Background
Methods
Protocol
Eligibility criteria
-
Population: People aged 13 years or older, with any type of diabetes (type 1, type 2, or gestational) or pre-diabetes
-
Intervention/exposure: Consumption of cannabis (all constituents) ingested in any product form (inhaled, brewed, eaten) for recreational purposes
-
Comparator: Non-exposure to cannabis
-
Outcomes: Metabolic factors related to diabetes (e.g., HbA1c, blood glucose, weight, blood pressure, dyslipidemia), including complications (e.g., [DKA]) or diabetes self-management behaviours (e.g., monitoring blood glucose, taking medication appropriately, exercise, attending medical appointments)
-
Study design: Any experimental, quasi-experimental or observational design with the exception of case studies/case series and knowledge syntheses, such as systematic or literature reviews
Literature search
Study selection, data extraction, quality assessment
Data synthesis
Results
Description of studies
First author, year | Country | Study design | Setting | Study period | Funding source |
---|---|---|---|---|---|
Akturk, 2019 [21] | USA | Cross-sectional | Single site | June 2017 and January 2018 | Not reported |
Helgeson, 2016 [22] | USA | Cohort | Single site | Not reported | National Institutes of Health |
Hogendorf, 2016 [23] | Poland | Cross-sectional | Multiple sites | May to June 2013 | Medical University of Lodz |
Thurheimer-Cacciotti, 2017 [24] | USA | Cross-sectional | Not reported | Not reported | American Diabetes Association |
Winhusen, 2018 [25] | USA | Case-control | City-wide | Not reported | Not reported |
Wisk, 2018 [26] | USA; Canada | Cross-sectional | Multinational | Not reported | Not reported |
First author, year | Sample size, mean age (SD), gender/sex distributiona | Type of diabetes; duration of illnessb [years, mean (SD)] | Diabetes treatment | Comorbidities/complications | Exclusion criteria |
---|---|---|---|---|---|
Akturk, 2019 [21] | Group 1: 134, 31.3 (11.1), 55.2% female Group 2: 316, 39.1 (14.2), 40.2% female | Group 1: TID; 16.3 (–) years Group 2: TID; 20.9 (–) | Continuous glucose monitoring: 45.% Group 1, 55.1% Group 2 Insulin pump: 50.7% Group 1, 66.5% Group 2 | – | Patients with diabetes other than T1D, pregnancy, and repeat follow-up visits within the study duration |
Helgeson, 2016 [22] | 132, 23 (–), 56% female | TID; 15.8 (--) | Insulin (units/day): median 50 (IQR 30-65) Insulin pump therapy: 79/183 (43%) | Indications of neuropathy (vibratory thresholds above age-specific norms): 17%; uACR above 30 mg/g: 7% | – |
Hogendorf, 2016 [23] | 209, 16.5 (1), 48.8% female, 51.2% male | TID; 6.5 (4.4) | – | – | Age 15–18 years |
Thurheimer-Cacciotti, 2017 [24] | 75, 24.3 (4.1), 100% female | TID– | – | – | – |
Winhusen, 2018 [25] | 1184 – (–), – | T2D – | – | – | – |
Wisk, 2018 [26] | 138 20.5 (1.5) 80.40% female | T2D, 10.9 (5.2)^ | Insulin pump users: 84.1% Testing blood sugar > 5 times/day: 44.2% > 3 HbA1c tests in the past year: 65.9% Average last HbA1c: 7.6 | – | – |
First author, year | Consumption method; description | Quantity and frequency consumed | Outcome name | Results |
---|---|---|---|---|
Akturk, 2019 [21] | Group 1: multiple Smoking: 97 (72.4%) Edible: 65 (48.5%) Vaporization: 54 (40.3%) Other: 19 (14.2%) Group 2: not applicable (non-users) | Group 1: < 1 time/month: 48 (35.8%) 2–4 times/month: 14 (10.4%) 2–3 times/week: 17 (12.7%) > 4 times/week: 54 (40.3%) Group 2: not applicable (non-users) | Outcome 1: risk of DKA | Cannabis use within the previous 12 months was associated with an increased risk of DKA compared with no cannabis use (entire cohort OR 1.98, 95% CI 1.01–3.91) |
Outcome 2: HbA1c mean level | Group 1: HbA1c: 8.4% (SD 2.0) [p < .01] Cannabis users had a mean 0.41% higher HbA1c level than nonusers when adjusted for insulin delivery method, income and age (95% CI, 0.38-0.43) Group 2: HbA1c: 7.6% (SD 1.6) [p < .01] | |||
Outcome 3: episodes of severe hypoglycemia | Group 1: 15.6% (21 of 134) [p = .17] Group 2: 20.3% (64 of 316) [p = .17] | |||
Helgeson, 2016 [22] | Inhaled; smoked cannabis | – | Outcome 1: HbA1c level | Average HbA1c was 8.8% (12% with HbA1c > 11%) [at baseline] Smoking cannabis was related to higher HbA1c (r = .30, p < 0.01) |
Outcome 2: albumin-to-creatinine ratio | Smoking cannabis was related to higher uACR (r = .22, p < 0.05) | |||
Hogendorf, 2016 [23] | – | – | Outcome 1: glycemic control | “Half of the [T1D] patients (53%) had HbA1c levels above 8% [at baseline]; lifetime and last 12-month use of cannabis were associated with poorer glycemic control (HbA1c ≥ 8%), p < 0.01 and p < 0.02, respectively.” |
Outcome 2: glycemic control | HbA1c of 6-8%: 14/89 tried cannabis HbA1c of 8-10% : 11/62 HbA1c of 10-12%: 9/30 HbA1c >12%: 4/8 tried cannabis (p = 0.03) | |||
Thurheimer-Cacciotti, 2017 [24] | – | – | Outcome 1: glycemic control | “Women who reported multiple (> 1) risk-taking behaviours were more likely to have a higher HbA1c (> 8%) compared to women who reported 0–1 risky behaviour (RR = 1.29, 95% CI 0.605–2.742).” |
Winhusen, 2018 [25] | – | – | Outcome 1: diabetic renal disease | Cannabis use was associated with a statistically significant increased risk of diabetic renal disease |
Outcome 2: myocardial infarction | Cannabis use was associated with a statistically significant increased risk of myocardial infarction | |||
Outcome 3: peripheral arterial occlusion | Cannabis use was associated with a statistically significant increased risk of peripheral arterial occlusion | |||
Outcome 4: neuropathy | Cannabis use was not associated with a statistically significant increased risk of neuropathy | |||
Outcome 5: cerebrovascular accident | Cannabis use was not associated with a statistically significant increased risk of cerebrovascular accident | |||
Wisk, 2018 [26] | – | – | Outcome 1: diabetes self-management | “Much like their peers, college students with T1D frequently consume alcohol and cannabis; those with T1D who use more frequently experience higher HbA1c and are less likely to achieve glycemic targets, independent of blood glucose testing and diabetes burden.” |
Outcome 2: most recent HbA1c level | “Multivariable analyses revealed that those who drank 3+ days in the past month (50.7% of sample) had significantly higher HbA1c (by 0.63%, p < 0.01) and were significantly more likely to be above recommended glycemic targets (OR 3.08, 95% CI 1.59–5.98). Similar results were observed for cannabis and cigarettes.” |
Quality appraisal results
Cross-sectional studies (n = 4) | ||||||||
---|---|---|---|---|---|---|---|---|
First author, year | Is the sample definition adequate? | Representativeness of the exposed cohort | Ascertainment of exposure | Assessment of outcome | Adequacy of response rate | |||
Akturk, 2019 [21] | B–records/self-report | B–somewhat representative | C–written self-report | A–independent or blind | C–inadequate | |||
Hogendorf, 2016 [23] Thurheimer-Cacciotti, 2017 [24] | B–records/self-report | B–somewhat representative | C–written self-report | C–self-report | B–small number lost | |||
C–no description | C–selected group | D–no description | C–self-report | Unclear | ||||
Wisk, 2018 [26] | B–records/self-report | B–somewhat representative | C–written self-report | C–self-report | D–no description | |||
Case-control studies (n = 1) | ||||||||
First author, year | Is the case definition adequate? | Representativeness of the cases | Selection of controls | Definition of controls | Comparability of cases and controls | Ascertainment of exposure | Same method for cases and controls | Non-response rate |
Winhusen, 2018 [25] | B–records/self-report | A–consecutive or representative | B–hospital | A–no history | A–age and other factor | A–secure record | A–yes | Unclear |
Cohort studies (n = 1) | ||||||||
First author, year | Representativeness of the exposed cohort | Selection of the non-exposed cohort | Ascertainment of exposure | Demonstration outcome was not present at thestart | Comparability of cohorts | Assessment of outcome | Was follow-up long enough for outcomes to occur? | Adequacy of follow-up of cohorts |
Helgeson, 2016 [22] | B–somewhat representative | A–same community | D–no description | A–yes | D–no description | A–independent or blind | A–yes | B–small number lost |