkman GA, van den Brink W, Pierce M, Atsma F, Vissers KCP, Schers HJ, et al. Monitoring Opioids in Europe: The Need for Shared Definitions and Measuring Drivers of Opioid Use and Related Harms. Eur Addict Res. 2022;28(3):231–40. [DOI] [PubMed] [Google Scholar]
- 20.Bradley H, Hall EW, Asher A, Furukawa NW, Jones CM, Shealey J, et al. Estimated Number of People Who Inject Drugs in the United States. Clin Infect Dis. 2023;76(1):96–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Degenhardt L, Peacock A, Colledge S, Leung J, Grebely J, Vickerman P, et al. Global prevalence of injecting drug use and sociodemographic characteristics and prevalence of HIV, HBV, and HCV in people who inject drugs: a multistage systematic review. Lancet Glob Health. 2017;5(12):e1192–e207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Larney S, Peacock A, Leung J, Colledge S, Hickman M, Vickerman P, et al. Global, regional, and country-level coverage of interventions to prevent and manage HIV and hepatitis C among people who inject drugs: a systematic review. Lancet Global Health. 2017;5(12):E1208–E20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Weber R, Ruppik M, Rickenbach M, Spoerri A, Furrer H, Battegay M, et al. Decreasing mortality and changing patterns of causes of death in the Swiss HIV Cohort Study. Hiv Medicine. 2013;14(4):195–207. [DOI] [PubMed] [Google Scholar]
- 24.Cox JA, Lukande RL, Lucas S, Nelson AM, Van Marck E, Colebunders R. Autopsy Causes of Death in HIV-Positive Individuals in Sub-Saharan Africa and Correlation with Clinical Diagnoses. Aids Rev. 2010;12(4):183–94. [PubMed] [Google Scholar]
- 25.Roulson J, Benbow EW, Hasleton PS. Discrepancies between clinical and autopsy diagnosis and the value of post mortem histology; a meta-analysis and review. Histopathology. 2005;47(6):551–9. [DOI] [PubMed] [Google Scholar]
- 26.Petoumenos K, Law MG. Smoking, alcohol and illicit drug use effects on survival in HIV-positive persons. Curr Opin Hiv Aids. 2016;11(5):514–20. [DOI] [PubMed] [Google Scholar]
- 27.Gunthard HF, Saag MS, Benson CA, del Rio C, Eron JJ, Gallant JE, et al. Antiretroviral Drugs for Treatment and Prevention of HIV Infection in Adults 2016 Recommendations of the International Antiviral Society-USA Panel. Jama-J Am Med Assoc. 2016;316(2):191–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mugavero MJ. Elements of the HIV Care Continuum: Improving Engagement and Retention in Care. Top Antivir Med. 2016;24(3):115–9. [PMC free article] [PubMed] [Google Scholar]
- 29.Croxford S, Kitching A, Desai S, Kall M, Edelstein M, Skingsley A, et al. Mortality and causes of death in people diagnosed with HIV in the era of highly active antiretroviral therapy compared with the general population: an analysis of a national observational cohort. Lancet Public Health. 2017;2(1):E35–E46. [DOI] [PubMed] [Google Scholar]
- 30.Trickey A, van Sighem A, Stover J, Abgrall S, Grabar S, Bonnet F, et al. Parameter estimates for trends and patterns of excess mortality among persons on antiretroviral therapy in high-income European settings. Aids. 2019;33:S271–S81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hotton A, Mackesy-Amiti ME, Boodram B. Trends in homelessness and injection practices among young urban and suburban people who inject drugs: 1997–2017. Drug Alcohol Depen. 2021;225. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lim J, Pavalagantharajah S, Verschoor CP, Lentz E, Loeb M, Levine M, et al. Infectious diseases, comorbidities and outcomes in hospitalized people who inject drugs (PWID) infections in persons who inject drugs. Plos One. 2022;17(4). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Paquette CE, Syvertsen JL, Pollini RA. Stigma at every turn: Health services experiences among people who inject drugs. Int J Drug Policy. 2018;57:104–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Due to the data sharing agreements between individual cohorts and ART-CC, the data collected for this study cannot be shared. Data are owned by the individual cohorts and those wishing to access these data should contact the individual cohorts, details of which are given in the appendix.