Table 3 Triangulation of included studies with reviews of the broader observational literature
Outcome | Study/ies first author (year) | Simplified summary of findings | Consistent with broader literature? | Details |
|---|---|---|---|---|
Prostate cancer | Dickerman (2018) | -Reverse J-shaped relationship | No | A 2016 meta-analysis found a monotonically increasing relationship [60] |
All-cause mortality | Sipila (2016) | -Monotonically increasing | Yes | A 2016 meta-analysis found a monotonically increasing relationship [16] |
Diabetes biomarkers: | ||||
–FBG | Peng (2016) | -Positive linear relationship | Yes | A 2017 review found results consistent with a positive linear relationship [61] |
–Glucose tolerance, insulin sensitivity and glycated haemoglobin | Peng (2016) | -Positive linear relationships for P2hBG (glucose tolerance) and HOMA-IR (insulin sensitivity) -Lack of relationship for HbA1c (glycated haemoglobin) and HOMA-beta (insulin sensitivity) | N/A | A 2017 review found that, for the other diabetes biomarkers included in this review, there was not enough/consistent evidence from which to draw conclusions [61] |
CVD biomarkers: | ||||
–Blood pressure | Silverwood (2014) Peng (2019) | -Non-linear relationship (small protective effect of light drinking) for SBP -Positive linear relationships for both SBP and DBP | No Yes | A 2018 meta-analysis found no protective effect for hypeternsion [62] |
–Obesity anthropometrics | Silverwood (2014) Peng (2019) | -Non-linear relationship (small protective effect of light drinking) for BMI and waist circumference -Positive linear relationships for BMI, waist circumference, and hip circumference, with no relationship for waist-to-hip ratio | Mixed Mixed | A 2011 review found mixed evidence on the relationship between alcohol and weight/measures of abdominal adiposity, with heavy drinking generally associated with increased values, but moderate consumption either associated with lower values or not associated at all (suggested that discrepancies may depend on type of alcohol consumed) [63] |
–Inflammatory markers | Silverwood (2014) | -Non-linear relationship (small protective effect of light drinking) for CRP but a positive linear relationship for IL-6 | N/A | A 2021 review of studies from the previous decade found no relevant longitudinal observational studies that were capable of detecting non-linear relationships [64] |
–Lipids: HDL-C | Silverwood (2014) Peng (2019) Vu (2016) | -No relationship for HDL-C -No relationship for HDL-C -No relationship for HDL-C, or HDL-3C, but an inverted reverse J-shaped relationship for HDL-2C (a beneficial effect) | No No Mixed | A 2021 review of studies from the previous decade found just one longitudinal observational study, which reported a reverse J-shaped relationship in decreases in HDL-C from baseline [64] |
–Lipids: other | Silverwood (2014) Peng (2019) Vu (2016 | -No relationship for TG, but a non-linear relationship (small protective effect of light drinking) for non-HDL-C -Positive linear relationships for TG, non-HDL-C, and TC -Reverse J-shaped relationships for LDL-C, TG and TC, no relationship for Lp(a), and monotonically decreasing relationships for sdLDL-C and apoB | N/A | A 2021 review of studies from the previous decade found no relevant longitudinal observational studies [64] |
CVD events/diagnoses: | ||||
–Myocardial infarction | Ilomaki (2011) Millwood (2019) | -J-shaped relationship -No relationship | No | A 2017 review found alcohol was detrimental for myocarditis [2], but no reviews on myocardial infarction were found |
–Stroke | Kadlecova (2015) Millwood (2019) | -Reverse J-shaped relationship for stroke and TIA -Monotonically increasing relationships for ischaemic stroke, intracerebral haemorrhage and total stroke | Mixed | A 2016 meta-analysis found J-shaped relationships for ischaemic stroke, but monotonically increasing relationships for intracerebral haemorrhage and subarachnoid haemorrhage [65] |
–Total coronary heart disease | Millwood (2019) | -No relationship | No | A 2016 meta-analysis found a U-shaped relationship [66] |
–Circulatory system diagnoses | Ropponen (2014) | -No clear functional form | N/A | Most reviews evaluate CVD sub-conditions, finding divergent functional forms |
Dementia | Handing (2015) | -J-shaped relationship | Yes | A 2021 review found most recent evidence is consistent with a J-shaped relationship [67] |
Mental health | Gemes (2019) Samuelsson (2013) | -U−/J-shaped relationship -Non-linear relationship with abstainers at increased risk over light frequent drinkers | Yes | A 2020 meta-analysis found a J-shaped relationship with depressive symptoms [68] |
HIV seroconversion | Sander (2013) | -Monotonically increasing relationship | N/A | There is a lack of dose-response information on alcohol’s relationship with HIV [69, 70] |
Musculoskeletal disorders | Pietikainen (2011) Ropponen (2014) Ropponen (2011) | -Monotonically increasing relationship -No clear functional form -No clear functional form, but abstainers have lowest risk | N/A | No reviews on the relationship between alcohol and musculoskeletal disorders were found |