Table 1.
First author | Location | Years | Median baseline age | % female | % on ART a | Comparison against general population | AIDS‐related mortality definition | Quality assessment score | Study design |
---|---|---|---|---|---|---|---|---|---|
Croxford [41] | England and Wales | 1997–2012 | 34 (IQR: 28–41) | 36% | 66% | Age‐ and sex‐matched general population mortality rates | Death reports from HIV clinicians through routine surveillance and death auditing | 9 | Cohort |
De Coninck [42] | Sweden | 1996–2011 | Mean: 37.6 | 36% | >56% | 1:2 matching with HIV‐negative people from population register by age, sex and birth region | ICD9/ICD10. Only rates for non‐AIDS cancer, cardiovascular disease and violent deaths given; the remainder were assumed to be AIDS‐related | 8 | Cohort |
Eyawo [29] | British Columbia, Canada | 1996–2012 | 38 (IQR: 32–46) | 20% | 74% | Direct comparison with a randomly sampled (age/sex‐matched) HIV‐negative control group in region who had personal health numbers | ICD9/ICD10 categories | 8 | Cohort |
Fontela [43] | Navarra, Spain | 1999–2018 |
1999–2003: 37 (IQR: 34–41) 2014–2018: 47 (IQR: 40–52) |
33% |
1999–2003: 72% 2014–2018: 88% |
Age‐ and sex‐matched general population mortality rates | HIV‐related ICD10s: B20‐B24 and R75 | 8 | Cohort |
Karnite [26] | Latvia | 1987–2010 | NA | NA | NA | Age‐matched general population mortality rates | HIV‐related ICD10s: B20‐B24 | 6 | Cohort |
Nishijima [44] | Tokyo, Japan | 2005–2016 | 36 (IQR: 32–46) | 8% | 32% at enrolment | Age‐ and sex‐matched general population mortality rates | Adapted Cause of Death (CoDe) project protocol | 8 | Cohort |
Park [45] | South Korea | 2004–2019 | Mean: 40.3 | 11% | 91% | Age‐ and sex‐matched general population mortality rates | ICD10 categories | 8 | Cohort |
Trickey [27] | Europe (Multi‐country) | 2000–2015 | 37 (IQR: 31–45) | 27% | 100% | Age‐, sex‐ and country‐matched general population mortality rates | Adapted Cause of Death (CoDe) project protocol | 7 | Cohort |
Abbreviations: ICD, International Classification of Disease; IQR, interquartile range; NA, not available.
People off ART at start of follow‐up could have subsequently started ART.