Supplementary Online Content Leibowitz M, Karpati T, Cohen-Stavi CJ, et al. Association between achieved low-density lipoprotein levels and major adverse cardiac events in patients with stable ischemic heart disease taking statin treatment. JAMA Intern Med. Published online June 20, 2016. doi:10.1001/jamainternmed.2016.2751 etable 1. Statin potency classification based on US Food and Drug Administration s classification etable 2. ICD-9 codes for outcomes of MACE etable 3. Number of events by LDL-C group for each individual component of MACE etable 4. Variables included in PS and Cox models efigure 1a. Kaplan-Meier survival analysis for low versus moderate LDL-C groups efigure 1b. Kaplan-Meier survival analysis for moderate versus high LDL-C groups efigure 2a. Distribution of the PS before and after matching: Low versus Moderate LDL-C groups efigure 2b. Distribution of the PS before and after matching: Moderate versus High LDL-C group efigure 3. Calculation of the standardized difference for each kind of variable etable 5a. Standardized difference before and after matching Low versus Moderate LDL- C groups etable 5b. Standardized difference before and after matching Moderate versus High LDL- C groups etable 6a. After propensity-matching Low versus Moderate LDL-C groups etable 6b. After propensity-matching Moderate versus High LDL-C groups This supplementary material has been provided by the authors to give readers additional information about their work.
etable 1. Statin potency classification based on US Food and Drug Administration s classification Potency Category Atorva a Fluva b Pitava c Lova d Prava e Rosuva f Simva g % LDL-C h Low 40 mg 1 mg 20 20 mg 10 mg 30% mg 10 mg 80 mg 2 mg 40 40 mg 20 mg 38% mg Moderate 20 mg 4 mg 80 80 mg 5 mg 40 mg 41% mg High 40 mg 10 mg 80 mg 47% 80 mg 20 mg 55% 40 mg 63% Preparations equivalent to Simvastatin 20mgm (38% LDL-C lowering) are classified as low intensity, preparations equivalent to Simvastatin 40mgm (41% LDL-C lowering) are moderate intensity, and all preparations with higher lipid lowering effects are classified as high intensity, equivalent to Simvastatin 80mgm. a. Atorva Atorvastatin; b. Fluva Fluvastatin; c. Pitava Pitavastatin; d. Lova Lovastatin; e. Prava Pravastatin; f. Rosuva Rosuvastatin; g. Simva Simvastatin; h. LDL- C Low-density lipoprotein cholesterol
etable 2. ICD-9 a codes for outcomes of MACE b Outcome ICD-9 codes Myocardial infarction 410 and related sub-classifications Unstable angina 411.1, 411.8, 411.81, 413.0, 413.1 Stroke 433.01, 433.11, 433.21, 433.31, 433.81, 433.91 Percutaneous coronary intervention procedural code 36 Coronary artery bypass grafting procedural code 36.1 a. ICD-9 International Classification of Diseases, Ninth Revision; b. MACE Major adverse cardiac events
etable 3. Number of events by LDL-C a group for each individual component of MACE b LDL-C group Low ( 70mg/dl) Moderate (70.1-100mg/dl) High (100.1-130mg/dl) Population N 9,086 16,782 5,751 All-cause mortality 693 1,021 437 Acute MI c 421 778 307 CABG d 113 200 71 PTCA e 794 1,442 511 CVA f 189 302 144 Angina 471 852 289 Total MACE 2,681 4,595 1,759 a. LDL-C Low-density lipoprotein; b. MACE Major adverse cardiac events; c. MI Myocardial infarction; d. CABG Coronary artery bypass grafting; e. PTCA Percutaneous transluminal coronary angioplasty; f. CVA Cerebrovascular accident
etable 4. Variables included in PS a and Cox models Variables Low vs Moderate Moderate vs High Cox PS Cox PS Age X X X X Sex X X X X Ethnicity (Jews / Arabs) X X X X SES b (Low / Medium-High) X X X X Residence (Urban / Rural) X X BMI c (<18.5 / 18.5-24.9 / 25.0-29.9 / 30+ / No X X measure) Smoking status (Non-smoker / Actual smoker) X X X X Glucose level (Low / Normal / Borderline / High / X X Very high / No measure) Diabetic (No / Yes) X X X X Insulin use (No / Yes) X X X Statin potency (Low / Medium / High) X X Time on statin (Categorical: 12-23m / 24-71m / X X 72m+) Time on statin (Continuous in months) X X Other non-statin lipid lowering meds (No / Yes) X X HDL X Total Cholesterol Triglycerides Visits above 24 (No / Yes) X X X X Lipid profiles above 9 (No / Yes) X X Chronic meds count (0-4 / 5-7 / 8+) X X CKD d stage (1 / 2 / 3A / 3B / 4 / 5 / Unclassified) X X X Number of angioplasties (PTCA) X X X Charlson risk score (0-1 / 2 / 3-4 / 5+) X X X X a. PS Propensity score; b. SES Socioeconomic status; c. BMI Body mass index; d. CKD Chronic Kidney Disease. For the Cox regression, variables that were thought to be clinical predictors of the appearance of a MACE event were included into the model. For the PS secondary analysis, variables that were considered to be predictors of being in a specific LDL-C group were included in the PS model. The effect of each variable on the multivariable model was assessed for instability and collinearity. Forward stepwise analysis was used to select the variables that were included in the final model. Age and sex were forced into the model if the stepwise analysis excluded them.
efigure 1a. Kaplan-Meier survival analysis for low versus moderate LDL-C a groups a. LDL-C Low-density lipoprotein
efigure 1b. Kaplan-Meier survival analysis for moderate versus high LDL-C a groups a. LDL-C. Low-density lipoprotein
efigure 2a. Distribution of the PS a before and after matching: Low versus Moderate LDL- C b groups a. PS Propensity score; b. LDL-C Low-density lipoprotein
efigure 2b. Distribution of the PS a before and after matching: Moderate versus High LDL- C b group a. PS Propensity score; b. LDL-C Low-density lipoprotein
efigure 3. Calculation of the standardized difference for each kind of variable For continuous variables the formula is based on the means and standard deviation: 2 Where denote the sample mean and denote the variance of the variables in the treatment and control groups. For binary categorical variables, the standard difference is calculated as follows: 1 1 2 Where denotes the proportion of the binary variable in the treatment and control groups. For multi-categorical variables, the standard difference is calculated as in the case of binary categories, but for each category in the variable separately.
etable 5a. Standardized difference before and after matching Low versus Moderate LDL-C a groups Variable Difference Before Difference After 30-49y -0.692 0.308 Age 65-74y 3.572-1.047 75y+ 2.131 1.915 Sex Female -8.499 0.797 Smoking Smoker -3.65-0.446 Missing -1.781 0.079 Diabetes Yes 21.216 0.489 2-6.004 0.197 3A -4.767-0.634 CKD 3B 5.159 0.125 4 7.113 0.163 5 3.443 0.327 Unknown 5.914 0.437 Time on statins 12-23 11.535 0.507 24-71 16.272-0.925 Visits >24 12.088-0.628 Lipid Profiles >9 9.694-1.517 Low 21.803 0.311 HDL High -10.785 0.109 Missing 1.837 0.266 Low 1.102 0.000 Borderline -3.298-0.752 Glucose High 12.749 0.347 Very high 11.009-0.714 Not measured 1.695-0.336 Medium 3.77 0.544 Statin Potency High -2.076-0.549 missing -0.255-0.766 Chronic meds 0-4 -12.659-1.112 Charlson score 4-5 -11.053-0.166 0-1 -6.473-0.059 2-9.557-0.122 3-4 -8.998-1.254 Unknown 3.885 0.795 Non-statins meds Yes 8.27-0.331 a. LDL-C Low-density lipoprotein; b. SES Socioeconomic status; c. CKD Chronic kidney disease; d. PTCA - Percutaneous transluminal coronary angioplasty The groups were matched 1:1 using a greedy algorithm technique, and the best fit model was chosen by the Akaike information criterion measure. After matching, the standardized difference for each of the variables in the PS model was calculated using different formulas depending on whether the variables were continuous, binary categorical or multi-categorical, and then
compared to the differences before matching. A difference greater than the absolute value of 10 is considered unbalanced.[8]
etable 5b. Standardized difference before and after matching Moderate versus High LDL-C a groups Variable Difference Before Difference After Age 30-49 -1.61-0.27 65-74 3.53-0.55 75+ 3.18 1.39 Sex 2-8.17 0.00 Residence Rural -3.86-0.60 Unknown 0.40 0.12 Smoker Smoker -3.60-1.73 Missing -1.14-0.68 Diabetes Diabetes 21.77-0.20 CKD 2-6.00-1.13 3A -5.15 0.53 3B 4.63-0.10 4 6.77 0.08 5 4.32 1.38 Unknown 5.85-0.52 Time on Statin 12-23 9.92-1.18 24-71 16.94 0.74 Visits >24 11.50-0.70 Lipid Profiles >9 9.70-0.81 PTCA 1 5.74-0.57 2 5.74-0.31 3+ 3.49-0.32 HDL Low 23.14-0.87 High -11.81 0.00 Missing 1.56-0.53 Total Cholesterol Low 72.56 0.84 High -15.30-1.55 Missing 1.45-0.48 Triglycerides High 5.42-0.81 Missing 1.47-0.53 Glucose Low 0.33 0.37 Borderline -3.54 0.74 Hyperglycemia 13.78 0.02 Diabetic 10.28-1.15 Unknown 0.96 0.29 Statin potency Medium 0.30 1.38 High -0.61-0.94 missing -0.95 0.18 Chronic Meds 0-4 -13.00-0.09 4-5 -10.25-0.33 Charlson score 0-1 -6.64-1.20 2-9.94-0.06
3-4 -9.28-0.69 Unknown 3.26-0.29 Non-statin meds Yes 5.49 0.91 a. LDL-C Low-density lipoprotein; b. SES Socioeconomic status; c. BMI Body mass index
etable 6a. After propensity-matching Low versus Moderate LDL-C a groups Characteristics Moderate LDL- Low LDL-C, p value C, 70.1-100 mg/dl* (n=7,879) 70 mg/dl** (n=7,879) Age Mean (SD b ) 67.8 (9.6) 67.7 (9.6) 0.210 Median (IQR c ) 69.0 (61.0-76.0) 68.0 (61.0-75.0) 0.162 30-49 312 (3.5%) 307 (3.5%) 50-64 2,966 (33.6%) 3,003 (34.0%) 65-74 3,037 (34.4%) 3,081 (34.9%) 75-84 2,518 (28.5%) 2,442 (27.6%) 0.626 Sex Male 6,711 (76.0%) 6,741 (76.3%) Female 2,122 (24.0%) 2,092 (23.7%) 0.609 Ethnicity Jewish 7,998 (90.5%) 7,863 (89.0%) Non-Jewish 835 (9.5%) 970 (11.0%) 0.001 SES d Mid-High 6,026 (68.2%) 5,881 (66.6%) Low 2,807 (31.8%) 2,952 (33.4%) 0.021 Index LDL-C Mean (SD) 58.5 (9.7) 83.7 (8.3) 0.000 Median (IQR) 60.9 (53.7-66.0) 83.0 (76.6-90.4) 0.000 HDL e Mean (SD) 45.2 (11.9) 45.9 (11.5) 0.000 Median (IQR) 43.0 (37.0-51.0) 44.0 (38.0-52.0) 0.000 BMI f Mean (SD) 29.1 (5.1) 29.3 (5.2) 0.089 Median (IQR) 28.4 (25.7-31.6) 28.4 (25.9-31.8) 0.052 <18.5 14 (0.2%) 20 (0.2%) 18.5-24.9 1,647 (18.6%) 1,553 (17.6%) 25.0-29.9 3,941 (44.6%) 3,908 (44.2%) 30+ 3,115 (35.3%) 3,242 (36.7%) Unknown 116 (1.3%) 110 (1.2%) 0.155 Chronic diseases Diabetics 5,446 (61.7%) 5,425 (61.4%) 0.757 Use Insulin 688 (7.8%) 629 (7.1%) 0.097 Hypertension 3,975 (45.0%) 4,002 (45.3%) 0.694 CHF g 1,611 (18.2%) 1,582 (17.9%) 0.584 AF h 1,185 (13.4%) 1,090 (12.3%) 0.035 CVA i 61 (0.7%) 62 (0.7%) 1.000 Charlson score 0-1 332 (3.8%) 333 (3.8%) 2 741 (8.4%) 744 (8.4%) 3-4 2,495 (28.2%) 2,545 (28.8%) 5+ 5,197 (58.8%) 5,149 (58.3%) Unknown 68 (0.8%) 62 (0.7%) 0.909 CKD j stage 1 196 (2.2%) 192 (2.2%) 2 1,789 (20.3%) 1,782 (20.2%) 3A 4,349 (49.2%) 4,377 (49.6%)
3B 1,383 (15.7%) 1,379 (15.6%) 4 743 (8.4%) 739 (8.4%) 5 275 (3.1%) 270 (3.1%) Unknown 98 (1.1%) 94 (1.1%) 1.000 PTCA k 0 7,593 (86.0%) 7,603 (86.1%) 1-2 1,186 (13.4%) 1,187 (13.4%) 3+ 54 (0.6%) 43 (0.5%) 0.532 Number of visits >24 5,414 (61.3%) 5,441 (61.6%) 0.688 Numper of lipid profiles >9 4,396 (49.8%) 4,463 (50.5%) 0.321 Time on statins (years) Mean (SD) 6.7 (3.4) 6.9 (3.4) 0.001 Median (IQR) 7.0 (4.0-10.0) 7.0 (4.0-10.0) 0.001 Statin potency Low 439 (5.0%) 435 (4.9%) Medium 4,599 (52.1%) 4,575 (51.8%) High 3,766 (42.6%) 3,790 (42.9%) missing 29 (0.3%) 33 (0.4%) 0.937 Simvastatin 4,706 (53.3%) 4,557 (51.6%) 0.026 Atorvastatin 2,414 (27.3%) 2,886 (32.7%) 0.000 Rosuvastatin 1,578 (17.9%) 1,165 (13.2%) 0.000 Pravastatin 135 (1.5%) 224 (2.5%) 0.000 MACE l (outcome) 2,583 (29.2%) 2,625 (29.7%) 0.499 a. LDL-C Low-density lipoprotein; b. SD Standard deviation; c. IQR Interquartile range;; d. SES Socioeconomic status; e. HDL High-density lipoprotein; f. BMI Body mass index; g. CHF Congestive heart failure; h. AF Atrial fibrillation; i. CVA Cerebrovascular accident; j. CKD Chronic kidney disease; k. PTCA Percutaneous transluminal coronary angioplasty; l. MACE Major adverse cardiac events; *70.1-100mg/dl 1.8mmol/l-2.6mmol/l; **100.1-130mg/dl 2.6-3.4mmol/l;
etable 6b. After propensity-matching Moderate versus High LDL-C a groups Characteristics Moderate LDL-C, High LDL-C, p value 70.1-100 mg/dl* (n=5,739) 100.1-130 mg/dl ** (n=5,739) Age Mean (SD b ) 66.6 (9.9) 66.5 (10.3) 0.445 Median 67.0 (59.0-75.0) 67.0 (59.0-75.0) 0.537 (IQR c ) 30-49 260 (4.5%) 304 (5.3%) 50-64 2,216 (38.6%) 2,246 (39.1%) 65-74 1,792 (31.2%) 1,713 (29.8%) 75+ 1,471 (25.6%) 1,476 (25.7%) 0.143 Sex Male 3,893 (67.8%) 3,884 (67.7%) Female 1,846 (32.2%) 1,855 (32.3%) 0.873 Ethnicity Jewish 4,959 (86.4%) 4,970 (86.6%) Non Jewish 780 (13.6%) 769 (13.4%) 0.785 SES d Med-High 3,668 (63.9%) 3,688 (64.3%) Low 2,071 (36.1%) 2,051 (35.7%) 0.712 Index LDL-C Mean (SD) 84.3 (8.3) 111.2 (7.9) 0.000 Median (IQR) 84.0 (77.2-91.0) 109.8 (104.4-117.0) 0.000 HDL e Mean (SD) 47.6 (11.8) 48.9 (12.5) 0.000 Median (IQR) 46.0 (39.0-54.0) 47.0 (40.0-55.0) 0.000 BMI f Mean (SD) 29.2 (5.1) 29.3 (5.3) 0.249 Median (IQR) 28.4 (25.8-31.8) 28.6 (25.8-31.9) 0.589 <18.5 8 (0.1%) 9 (0.2%) 18.5-24.9 1,050 (18.3%) 1,061 (18.5%) 25.0-29.9 2,432 (42.4%) 2,434 (42.4%) 30+ 2,160 (37.6%) 2,148 (37.4%) Unknown 89 (1.6%) 87 (1.5%) 0.996 Chronic diseases Diabetics 2,894 (50.4%) 2,859 (49.8%) 0.526 Use Insulin 304 (5.3%) 297 (5.2%) 0.801 Hypertension 2,439 (42.5%) 2,456 (42.8%) 0.763 CHF g 882 (15.4%) 897 (15.6%) 0.718 AF h 604 (10.5%) 645 (11.2%) 0.231 CVA i 33 (0.6%) 45 (0.8%) 0.211 Charlson score 0-1 369 (6.4%) 369 (6.4%) 2 669 (11.7%) 660 (11.5%) 3-4 1,804 (31.4%) 1,784 (31.1%) 5+ 2,867 (50.0%) 2,897 (50.5%) Unknown 30 (0.5%) 29 (0.5%) 0.987 CKD j stage 1 164 (2.9%) 163 (2.8%) 2 1,402 (24.4%) 1,374 (23.9%) 3A 2,844 (49.6%) 2,780 (48.4%)
3B 758 (13.2%) 806 (14.0%) 4 386 (6.7%) 414 (7.2%) 5 148 (2.6%) 164 (2.9%) Unknown 37 (0.6%) 38 (0.7%) 0.636 PTCA k 0 5,078 (88.5%) 5,090 (88.7%) 1-2 637 (11.1%) 621 (10.8%) 3+ 24 (0.4%) 28 (0.5%) 0.895 Number of visits >24 3,065 (53.4%) 3,067 (53.4%) 0.985 Numper of lipid profiles >9 2,438 (42.5%) 2,439 (42.5%) 1.000 Time on statins (years) Mean (SD) 7.5 (3.1) 7.6 (3.1) 0.073 Median (IQR) 8.0 (5.0-10.0) 8.0 (5.0-11.0) 0.034 Statin potency High 2,473 (43.1%) 2,459 (42.8%) Low 496 (8.6%) 498 (8.7%) Medium 2,733 (47.6%) 2,748 (47.9%) missing 37 (0.6%) 34 (0.6%) 0.976 Simvastatin 2,823 (49.2%) 2,736 (47.7%) 0.108 Atorvastatin 1,806 (31.5%) 1,999 (34.8%) 0.000 Rosuvastatin 844 (14.7%) 668 (11.6%) 0.000 Pravastatin 265 (4.6%) 335 (5.8%) 0.004 MACE l (outcome) 1,614 (28.1%) 1,757 (30.6%) 0.004 a. LDL-C Low-density lipoprotein; b. SD Standard deviation; c. IQR Interquartile range;; d. SES Socioeconomic status; e. HDL High-density lipoprotein; f. BMI Body mass index; g. CHF Congestive heart failure; h. AF Atrial fibrillation; i. CVA Cerebrovascular accident; j. CKD Chronic kidney disease; k. PTCA Percutaneous transluminal coronary angioplasty; l. MACE Major adverse cardiac events; *70.1-100mg/dl 1.8mmol/l-2.6mmol/l; **100.1-130mg/dl 2.6-3.4mmol