Background
Methods
Exposure and mortality data
Risk relations and calculating the alcohol-attributable fraction (AAF)
Application of the alcohol-attributable fractions to causes of death
Uncertainty
Differences of the current analyses to the WHO 2014 GSRAH and their scientific basis
New algorithms | Algorithms from GSRAH | |
---|---|---|
Risk relation curve up to 100 g/day | based on [5]; RR ≥ 1 after 60 g/day | based on [5]; RR ≥ 1 after 60 g/day |
Risk relation curve for 100 g+/day | Not included, set to 1 | |
Impact of binge drinking in persons who drink on average less than 60 g/day | Yes [6] | Yes [6] |
Modelling the sick quitter effect in former drinkers | For all countries, ex-drinkers were modelled with the increased risk of all-cause mortality RRs | |
Age-specific risk relations | Only for global |
-
4) The effect of former drinking is detrimental according to meta-analyses [8, 36, 37]; however, most of these meta-analyses originated in high-income countries, where drinking over a lifetime is the norm [19], and where a substantial proportion of people in later adulthood quit drinking for health reasons [38‐40]. As a consequence, we applied the specific RR for former drinkers [36] to the total population of former drinkers only in these regions, and artificially capped the prevalence of former drinkers for other regions (see Additional file 1: Web-Appendix 2 for details and Additional file 1: Web Appendix 3 for the countries within each region).
-
The reasoning for capping was mainly because we wanted to be conservative and restrict the detrimental impact of former drinking to region, where we actually have evidence of the phenomenon (but see [41]). Unless we find better evidence for increased health risks of former drinkers due to alcohol in low- to mid income countries, this seems to be the cautious choice. This decision is also in line with the general rule of the comparative risk assessments for alcohol to always choose the more conservative option (e.g., [18]).
-
5) Additionally, the impact of alcohol on all-cause mortality and on IHD and IS is age-specific [42]. Accordingly, the alcohol risk for IHD and IS was modelled based on age-specific RRs estimated based on the observed effect modification of age on the relationship between alcohol and IHD and IS [6, 8].
Ethical approval
Results
Relative risk relations between alcohol consumption and ischaemic diseases
Resulting alcohol-attributable global mortality burden from ischaemic diseases
Disease and region | Men | Women | Total | ||||
---|---|---|---|---|---|---|---|
Deaths | 95 % CI | Deaths | 95 % CI | Deaths | 95 % CI | ||
Ischaemic heart disease | |||||||
Africa | −3,306 | (−9,609 to 2,997) | 2,302 | (−1,800 to 6,405) | −1,004 | (−8,525 to 6,517) | |
Americas | −2,433 | (−27,910 to 23,043) | 3,132 | (−12,914 to 19,178) | 699 | (−29,410 to 30,808) | |
Eastern-Mediterranean | 1,211 | (−535 to 2,957) | 1,511 | (728 to 2,295) | 2,722 | (808 to 4,636) | |
Europe without Russia and surrounding countriesb
| 910 | (−35,567 to 37,387) | 9,290 | (−15,998 to 34,578) | 10,200 | (−34,185 to 54,586) | |
Russia and surrounding countriesb
| 72,656 | (21,148 to 124,164) | 99,080 | (37,263 to 160,896) | 171,736 | (91,193 to 252,278) | |
South-East Asia | −1,377 | (−23,721 to 20,966) | 10,710 | (4,519 to 16,900) | 9,333 | (−13,852 to 32,518) | |
Western-Pacific | −13,046 | (−67,728 to 41,637) | 23,919 | (−15,270 to 63,107) | 10,873 | (−56,402 to 78,148) | |
Totalc
| 54,499 | (22,551 to 86,446) | 150,121 | (122,037 to 178,206) | 204,620 | (162,064 to 247,176) | |
Ischaemic stroke | |||||||
Africa | 1,037 | (−848 to 2,922) | −640 | (−3,450 to 2,171) | 397 | (−2,987 to 3,782) | |
Americas | 3,375 | (−3,755 to 10,506) | −5,626 | (−12,292 to 1,040) | −2,251 | (−12,012 to 7,511) | |
Eastern-Mediterranean | 871 | (−581 to 2,323) | 132 | (−769 to 1,034) | 1,003 | (−706 to 2,712) | |
Europe without Russia and surrounding countriesb
| 8,691 | (−5,981 to 23,363) | −9,399 | (−24,203 to 5,405) | −708 | (−21,551 to 20,135) | |
Russia and surrounding countriesb
| 22,374 | (11,573 to 33,174) | 46,155 | (34,239 to 58,072) | 68,529 | (52,467 to 84,591) | |
South-East Asia | 7,264 | (−1,694 to 5,501) | 1,766 | (−5,836 to 9,368) | 9,030 | (−394 to 18,453) | |
Western-Pacific | 20,239 | (−2,534 to 43,011) | −9,284 | (−40,890 to 22,322) | 10,954 | (−28,001 to 49,910) | |
Totalc
| 64,040 | (53,211 to 74,868) | 23,042 | (9,481 to 36,602) | 87,082 | (69,730 to 104,443) |
Disease | Men | Women | Total | |
---|---|---|---|---|
Conduction disorder and other dysrhythmias | 7,373 | 7,835 | 15,208 | |
Hypertension | 70,051 | 24,664 | 94,714 | |
Haemorrhagic stroke | 245,930 | 132,828 | 378,757 | |
Ischaemic heart disease | 54,499 | 150,121 | 204,620 | |
Ischaemic stroke | 64,040 | 23,042 | 87,082 | |
Total CVD current analysis | 441,893 | 338,490 | 780,381 | |
From [19] | ||||
Ischaemic heart disease
|
111,755
|
417,469
|
529,225
| |
Ischaemic stroke
|
64,390
|
45,979
|
110,369
| |
Total CVD
|
499,499
|
628,775
|
1,128,273
|