Supplementary Online Content Pedersen SB, Langsted A, Nordestgaard BG. Nonfasting mild-to-moderate hypertriglyceridemia and risk of acute pancreatitis. JAMA Intern Med. Published online November 7, 2016. doi:10.1001/jamainternmed.2016.6875. emethods. etable 1. Baseline characteristics of individuals with measurement of either lipase or pancreatic amylase in the Co-penhagen General Population Study efigure 1. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides before correction for regression dilution bias efigure 2. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for time since last meal efigure 3. Hazards ratios for acute pancreatitis by higher levels of triglycerides after additional exclusion of individuals with chronic pancreatitis events before the exami-nation day efigure 4. Hazards ratios for acute pancreatitis in the two separate study cohorts efigure 5. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for high-density lipoprotein cholesterol efigure 6. Hazards ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglyc-erides in different strata before correction for regression dilution bias efigure 7. Hazard ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglycer-ides with postponed start of follow-up efigure 8. Association between plasma levels of triglycerides and lipase or pan-creatic amylase after exclusion of individuals with triglycerides levels above the recommended levels for the assay This supplementary material has been provided by the authors to give readers additional information about their work. 1
Supplementary Online Content Nonfasting mild-to-moderate hypertriglyceridemia and risk of acute pancreatitis Simon B. Pedersen, BMSc; Anne Langsted, MD, PhD; Børge G. Nordestgaard, MD, DMSc emethods Table of contents etable. Baseline characteristics of individuals with measurement of either lipase or pancreatic amylase in the Copenhagen General Population Study efigure 1. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides before correction for regression dilution bias efigure 2. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for time since last meal efigure 3. Hazards ratios for acute pancreatitis by higher levels of triglycerides after additional exclusion of individuals with chronic pancreatitis events before the examination day efigure 4. Hazards ratios for acute pancreatitis in the two separate study cohorts efigure 5. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for high-density lipoprotein cholesterol efigure 6. Hazards ratios for acute pancreatitis by higher levels of triglycerides in different strata efigure 7. Hazards ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglycerides in different strata before correction for regression dilution bias efigure 8. Hazard ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglycerides with postponed start of follow-up efigure 9. Association between plasma levels of triglycerides and lipase or pancreatic amylase after exclusion of individuals with triglycerides levels above the recommended levels for the assay 2
emethods Laboratory analysis Plasma levels of high-density lipoprotein cholesterol were measured in the Copenhagen General Population Study using standard hospital assays. Covariates Level of education was defined as elementary ( 9 years of education), high school (9-12 years), and academic (>12 years). Smokers were current smokers. Hypertension was a systolic blood pressure above 140 mmhg, a diastolic blood pressure above 90 mmhg, and/or use of antihypertensive medication. Use of statins was selfreported. Alcohol intake was alcohol units (of 12 g) per week. High alcohol intake was men consuming 21 units/week ( 36 g of alcohol per day) or women consuming 14 units/week ( 24 g of alcohol per day), according to the Danish national recommendations. Diabetes was self-reported disease, use of insulin or oral hypoglycemic medication, a non-fasting plasma glucose >198 mg/dl (>11 mmol/liter), and/or registry-based diagnosis of diabetes (ICD-8 code 250 and ICD10 codes E11, E13-E14). Body mass index was measured weight in kilograms divided by measured height in meters squared. Time since last meal was self-reported hours since the individuals had eaten their last meal before blood sampling (0-1, 1-2, 2-3, 3-4, 4-5, 5-6, 6-7, 7-8, and >8 hours). For calculation of median and interquartile range, >8 was estimated to 10 hours. Information on time since last meal was only available in the Copenhagen General Population Study. Gallstone disease was any record of gallstone disease (8th edition codes 574-575 and 10th edition codes K80-K81) in the national Danish Patient Registry or the national Danish Causes of Death Registry.
etable. Baseline characteristics of individuals with measurement of either lipase or pancreatic amylase in the Copenhagen General Population Study Variable Lipase tertiles Pancreatic amylase tertiles All Lowest Middle Highest All Lowest Middle Highest Number of individuals 15,856 5606 4977 5273 92,672 32,032 30,348 30,292 Lipase/pancreatic amylase, units/l Age, years 31 (22-43) 56 (47-67) 18 (13-22) 57 (47-67) 32 (28-35) 56 (47-67) 50 (43-62) 56 (47-66) 32 (26-40) 58 (48-67) 24 (20-26) 58 (48-67) 33 (31-35) 57 (47-66) Men 46% 49% 45% 43% 45% 49% 45% 40% Education 9 years a 38% 40% 37% 36% 25% 27% 24% 23% Current smokers 27% 30% 25% 26% 17% 21% 16% 14% Hypertension b 65% 66% 64% 64% 66% 70% 65% 64% Statin use c 7% 7% 8% 7% 12% 13% 11% 12% Diabetes d 8% 8% 7% 7% 5% 8% 4% 4% Body mass index, kg/m2 26 (23-29) 26 (23-29) 26 (23-29) 26 (23-28) 26 (23-28) 26 (24-29) 26 (23-28) 44 (40-51) 58 (48-67) 25 (23-27) Alcohol intake, units per e 8 (3-15) 8 (3-16) 8 (3-15) 8 (3-15) 8 (3-15) 9 (4-17) 8 (3-14) 7 (3-13) week Time since last meal, hours 3 (2-4) 3 (2-5) 3 (2-4) 3 (2-4) 3 (2-4) 3 (2-4) 3 (2-4) 3 (2-4) On the basis of individuals from the Copenhagen General Population Study. Lipase, pancreatic amylase, age, body mass index, alcohol intake, and time since last meal are expressed as median (interquartile range). a 9 years of education = elementary education. b Hypertension was a systolic blood pressure above 140 mmhg, a diastolic blood pressure above 90 mmhg, and/or use of antihypertensive medication. c Statin use was self-reported. d Diabetes was self-reported disease, use of insulin or oral hypoglycemic medication, a non-fasting plasma glucose >198 mg/dl (>11 mmol/l), or a registry-based diagnosis of diabetes. e 1 unit alcohol = 12g. 4
efigure 1. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides before correction for regression dilution bias Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, birth year, and study cohort. CI = Confidence Interval. BMI = Body mass index.
efigure 2. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for time since last meal On the basis of individuals from the Copenhagen General Population Study only. Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use and birth year. CI = Confidence Interval.
efigure 3. Hazards ratios for acute pancreatitis by higher levels of triglycerides after additional exclusion of individuals with chronic pancreatitis events before the examination day Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, birth year, and study cohort. Corrected for regression dilution bias. CI = Confidence Interval. BMI = Body mass index.
efigure 4. Hazards ratios for acute pancreatitis in the two separate study cohorts Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, and birth year. CI = Confidence Interval. BMI = Body mass index.
efigure 5. Hazards ratios for acute pancreatitis and myocardial infarction by higher levels of triglycerides further adjusted for high-density lipoprotein cholesterol On the basis of individuals from the Copenhagen General Population Study only. Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use and birth year. CI = Confidence Interval.
efigure 6. Hazards ratios for acute pancreatitis by higher levels of triglycerides in different strata Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, birth year, and study cohort. Gallstone disease was any record of gallstone disease. CI = Confidence Interval. 10
efigure 7. Hazards ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglycerides in different strata before correction for regression dilution bias Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, birth year, and study cohort. Tests for interaction were performed by introducing a 2- factor interaction term in the Cox regression model. High alcohol intake was men consuming 21 units/week ( 36 g of alcohol per day) or women consuming 14 units/week ( 24 g of alcohol per day). Gallstone disease was any record of gallstone disease in Danish national registries. Individuals and events differ slightly in different strata due to availability of data. CI = Confidence interval. CGPS = Copenhagen General Population Study. CCHS = Copenhagen City Heart Study. 11
efigure 8. Hazard ratios for acute pancreatitis per 89 mg/dl (1 mmol/l) higher triglycerides with postponed start of follow-up Analyzes were performed by only allowing individuals to enter the Cox regression model at 1, 2, 3, 4, and 5 years after the initial start of follow-up and triglyceride measurement. Hazard ratios were multivariable adjusted for age, sex, education, smoking, hypertension, statin use, birth year, and study cohort. CI = Confidence Interval.
efigure 9. Association between plasma levels of triglycerides and lipase or pancreatic amylase after exclusion of individuals with triglycerides levels above the recommended levels for the assay On the basis of individuals from the Copenhagen General Population Study. Multivariable adjusted was for age, sex, education, smoking, hypertension, statin use, and birth year. Mean was geometric mean. P values for trend were by the Cuzick nonparametric test for trend. For the lipase assay we excluded individuals with plasma triglycerides above 1000 mg/dl (>11.3 mmol/l) and for the pancreatic amylase assay we excluded individuals with plasma triglycerides above 2000 mg/dl (>22.6 mmol/l).