Association of Statin Therapy Initiation With Diabetes Progression: A Retrospective Matched-Cohort Study

Ishak A Mansi; Matthieu Chansard; Ildiko Lingvay; Song Zhang; Ethan A Halm; Carlos A Alvarez

doi:10.1001/jamainternmed.2021.5714

. 2021 Oct 4;181(12):1–14. doi: 10.1001/jamainternmed.2021.5714

Association of Statin Therapy Initiation With Diabetes Progression

A Retrospective Matched-Cohort Study

Ishak A Mansi ^1,^2,^3,^✉, Matthieu Chansard ⁴, Ildiko Lingvay ^2,³, Song Zhang ³, Ethan A Halm ^2,³, Carlos A Alvarez ^3,⁵

¹Department of Medicine, VA North Texas Health Care System, Dallas

²Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas

³Department of Population and Data Sciences, University of Texas Southwestern Medical Center, Dallas

⁴Department of Anesthesiology and Pain Management, University of Texas Southwestern Medical Center, Dallas

⁵Department of Pharmacy Practice, Texas Tech University Health Sciences Center, Dallas

Accepted for Publication: October 4, 2021.

Published Online: October 4, 2021. doi:10.1001/jamainternmed.2021.5714

^✉

Corresponding Author: Ishak A. Mansi, MD, VA North Texas Health System, 4500 S Lancaster Rd, #111E, Dallas, TX 75216 (ishak.mansi@va.gov).

Author Contributions: Dr Mansi, Mr Chansard, and Dr Alvarez had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Mansi, Lingvay, Halm, Alvarez.

Acquisition, analysis, or interpretation of data: Mansi, Chansard, Lingvay, Zhang, Alvarez.

Drafting of the manuscript: Mansi, Alvarez.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Mansi, Zhang, Alvarez.

Administrative, technical, or material support: Mansi.

Supervision: Mansi, Halm, Alvarez.

Other−data management: Chansard.

Conflict of Interest Disclosures: Dr Lingvay reports research funding, advisory and consulting fees, and/or other support from Novo Nordisk, Eli Lilly, Sanofi, AstraZeneca, Boehringer Ingelheim, Janssen, Intercept, Intarcia, TARGETPharma, Merck, Pfizer, Novartis, GI Dynamics, Mylan, Mannkind, Valeritas, Bayer, and Zealand Pharma. Dr Alvarez reports a grant (No. K08 DK101602) from the National Institutes of Diabetes and Digestive and Kidney Diseases during the conduct of the study and grants from Merck outside the submitted work. No other disclosures were reported.

Disclaimer: This material is the result of work supported with resources and the use of facilities at the North Texas VA and University of Texas Southwestern. The views expressed herein are those of the authors and do not reflect the official policy or position of the US Department of the Army, US Department of Defense, the US Department of Veterans Affairs, or the US government.

^✉

Corresponding author.

PMCID: PMC8491130 PMID: 34605849

Key Points

Question

What is the association of statin treatment initiation and diabetes progression in patients with diabetes?

Findings

This large retrospective cohort study included 83 022 propensity-scored matched pairs of statin users and nonusers and found that the diabetes-progression composite outcome was significantly higher among patients with diabetes who used statins than among patients with diabetes who did not use statins. The study examined 12 years of data on patients covered by the Veterans Affairs health system and new-user and active-comparator designs to assess associations between statin initiation and diabetes progression from 2003 to 2015.

Meaning

Statin use was associated with diabetes progression in patients with diabetes—statin users had a higher likelihood of insulin treatment initiation, developing significant hyperglycemia, experiencing acute glycemic complications, and being prescribed an increased number of glucose-lowering medication classes.

Abstract

Importance

Statin therapy has been associated with increased insulin resistance; however, its clinical implications for diabetes control among patients with diabetes is unknown.

Objective

To assess diabetes progression after initiation of statin use in patients with diabetes.

Design, Setting, and Participants

This was a retrospective matched-cohort study using new-user and active-comparator designs to assess associations between statin initiation and diabetes progression in a national cohort of patients covered by the US Department of Veterans Affairs from fiscal years 2003-2015. Patients included were 30 years or older; had been diagnosed with diabetes during the study period; and were regular users of the Veterans Affairs health system, with records of demographic information, clinical encounters, vital signs, laboratory data, and medication usage.

Interventions

Treatment initiation with statins (statin users) or with H2-blockers or proton pump inhibitors (active comparators).

Main Outcomes and Measures

Diabetes progression composite outcome comprised the following: new insulin initiation, increase in the number of glucose-lowering medication classes, incidence of 5 or more measurements of blood glucose of 200 mg/dL or greater, or a new diagnosis of ketoacidosis or uncontrolled diabetes.

Results

From the 705 774 eligible patients, we matched 83 022 pairs of statin users and active comparators; the matched cohort had a mean (SD) age of 60.1 (11.6) years; 78 712 (94.9%) were men; 1715 (2.1%) were American Indian/Pacific Islander/Alaska Native, 570 (0.8%) were Asian, 17 890 (21.5%) were Black, and 56 633 (68.2 %) were White individuals. Diabetes progression outcome occurred in 55.9% of statin users vs 48.0% of active comparators (odds ratio, 1.37; 95% CI, 1.35-1.40; P < .001). Each individual component of the composite outcome was significantly higher among statin users. Secondary analysis demonstrated a dose-response relationship with a higher intensity of low-density lipoprotein-cholesterol lowering associated with greater diabetes progression.

Conclusions and Relevance

This retrospective matched-cohort study found that statin use was associated with diabetes progression, including greater likelihood of insulin treatment initiation, significant hyperglycemia, acute glycemic complications, and an increased number of prescriptions for glucose-lowering medication classes. The risk-benefit ratio of statin use in patients with diabetes should take into consideration its metabolic effects.

This national matched-cohort study explores the association between statin therapy initiation and progression of diabetes among patients with diabetes.

Introduction

Guidelines recommend statin therapy for all patients with diabetes mellitus type 2 (diabetes) who are 40 to 75 years old and have a low-density lipoprotein (LDL) cholesterol level of 70 mg/dL or greater (to convert to mmol/L, multiply by 0.0259) for primary prevention of cardiovascular diseases (CVD).^1,2 However, statin use has been associated with increased insulin resistance and higher blood glucose levels. Several randomized controlled trials (RCTs)^3,4,5 and large prospective and retrospective observational studies^6,7,8 have noted that patients treated with statin therapy (hereafter, statin users) had increased insulin resistance, hemoglobin A_1C (HbA_1C) levels_, and fasting plasma glucose levels. In a large observational study, statin use was associated with a 24% reduction in insulin sensitivity.⁸ Despite the significant reduction in insulin sensitivity and the increase in fasting plasma insulin, the difference in fasting plasma glucose and HbA_1C between statin users and nonusers appears modest.⁸ For example, a large RCT noted that an increase in HbA_1C was 0.30% in the rosuvastatin group and 0.22% in the placebo group (P < .001), and there was no significant difference in fasting serum glucose between the groups.⁹

A modest change in fasting blood glucose after statin initiation despite a relatively large change in insulin resistance and fasting insulin levels deserves further study. Clinicians might escalate antidiabetes therapy to offset the rising blood glucose; hence, HbA_1C levels may underestimate how statin therapy influences diabetes control. Yet increased insulin resistance is of concern because it may fuel diabetes disease progression.^10,11 It is important to understand the clinical importance of increased insulin resistance to actual patient care.

The statin Diabetes Safety Task Force has remarked on the paucity of data regarding how statin use affects glycemic control.¹² The present study’s objective was to compare diabetes progression (by assessing new insulin treatment initiations, changes in the number of glucose-lowering medication classes, and new persistent hyperglycemia or acute glycemic complications) after statin initiation with progression among nonusers in a national cohort of patients covered by the US Department of Veterans Affairs (VA).

Methods

Study Design

This was a retrospective matched-cohort study that used new-user and active-comparator designs to assess associations between statin initiation and diabetes progression among a national cohort of patients covered by the VA from fiscal year (FY) 2003-FY 2015. The study was reviewed and approved by the institutional review boards of the VA North Texas Health Care System and the Texas Tech University Health Sciences Center. Informed consent was waived because the study used only preexisting deidentified data. The study followed the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline and the reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiologic research (RECORD-PE).

We extracted medical record data from the national VA Corporate Data Warehouse FY 2003-FY 2015 (October 1, 2002-September 30, 2015), which captures several domains under published protocols.¹³ It includes inpatient and outpatient diagnoses and procedure codes, pharmacy and medication usage, vital signs records, and laboratory data. We used all available medical encounters for a national VA cohort of patients diagnosed with diabetes during the study period, identified using a validated algorithm.^14,15

The overall cohort included patients who met these criteria, were age 30 years or older at the index date and were regular VA health system users. We defined regular VA system users as having the following during both the baseline and follow-up periods: (1) at least 1 VA encounter; (2) blood pressure and weight measurements; (3) VA pharmacy dispensing; and (4) laboratory data, including blood or serum glucose, creatinine, and LDL cholesterol measurements.

Study Groups

We used an active comparator, new user design, to minimize unmeasured confounders and confounding by indication.^16,17 The statin user group was composed of patients who initiated statin therapy within the study period. The active comparator group was composed of patients who initiated an H2-blocker or proton pump inhibitor (H2/PPI) and were not concurrently prescribed a statin. To exclude prevalent users, any patient who during the 12 months prior to cohort entry had filled a statin prescription was excluded from the statin user group and any patient who had filled an H2-blocker or PPI prescription was excluded from the active comparator group. For any patient in the active comparator group who filled a statin prescription during the follow-up period, their follow-up ended as a nonuser (on the date of statin initiation) and they were crossed over to the statin user group (the date of statin initiation became their new index date). This design mitigated confounding by indication and immortal time bias.¹⁸

The index date was the date on which the first prescription for statin therapy or an H2 or PPI was filled. Because the study data included all available encounters from FY 2003-FY 2015, regardless of when a patient was diagnosed with diabetes, the index date could precede, coincide with, or occur after the date of diabetes diagnosis.

Study Intervals

The baseline period, which was used to describe baseline characteristics, comprised the year preceding the index date. The follow-up period, which was used to ascertain outcomes, started at the index date and continued until: (1) the last date of VA care, (2) the end of the study period, (3) death of the patient, or (4) date of statin initiation in active comparators who subsequently used a statin. To minimize confounding, we excluded patients who had fewer than 60 days of follow-up because the study outcomes would be highly unlikely to occur with fewer than 60 days of statin exposure.^18,19,20

Primary Outcome—Diabetes Progression

This study’s prespecified outcome was a dichotomous composite outcome that comprised (1) therapy intensification, including new insulin initiation during the follow-up period or an increased number of glucose-lowering medication classes that were ever used during follow-up in comparison with the baseline (eTable 1 in the Supplement) and (2) new persistent hyperglycemia or acute glycemic complications, including: (a) the presence of 5 or more measurements with blood glucose levels of 200 mg/dL or greater (to convert to mmol/L, multiply by 0.0555) during follow-up (not present during the baseline period) and (b) receiving a new diagnosis of diabetes with ketoacidosis or uncontrolled diabetes during the follow-up period (not present during the baseline period). The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes were used to identify diabetic ketoacidosis or uncontrolled diabetes (eTable 2 in the Supplement)²¹ and have been widely used in the literature.^22,23,24 Overall, administrative health databases have found ICD-9-CM codes to be useful for identifying diabetes and its complications, with specificity from 94.3% to 100%, although sensitivity can vary.^25,26

Secondary Outcomes

The study had 4 secondary outcomes. The first was individual components of diabetes progression outcome; second, the difference in the number of glucose-lowering medication classes ever used during the follow-up period by each individual in comparison with the number during their baseline; third, the proportion of patients with a decreased number of glucose-lowering medication classes during follow-up vs baseline; and fourth, the change in mean blood glucose (mg/dL) during follow-up vs baseline.

Cohort Characterization

Patients’ comorbidities were identified using ICD-9-CM codes as defined by the Agency for Health Research and Quality Clinical Classifications Software disease categories.²¹ We calculated each patient’s Charlson comorbidity index²⁷ score and cardiovascular risk (eTables 3 and 4 in the Supplement).^28,29,30 We created a propensity score (PS) to match statin users and active comparators (nonusers) at a ratio of 1:1 using 93 variables chosen a priori (Table 1).³¹ We used the routine by Leuven and Sianesi³² to perform multivariable logistic regression to estimate the PS and to perform nearest number matching using the logit model.³³ We subsequently examined the balance in baseline characteristics between treatment groups. A caliper of 0.00014 was found to balance differences between treatment groups and maximize sample size (eMethods in the Supplement).

Table 1. Baseline Characteristics of Propensity Score–Matched Statin Users and Active Comparators in the Overall and Diabetes-Prevalent Cohorts.

Characteristics	Overall cohort			Overall cohort before matching			Diabetes-prevalent cohort
	No. (%)		P value	No. (%)		P value	No. (%)		P value
	Statin users (n = 83 022)	Active comparators (n = 83 022)	P value	Statin users (n = 595 579)	Active comparators (n = 110 195)	P value	Statin users (n = 51 467)	Active comparators (n = 51 467)	P value
Baseline characteristics in the propensity score
Demographic characteristics
Age, mean (SD), y	60.1 (11.7)	60.1 (11.6)	.72	60.7 (10.5)	59.3 (11.6)	<.001	61.5 (11.3)	61.5 (11.2)	.47
Sex (self-reported)
Men	78 660 (94.8)	78 763 (94.9)	.26	573 455 (96.3)	104 235 (94.6)	<.001	49 246 (95.7)	49 384 (96.0)	.03
Women	4362 (5.2)	4259 (5.1)		22 124 (3.7)	5960 (5.4)	<.001	2221 (4.3)	2083 (4.1)
Race (self-reported)
American Indian/Alaska Native/Pacific/Islander	1582 (1.9)	1848 (2.2)	<.001	12 170 (2.0)	2389 (2.2)	<.001	988 (1.9)	1189 (2.3)	<.001
Asian	672 (0.8)	468 (0.6)	<.001	4543 (0.8)	633 (0.6)	<.001	447 (0.9)	323 (0.6)	<.001
Black/African American	17 904 (21.6)	17 876 (21.5)	.87	108 178 (18.2)	24 378 (22.1)	<.001	10 616 (20.6)	10 469 (20.3)	.26
White	56 542 (68.1)	56 724 (68.3)	.34	426 253 (71.6)	74 946 (68.0)	<.001	35 304 (68.6)	35 564 (69.1)	.08
Unknown/missing	6322 (7.6)	6106 (7.4)	.04	44 435 (7.5)	7849 (7.1)	<.001	4112 (8.0)	3922 (7.6)	.03
Ethnicity
Hispanic Latino	5466 (6.6)	5643 (6.8)	.08	37 152 (6.2)	7648 (6.9)	<.001	3512 (6.8)	3628 (7.1)	.16
Non-Hispanic Latino	73 345 (88.3)	73 320 (88.3)	.85	529 171 (88.9)	97 360 (88.4)	<.001	45 276 (88.0)	45 225 (88.0)	.63
Unknown/missing	4211 (5.1)	4059 (4.9)	.09	29 256 (4.9)	5187 (4.7)	.004	2679 (5.2)	2614 (5.1)	.36
Social and family medical history
Family history of CVD	991 (1.2)	1025 (1.2)	.45	9743 (1.6)	1262 (1.2)	<.001	531 (1.0)	530 (1.0)	.98
Smoking	16 389 (19.7)	16 347 (19.7)	.80	110 710 (18.6)	22 723 (20.6)	<.001	9155 (17.8)	9265 (18.0)	.37
Alcohol-related disorder	7591 (9.1)	7670 (9.2)	.50	35 133 (5.9)	11 800 (10.7)	<.001	4128 (8.0)	4139 (8.0)	.90
Substance-related disorder	5671 (6.8)	5724 (6.9)	.61	25 123 (4.2)	8682 (7.9)	<.001	3013 (5.9)	3030 (5.9)	.82
Duration of follow-up, mean (SD), d	1742 (1101)	1749 (1101)	.19	2430 (1092)	1467 (1105)	<.001	1439 (985)	1445 (992)	.31
Vital signs during baseline period
Blood pressure, mean (SD), mm Hg
Systolic	135 (15)	135 (15)	.92	137 (15)	135 (14)	<.001	135 (15)	135 (15)	.72
Diastolic	78 (10)	78 (10)	.92	79 (10)	78 (10)	<.001	77 (10)	77 (10)	.72
BMI
<25	9850 (11.9)	9949 (12.0)	.45	50 118 (8.4)	14 233 (12.9)	<.001	5619 (10.9)	5677 (11.0)	.56
25 to <30	22 852 (27.5)	22 832 (27.5)	.91	161 330 (27.1)	30 631 (27.8)	<.001	13 664 (26.7)	13 767 (26.8)	.47
30 to <35	21 373 (25.7)	21 415 (25.8)	.81	166 751 (28.0)	28 158 (25.6)	<.001	13 295 (25.8)	13 330 (25.9)	.80
35 to <40	11 097 (13.4)	11 032 (13.3)	.64	86 780 (14.6)	14 472 (13.1)	<.001	7285 (14.2)	7256 (14.1)	.80
40 to <45	4277 (5.2)	4281 (5.2)	.97	33 530 (5.6)	5547 (5.0)	<.001	2935 (5.7)	2906 (5.7)	.70
≥45	2438 (2.9)	2392 (2.9)	.50	17 983 (3.0)	3065 (2.8)	<.001	1710 (3.3)	1698 (3.3)	.83
Data missing	11 135 (13.4)	11 121 (13.4)	.92	79 087 (13.3)	14 089 (12.8)	<.001	6959 (13.5)	6833 (13.3)	.25
Health care utilization
No. of inpatient admissions
Mean (SD)	1.29 (3.71)	1.32 (3.80)	.12	0.83 (2.95)	1.54 (4.03)	<.001	1.43 (3.99)	1.44 (3.99)	.35
Median (IQR)	0 (0-0)	0 (0-0)	.45	0 (0-0)	0 (0-0)	<.001	0 (0-0)	0 (0-0)	.13
No. of outpatient encounters
Mean (SD)	12.0 (19.7)	12.0 (19.4)	.93	9.6 (15.7)	12.6 (20.6)	<.001	12.0 (18.3)	12.0 (18.0)	.84
Median (IQR)	7 (3-14)	7 (3-14)	.13	5 (2-11)	7 (3-15)	<.001	7 (3-15)	7 (3-14)	.74
Immunization	32 553 (39.2)	32 486 (39.1)	.74	237 284 (39.8)	42 705 (38.8)	<.001	21 279 (41.3)	21 265 (41.3)	.93
Rehabilitation	8037 (9.7)	8040 (9.7)	.98	40 612 (6.8)	11 345 (10.3)	<.001	5088 (9.9)	5090 (9.9)	.98
Comorbidities^a
Obesity by ICD-9 codes^b	19 084 (23.0)	18 941 (22.8)	.40	148 522 (24.9)	24 267 (22.0)	<.001	12 908 (25.1)	12 828 (24.9)	.57
Diabetes mellitus	43 221 (52.1)	42 931 (51.7)	.15	371 324 (62.4)	51 136 (46.4)	<.001	51 467 (100)	51 467 (100)	>.99
Diabetes with complications	10 203 (12.3)	10 038 (12.1)	.22	82 454 (13.8)	11 964 (10.9)	<.001	10 212 (19.8)	10 142 (19.7)	.58
Ketoacidosis or uncontrolled	3860 (4.7)	3799 (4.6)	.48	33 222 (5.6)	4534 (4.1)	<.001	3781 (7.4)	3825 (7.4)	.60
Kidney manifestations	824 (1.0)	823 (1.0)	.98	7032 (1.2)	934 (0.9)	<.001	827 (1.6)	815 (1.6)	.77
Ophthalmic manifestations	1633 (2.0)	1624 (2.0)	.87	12 933 (2.2)	1943 (1.8)	<.001	1680 (3.3)	1638 (3.2)	.46
Neurologic manifestations	3888 (4.7)	3795 (4.6)	.28	28 536 (4.8)	4524 (4.1)	<.001	3884 (7.6)	3814 (7.4)	.41
Circulatory manifestations	341 (0.4)	347 (0.4)	.82	2575 (0.4)	405 (0.4)	.002	362 (0.7)	358 (0.7)	.88
Unspecified complications	1206 (1.5)	1197 (1.4)	.85	6522 (1.1)	862 (0.8)	<.001	697 (1.4)	744 (1.5)	.21
Valvular heart disease	2266 (2.7)	2251 (2.7)	.82	15 811 (2.7)	2958 (2.7)	.58	1454 (2.8)	1485 (2.9)	.56
Peri/endo/myocarditis; cardiomyopathy	1056 (1.3)	1052 (1.3)	.93	8129 (1.4)	1321 (1.2)	<001	721 (1.4)	732 (1.4)	.77
Hypertension	53 566 (64.5)	53 431 (64.4)	.49	416 803 (70.0)	68 260 (61.9)	<.001	35 770 (69.5)	35 820 (69.6)	.74
Secondary hypertension/hypertension complication	1822 (2.2)	1833 (2.2)	.86	12 371 (2.1)	2309 (2.1)	.70	1404 (2.7)	1366 (2.70)	.46
Acute myocardial infarction	266 (0.3)	249 (0.3)	.45	6298 (1.1)	261 (0.2)	<.001	184 (0.4)	181 (0.4)	.88
Coronary atherosclerosis	10 165 (12.2)	10 022 (12.1)	.28	114 541 (19.2)	11 505 (10.4)	<.001	7290 (14.2)	7352 (14.3)	.58
Nonspecific chest pain	5828 (7.0)	5819 (7.0)	.93	39 836 (6.7)	7977 (7.2)	<.001	3356 (6.5)	3394 (6.6)	.63
Pulmonary heart disease	767 (0.9)	711 (0.9)	.14	4437 (0.7)	943 (0.9)	<.001	505 (1.0)	514 (1.0)	.78
Ill-defined heart disease	1374 (1.7)	1325 (1.6)	.34	10 450 (1.8)	1757 (1.6)	<.001	831 (1.6)	856 (1.7)	.54
Conduction disorders	1732 (2.1)	1628 (2.0)	.07	12 021 (2.0)	2035 (1.9)	<.001	1168 (2.3)	1177 (2.3)	.85
Cardiac dysrhythmias	6898 (8.3)	6852 (8.3)	.68	46 382 (7.8)	8836 (8.0)	.01	4618 (9.0)	4595 (8.9)	.80
Cardiac arrest/ventricular fibrillation	30 (0.04)	39 (0.05)	.28	297 (0.05)	47 (0.04)	.32	23 (0.04)	26 (0.05)	.67
Congestive heart failure	3409 (4.1)	3369 (4.1)	.62	25 932 (4.4)	4136 (3.8)	<.001	2583 (5.0)	2511 (4.9)	.30
Acute cerebrovascular disease	1835 (2.2)	1818 (2.2)	.78	17 628 (3.0)	2095 (1.9)	<.001	1241 (2.4)	1259 (2.5)	.72
Cerebrovascular disease/transient cerebral ischemia	1264 (1.5)	1247 (1.5)	.73	12 402 (2.1)	1450 (1.3)	<.001	856 (1.7)	847 (1.7)	.83
Peripheral/visceral atherosclerosis	2699 (3.3)	2663 (3.2)	.62	23 125 (3.9)	3164 (2.9)	<.001	1896 (3.7)	1880 (3.7)	.79
Arterial aneurysms	725 (0.9)	719 (0.9)	.87	5589 (0.9)	883 (0.8)	<.001	504 (1.0)	469 (0.9)	.26
Aortic/peripheral arterial embolism/thrombosis	137 (0.2)	122 (0.2)	.35	936 (0.2)	168 (0.2)	.72	75 (0.2)	84 (0.2)	.48
COPD/bronchiectasis	10 076 (12.1)	10 012 (12.1)	.63	60 682 (10.2)	13 590 (12.3)	<.001	6041 (11.7)	5980 (11.6)	.55
Asthma	3541 (4.3)	3615 (4.4)	.37	21 187 (3.6)	5011 (4.6)	<.001	2102 (4.1)	2085 (4.1)	.79
Respiratory failure	536 (0.7)	559 (0.7)	.49	2582 (0.4)	793 (0.7)	<.001	432 (0.8)	461 (0.9)	.33
Nephritis/ nephrosis/renal sclerosis/chronic kidney disease	3266 (3.9)	3317 (4.0)	.52	21 702 (3.6)	4084 (3.7)	.31	2687 (5.2)	2585 (5.0)	.15
Acute/unspecified kidney failure	1728 (2.1)	1739 (2.1)	.85	8714 (1.5)	2514 (2.3)	<.001	1393 (2.7)	1421 (2.8)	.59
Rheumatoid arthritis/SLE/connective tissue disorders	1303 (1.6)	1326 (1.6)	.65	6226 (1.1)	1840 (1.7)	<.001	753 (1.5)	769 (1.5)	.68
Pathological fracture	78 (0.1)	86 (0.1)	.53	290 (0.05)	129 (0.1)	<.001	44 (0.1)	49 (0.1)	.60
Schizophrenia/psychosis	2922 (3.5)	2919 (3.5)	.97	15 130 (2.5)	4178 (3.8)	<.001	1521 (3.0)	1595 (3.1)	.18
Suicide/intentional self-inflicted injury	841 (1.0)	864 (1.0)	.58	3617 (0.6)	1240 (1.1)	<.001	485 (0.9)	476 (0.9)	.77
Severe liver disease^c	492 (0.6)	496 (0.6)	.90	945 (0.2)	1394 (1.3)	<.001	357 (0.7)	414 (0.8)	.04
Malignant neoplam^c	7069 (8.5)	7106 (8.6)	.75	39 941 (6.7)	9714 (8.8)	<.001	4639 (9.0)	4690 (9.1)	.58
Metastatic neoplasm^c	380 (0.5)	395 (0.5)	.59	1226 (0.2)	709 (0.6)	<.001	286 (0.6)	303 (0.6)	.48
AIDS^c	474 (0.6)	502 (0.6)	.37	1872 (0.3)	766 (0.7)	<.001	267 (0.5)	257 (0.5)	.66
Comorbidity score
Charlson comorbidity index total score^d
Mean (SD)	1.29 (1.42)	1.29 (1.43)	>.99	1.29 (1.25)	1.27 (1.50)	<.001	1.7 (1.4)	1.7 (1.4)	.76
Median (IQR)	1 (0-2)	1 (0-2)	.49	1 (1-2)	1 (0-2)	<.001	1 (1-2)	1 (1-2)	.58
Cardiovascular risk^e
<5%	19 871 (23.9)	19 771 (23.8)	.57	84 226 (14.1)	30 138 (27.4)	<.001	7221 (14.0)	7150 (13.9)	.52
5% to <10%	15 726 (18.9)	15 731 (19.0)	.98	103 753 (17.4)	21 310 (19.3)	<.001	8605 (16.7)	8659 (16.8)	.65
10% to <15%	18 740 (22.6)	18 916 (22.8)	.30	146 763 (24.6)	24 217 (22.0)	<.001	12 871 (25.0)	12 840 (25.0)	.82
15% to <20%	15 958 (19.2)	15 944 (19.2)	.93	141 990 (23.8)	19 215 (17.4)	<.001	12 593 (24.5)	12 602 (24.5)	.95
20% to <25%	7684 (9.3)	7614 (9.2)	.55	76 949 (12.9)	8712 (7.9)	<.001	6791 (13.2)	6815 (13.2)	.83
25% to <30%	1802 (2.2)	1726 (2.1)	.20	19 921 (3.3)	1884 (1.7)	<.001	1661 (3.2)	1657 (3.2)	.94
≥30%	139 (0.2)	146 (0.2)	.68	1895 (0.3)	151 (0.1)	<.001	155 (0.3)	140 (0.3)	.38
Missing	3102 (3.7)	3174 (3.8)	.35	20 082 (3.4)	4568 (4.2)	<.001	1570 (3.1)	1604 (3.1)	.54
Laboratory testing (for mmol/L, multiply by 0.0555)
Glucose in blood, mean (SD), mg/dL	133 (49)	133 (51)	.20	142 (57)	130 (48)	<.001	149 (56)	149 (57)	.62
Serum creatinine, mean (SD), mg/dL	1.11 (0.6)	1.11 (0.6)	.33	1.10 (0.4)	1.10 (0.6)	.02	1.1 (0.6)	1.1 (0.6)	.34
≥1 Blood glucose level ≥200 mg/dL	16 536 (19.9)	16 481 (19.9)	.74	133 872 (22.5)	20 994 (19.1)	<.001	16 715 (32.5)	16 918 (32.9)	.18
>5 Test results with blood glucose ≥200 mg/dL	3911 (4.7)	3934 (4.7)	.79	22 384 (3.8)	5558 (5.0)	<.001	3964 (7.7)	4069 (7.9)	.22
Mean eGFR, mL/min/1.73 m²
>90	22 231 (26.8)	22 013 (26.5)	.23	139 503 (23.4)	30 237 (27.4)	<.001	14 274 (27.7)	14 103 (27.4)	.23
60-89	45 693 (55.0)	45 868 (55.2)	.39	343 528 (57.7)	61 098 (55.5)	<.001	26 700 (51.9)	26 953 (52.4)	.11
45-59	10 282 (12.4)	10 341 (12.5)	.66	80 060 (13.4)	12 855 (11.7)	<.001	6818 (13.3)	6770 (13.2)	.66
30-44	3419 (4.1)	3356 (4.0)	.44	24 484 (4.1)	4104 (3.7)	<.001	2586 (5.0)	2530 (4.9)	.42
15-29	939 (1.1)	938 (1.1)	.98	5939 (1.0)	1205 (1.1)	.003	718 (1.4)	741 (1.4)	.54
<15	458 (0.6)	506 (0.6)	.12	2065 (0.35)	696 (0.6)	<.001	371 (0.7)	370 (0.7)	.97
eGFR, mean (SD)	78.36 (22.66)	78.34 (22.86)	.89	76.93 (21.18)	79.07 (23.01)	<.001	77.9 (23.6)	78.0 (24.0)	.57
Glucose-lowering medication classes^f
Metformin	15 742 (19.0)	15 608 (18.8)	.40	145 324 (24.4)	18 169 (16.5)	<.001	15 590 (30.3)	15 597 (30.3)	.96
Sulphonylurea	11 267 (13.6)	11 180 (13.5)	.53	98 344 (16.5)	13 440 (12.2)	<.001	11 384 (22.1)	11 276 (21.9)	.42
GLP1	15 (0.02)	16 (0.02)	.86	149 (0.03)	19 (0.02)	.12	26 (0.05)	18 (0.03)	.23
DDP4	91 (0.1)	95 (0.1)	.77	649 (0.1)	98 (0.1)	.06	116 (0.2)	95 (0.2)	.15
Thiazolidinediones	1201 (1.5)	1214 (1.5)	.79	13 460 (2.3)	1483 (1.4)		1167 (2.3)	1214 (2.4)	.33
α-Glucosidase inhibitors	1 (< .1)	1 (< .1)	>.99	28 (< .1)	1 (< .1)	.07	2 (< .1)	1 (< .1)	.56
Amylin analog	2 (< .1)	3 (< .1)	.66	20 (< .1)	3 (< .1)	.73	1 (< .1)	3 (< .1)	.32
SGLT2	0	0	NA	2 (0.0)	0	.54	1 (0.0)	0	.32
Any insulin	6763 (8.2)	6759 (8.1)	.97	56 007 (9.4)	7834 (7.1)	<.001	6683 (13.0)	6778 (13.2)	.38
No. of glucose-lowering classes
Mean (SD)	0.42 (0.74)	0.42 (0.74)	.50	0.53 (0.79)	0.37 (0.71)	<.001	0.68 (0.85)	0.68 (0.84)	.96
Median (IQR)	0 (0-1)	0 (0-1)	.25	0 (0-1)	0 (0-1)	<.001	0 (0-1)	0 (0-1)	.88
Other medication groups
ACEI	26 553 (32.0)	26 395 (31.8)	.41	214 428 (36.0)	33 482 (30.4)	<.001	18 890 (36.7)	19 031 (37.0)	.36
ARB	4237 (5.1)	4238 (5.1)	.99	31 076 (5.2)	5200 (4.7)	<.001	3311 (6.4)	3219 (6.3)	.24
β-Blockers	16 744 (20.2)	16 670 (20.1)	.65	133 168 (22.4)	21 385 (19.4)	<.001	10 565 (20.5)	10 652 (20.7)	.50
Nonloop diuretic	20 103 (24.2)	20 186 (24.3)	.64	148 691 (25.0)	26 617 (24.0)	<.001	12 590 (24.5)	12 549 (24.4)	.77
Loop diuretic	5908 (7.1)	5883 (7.1)	.81	41 169 (6.9)	7717 (7.0)	<.001	4279 (8.3)	4260 (8.3)	.83
Other antihypertensive agents^g	9138 (11.0)	9272 (11.2)	.30	58 503 (9.8)	12 028 (10.9)	<.001	5637 (11.0)	5600 (10.9)	.71
Antiarrhythmic	2886 (3.5)	2915 (3.5)	.70	21 710 (3.7)	3662 (3.3)	<.001	1850 (3.6)	1849 (3.6)	.99
Antithrombotic	2681 (3.2)	2684 (3.2)	.97	19 619 (3.3)	3399 (3.1)	<.001	1748 (3.4)	1796 (3.5)	.41
Antipsychotic	2888 (3.5)	2904 (3.5)	.83	16 001 (2.7)	3995 (3.6)	<.001	1508 (2.9)	1565 (3.0)	.30
Dopamine agonist	608 (0.7)	615 (0.7)	.84	3566 (0.6)	776 (0.7)	<.001	418 (0.8)	420 (0.8)	.95
Peripheral vascular disease^h	335 (0.4)	319 (0.4)	.53	2570 (0.4)	382 (0.4)	<.001	228 (0.4)	201 (0.4)	.19
Antismoking medications	4069 (4.9)	4138 (5.0)	.44	25 841 (4.3)	5714 (5.2)	<.001	2354 (4.6)	2347 (4.6)	.92
Nonstatin lipid-lowering	6950 (8.4)	7045 (8.5)	.40	59 670 (10.0)	8785 (8.0)	<.001	5115 (9.9)	5041 (9.8)	.44
Cardiovascular procedures
Electrocardiography	14 951 (18.0)	14 932 (18.0)	.90	103 775 (17.4)	20 690 (18.8)	<.001	9343 (18.2)	9451 (18.4)	.38
Echocardiography	4462 (5.4)	4404 (5.3)	.53	30 270 (5.1)	5926 (5.4)	<.001	3048 (5.9)	3086 (6.0)	.62
Stress test	2050 (2.5)	2009 (2.4)	.52	15 763 (2.7)	2704 (2.5)	<.001	1243 (2.4)	1268 (2.5)	.61
Cardiac catheterization	105 (0.1)	110 (0.1)	.73	1693 (0.3)	115 (0.1)	<.001	91 (0.2)	89 (0.2)	.88
PCI	45 (< .1)	41 (< .1)	.67	1493 (0.3)	41 (< .1)	<.001	34 (0.1)	34 (0.1)	>.99
CABG	0	1 (< .1)	NA	49 (< .1)	1 (0.0)	.008	0	1 (< .1)	.32
Pacemaker/defibrillator	69 (0.1)	69 (0.1)	>.99	497 (0.1)	85 (0.1)	.50	54 (0.1)	52 (0.10)	.85
Peripheral arterial revascularization procedures	8 (< .1)	7 (< .1)	.80	44 (0.01)	8 (< .1)	.96	2 (< .1)	5 (< .1)	.26
Baseline characteristics not included in the propensity score, mean (SD)ⁱ
Total cholesterol^j	195 (44)	175 (39)	<.001	197 (45)	178 (40)	<.001	186 (43)	170 (40)	<.001
LDL cholesterol	119 (38)	101 (31)	<.001	121 (38)	103 (32)	<.001	111 (35)	96 (31)	<.001
HDL cholesterol^k	43 (12)	41 (13)	<.001	41 (12)	42 (14)	<.001	42 (21)	41 (13)	<.001

Open in a new tab

Abbreviations: ACEI, angiotensin converting enzyme inhibitors; ARB, angiotensin-receptor blockers; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CABG, coronary artery bypass graft surgery; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular diseases; DPP 4, dipeptidyl peptidase 4 inhibitors; eGFR, estimated glomerular filtration rate using the Modification of Diet in Renal Disease Study equation³¹; GLP-1, glucagon-like peptide 1 agonists; HDL, high-density lipoprotein; LDL, low-density lipoprotein; PCI, percutaneous coronary intervention; SGLT2, sodium glucose cotransporter 2 inhibitors; SLE, systemic lupus erythematosus.

^{^a}

Diagnoses and procedures as defined by the Agency for Health Research and Quality Clinical Classifications Software disease categories. See Supplement for details.²¹

^{^b}

Diagnosis of obesity by ICD-9 codes have a very high specificity (≈ 98%) but low sensitivity (≈ 30%).^29,30

^{^c}

Malignant neoplasm, metastatic neoplasm, and AIDS were defined using Deyo and colleagues’ method for calculating the Charlson comorbidity index.²⁷

^{^d}

The Charlson comorbidity total score was calculated using Deyo and colleagues’ method.²⁷

^{^e}

Cardiovascular risk was calculated using D’Agostino and colleagues’ method for calculating the Framingham risk score.²⁸

^{^f}

Approximately half of the cohort was diagnosed with diabetes at the index date.

^{^g}

Other antihypertensive agents include α-blocker medications, clonidine, α-methyldopa, hydralazine, minoxidil, and reserpine.

^{^h}

Peripheral vascular disease medications include pentoxifylline, cilostazol, papaverine, tolazoline, cyclandelate, and ethaverine.

^ⁱ

We did not include lipid values in propensity score creation because we expected patients who would be initiated on statins (statin users) to have worse lipid panel values.

^{^j}

Results for total cholesterol testing were available for only 82 566 statin users and 82 702 active comparators.

^{^k}

Results for HDL cholesterol testing were available for only 79 949 statin users and 79 873 active comparators.

Primary and Secondary Analyses

In the primary analysis, we compared the primary and secondary outcomes in the PS matched cohort using conditional logistic regression to calculate odds ratios (OR) and 95% CI. For the secondary analyses, we compared the primary outcome in the following cohorts:

1. Overall cohort: all patients who fulfilled inclusion and exclusion criteria of the study before PS matching.
2. Healthy cohort: patients with no comorbidities in the Charlson comorbidity index at baseline.
3. High-intensity cholesterol-lowering statin users (decrease of ≥50% in mean LDL cholesterol during follow-up vs baseline) compared with nonusers in the overall cohort.¹
4. Moderate-intensity cholesterol lowering statin users (decrease of <50% and ≥30% in mean LDL cholesterol during follow-up vs baseline) compared with nonusers in the overall cohort.
5. Low-intensity cholesterol lowering statin users (a decrease of <30% in mean LDL cholesterol during follow-up vs baseline) compared with nonusers in the overall cohort.
6. Ever user vs never user cohort: we excluded patients who started as active comparators and were crossed over to the statin user group.

Sensitivity Analysis

We excluded patients who were diagnosed with incident diabetes, diabetic complications, ketoacidosis or uncontrolled diabetes, or CVD events within 60 days from the index date. Because it is highly unlikely that statin use would influence these outcomes within 60 days of statin initiation, excluding these patients further mitigated confounding by indication or residual confounding.^18,19,20

Post Hoc Analysis

We created another PS matched cohort that included only patients who had diabetes during the baseline period (PS matched prevalent diabetes cohort). All variables and techniques that were used in the primary analysis were included in this cohort. A caliper of 0.00002 was used to achieve balance-maximizing sample size (Table 1).

Statistical Analysis

Dichotomous variables were compared using χ², and continuous variables were compared using t tests. When the Kolmogorov-Smirnov test indicated unequal distribution, we used the Wilcoxon Mann-Whitney test. Statistical tests were 2-tailed, and significance was defined as P < .05.

Because multiple comparisons in the secondary analyses may produce a type I error, findings of secondary analyses should be interpreted as exploratory. Secondary and sensitivity analyses were performed using a separate logistic regression model for each dichotomous outcome adjusting for the PS. Data management and statistical analyses were conducted from September 2019 to October 2020 using STATA, version 15 (Stata Corp LLC).

Results

From the 705 774 eligible VA patients (595 579 statin users and 110 195 nonusers), we matched 83 022 pairs of statin users and active comparators; the matched cohort had a mean (SD) age of 60.1 (11.6) years; 78 712 (94.9%) were men; 1715 (2.1%) were American Indian/Pacific Islander/Alaska Native, 570 (0.8%) were Asian, 17 890 (21.5%) were Black, and 56 633 (68.2 %) were White individuals. To form the study cohort, we had excluded 441 778 patients who were prevalent users of statins, H2, or PPI; 6613 who were less than 30 years old; 36 085 whose follow-up period had been fewer than 60 days; 421 170 who were not regular VA users; and 1098 whose records were missing data on age (eFigure in the Supplement).

Statin users had filled prescriptions for statins for a mean (SD) duration of 5.3 (3.3) years; median (IQR) of 5.1 (2.6-8.0) years. Statin users filled 12 118 523 prescriptions for statins throughout the study; 63.4% of the prescriptions were for simvastatin, 12.4% for atorvastatin, 10.5% for rosuvastatin, and 9.5% for pravastatin. Among nonusers, 52 176 patients (47.4%) subsequently used a statin during the study period and 58 019 (52.7%) never used a statin.

Among the 83 022 matched pairs of statin users and nonusers (Table 1) there was only a small difference in race and ethnicity distribution. Based on the cohort characteristics, all patients had been diagnosed with diabetes by the end of the study period. At baseline, the statin user and nonuser groups in the PS matched cohort had similar proportions of patients with diabetes, diabetes complications, recurrent episodes of high blood glucose (≥200 mg/dL), and using glucose-lowering agents; they also had similar mean blood glucose levels and patient follow-up periods. During the follow-up period, the statin users decreased their mean LDL cholesterol by a mean (SD) of 25 (31.6) mg/dL compared with a 0.8 (23.7) mg/dL among nonusers (95% CI of mean difference, 23.9-24.4; P < .001), demonstrating that statin users had effectively used, not just filled, their statin prescriptions (eTable 5 in the Supplement). The mean (SD) number of outpatient encounters by the PS matched cohort during the follow-up period was 37.8 (43.1) for the statin user group and 41.4 (50.5) encounters for the nonuser group.

Primary Analysis

Statin users had significantly higher odds of diabetes progression (OR, 1.37; 95% CI, 1.35-1.40) compared with nonusers. There was significantly higher rate of each component of the diabetes progression outcome in statin users compared with nonusers (Table 2) including an increase in the number of glucose-lowering medication classes (OR, 1.41; 95% CI, 1.38-1.43), new insulin starts (OR, 1.16; 95% CI, 1.12-1.19), presence of persistent hyperglycemia (OR, 1.13; 95% CI, 1.10-1.16), and new diagnosis of ketoacidosis or uncontrolled diabetes (OR, 1.24; 95% CI, 1.19-1.30).

Table 2. Odds of Outcomes During Follow-up Period Between Statin Users and Active Comparators in Propensity Score–Matched Cohorts.

Outcomes	Overall cohort				Diabetes-prevalent cohort
	No. (%)		OR (95% CI)	P value	No. (%)		OR (95% CI)	P value
	Statin users (n = 83 022)	Active comparators (n = 83 022)	OR (95% CI)	P value	Statin users (n = 51 467)	Active comparators (n = 51 467)	OR (95% CI)	P value
Primary
Diabetes progression	46 434 (55.9)	39 868 (48.0)	1.37 (1.35 to 1.40)	<.001	30 494 (59.3)	27 189 (52.8)	1.30 (1.27 to 1.33)	<.001
Secondary
Components of diabetes progression outcome
New insulin starts during follow-up	11 947 (14.4)	10 540 (12.7)	1.16 (1.12 to 1.19)	<.001	9081 (17.6)	8215 (16.0)	1.13 (1.09 to 1.17)	<.001
Increased No. of glucose-lowering classes	42 579 (51.3)	35 533 (42.8)	1.41 (1.38 to 1.43)	<.001	27 299 (53.0)	23 443 (45.6)	1.35 (1.32 to 1.38)	<.001
Incident ≥5 measurements with blood glucose ≥200 mg/dL^a	13 963 (16.8)	12 601 (15.2)	1.13 (1.10 to 1.16)	<.001	9858 (19.2)	9167 (17.8)	1.09 (1.06 to 1.13)	<.001
Incident diabetes with ketoacidosis/uncontrolled diabetes	4468 (5.4)	3635 (4.4)	1.24 (1.19 to 1.30)	<.001	2837 (5.5)	2427 (4.7)	1.18 (1.12 to 1.25)	<.001
Difference in No. of glucose-lowering classes during follow-up vs baseline
Mean (SD)	0.77 (0.99)	0.59 (0.89)	−0.19 to −0.17^b	<.001	0.78 (1.0)	0.62 (0.94)	−0.18 to −0.15^b	<.001
Median (IQR)	1 (0 to 1)	0 (0 to 1)	NA	<.001^c	1 (0 to 1)	0 (0 to 1)	NA	<.001
Decreased No. of glucose-lowering medication classes	2132 (2.6)	2276 (2.7)	0.94 (0.88 to 0.99)	.03	2116 (4.1)	2272 (4.4)	0.93 (0.87 to 0.99)	.02
Change in mean blood glucose during follow-up vs baseline, mg/dL^a
Mean (SD)	6.3 (45.4)	5.5 (46.1)	−1.24 to −0.36^b	<.001	0.32 (52.8)	−0.60 (52.8)	−1.6 to −0.28^b	.005
Median (IQR)	6.1 (−8.5 to 24.2)	5.7 (−8.8 to 23.0)	NA	<.001^c	2.8 (−17.7 to 23.8)	2.1 (−18.5 to 22.6)	NA	<.001

Open in a new tab

Abbreviations: NA, not applicable; OR, odds ratio.

^{^a}

To convert to mmol/L, multiply by 0.0555.

^{^b}

95% CI of difference.

^{^c}

Calculated using Wilcoxon rank-sum test.

Secondary Analysis

Odds of diabetes progression outcome were consistently higher in statin users compared with nonusers among all secondary and sensitivity analyses (Table 3). The odds of diabetes progression among statin users in the healthy cohort were higher than the overall cohort (OR, 1.56 vs 1.40) (Table 3). Intensive cholesterol lowering was associated with highest odds of diabetes progression outcome among statin users in compared with nonusers.

Table 3. Secondary Analysis and Sensitivity Analysis Comparing the Diabetes Progression Composite Outcome During Follow-up Between Statin Users and Active Comparators.

Cohort type	No. (%)		Adjusted OR (95% CI)^a	P value
Cohort type	Statin users	Active comparators	Adjusted OR (95% CI)^a	P value
Overall cohort (all included patients)
No.	595 579	110 195	NA	NA
Diabetes progression	391 553 (65.7)	49 328 (44.8)	1.40 (1.38-1.42)	<.001
Healthy cohort (no comorbidities at baseline)
No.	148 509	39 009	NA	NA
Diabetes progression	95 612 (64.4)	16 146 (41.4)	1.56 (1.52-1.60)	<.001
Cholesterol-lowering intensity in overall cohort
High intensity
No.	38 823	110 195	NA	NA
Diabetes progression	26 547 (68.4)	49 328 (44.8)	1.83 (1.78-1.88)	<.001
Moderate intensity
No.	180 884	110 195	NA	NA
Diabetes progression	120 246 (66.5)	49 328 (44.8)	1.55 (1.53-1.58)	<.001
Low intensity
No.	375 872	110 195	NA	NA
Diabetes progression	244 760 (65.1)	49 328 (44.8)	1.45 (1.42-1.47)	<.001
Statin ever user vs never user cohort
No.	543 403	58 019	NA	NA
Diabetes progression	363 599 (66.9)	26 540 (45.7)	1.73 (1.70-1.76)	<.001
Sensitivity analysis^b
No.	353 065	77 657	NA	NA
Diabetes progression	217 808 (61.7)	31 814 (41.0)	1.36 (1.33-1.38)	<.001

Open in a new tab

Abbreviations: NA, not applicable; OR, odds ratio.

^{^a}

Odds ratio adjusted for propensity score.

^{^b}

Overall cohort after excluding patients with incident diabetes, diabetic complications, ketoacidosis, uncontrolled diabetes, or cardiovascular disease within 60 days from index date.

Post Hoc Analysis

In the PS-matched prevalent diabetes cohort, we matched 51 467 pairs of statin users and nonusers with only a small difference in proportion of men and some racial and ethnic minority subgroups (Table 1). The odds of the primary and secondary outcomes were significantly higher among statin users (Table 2).

Discussion

This study of a national cohort of VA patients with diabetes found that statin use was associated with an escalation of diabetes treatment, including a higher risk of initiating insulin and the use of more glucose-lowering medication classes. This escalation of diabetes treatment was associated with worse diabetes control, including new persistent hyperglycemia and acute glycemic complications. Moreover, there was a dose-response association between intensity of lowering LDL cholesterol and risk of the study outcomes, with higher intensity of LDL cholesterol-lowering associated with higher odds of diabetes progression. For example, the odds of diabetes progression among statin users vs nonusers were 1.83, 1.55, and 1.45 for high-, moderate-, and low-intensity cholesterol lowering, respectively.

Using the Bender and Blettner formula,³⁴ the number needed to be exposed to statins for 1 additional person to experience diabetes progression outcome was 13. Although the mean difference between blood glucose during follow-up and baseline among statin users in contrast with nonusers was modest, there was significant escalation in diabetes therapy that was not associated with better clinical outcomes. This statin-associated metabolic cost was not measured by the RCTs, which instead focused mainly on cardiovascular benefits.³⁵ From 2009 to 2015, annual emergency department visits for hyperglycemic crisis almost doubled, hospitalization increased by 73%, and related deaths increased by 55%.³⁶ This resurgence in diabetes-specific complications deserves attention and a call to scrutinize our practices and goals.³⁷

The higher risk of diabetes progression associated with statin use may seem less consequential, at least in the short and intermediate term, than the cardiovascular benefits of statin use, especially when used for secondary prevention. However, diabetes progression has long-term effects on quality of life and treatment burden, which warrant consideration when discussing the overall risk-benefit profile, especially when used for primary prevention. In this study, approximately 77% of the cohort had no known CVD at baseline. A recent cohort study³⁸ reported that patients with diabetes with 5 risk factors within a target control range (eg, LDL cholesterol <97 mg/dL) had little or no excess risk of CVD outcomes. Of note, 61.5% of patients in this cohort used statins, and HbA_1C was a stronger predictor than LDL cholesterol of risk of death from any cause. Moreover, some studies^39,40,41,42 have demonstrated that only patients who have had diabetes for at least 10 to 15 years have an increased risk of CVD events. In 1 of these studies,³⁹ only 31% of patients with diabetes and 66% of those with diabetes and coronary heart disease used statins at baseline.

Several RCTs,^3,4,43 Mendelian randomization studies,⁴⁴ observational studies,⁵ and animal studies^45,46 have demonstrated that statin use increases insulin resistance. The association of statin use with diabetes progression may be explained by its effect on insulin resistance (eDiscussion in the Supplement). Insulin resistance has been shown to increase the risk of diabetic complications,^10,47 endothelial dysfunction, inflammation, and increased platelet reactivity.^{48,49,50,51,52}

Strengths and Limitations

Strengths of this study are its large size and longitudinal follow-up period with detailed clinical data. Also of note, patients in the nonuser group did not receive statin therapy despite guidelines recommending statins, a finding consistent with other studies⁵³ that indicate that, in community practice, many patients do not receive statins according to guidelines. The present study’s definitions of the intensity of cholesterol lowering is guided by, but not identical to, the definitions used by the American College of Cardiology/American Heart Association (ACC/AHA) and the American Diabetes Association.¹ The ACC/AHA identifies high-intensity statins based on their expected efficacy in lowering LDL cholesterol by 50% or more. For this study, we defined intensive cholesterol lowering among statin users as decreasing mean LDL cholesterol during the follow-up period by 50% or more compared with the baseline mean. Similarly, our definitions for low- and moderate-intensity cholesterol lowering used the same cutoff limits as the ACC/AHA guidelines but used the actual decrease in mean LDL cholesterol during follow-up. Our approach incorporates a measure of adherence to treatment, not simply filling of prescriptions.

This study also had some limitations inherent to retrospective observational data; hence, there is always a chance of unmeasured residual confounding. It may be argued that statin users have closer follow-up, resulting in ascertainment bias. However, the outpatient encounters of the PS matched cohort during the follow-up period were not more frequent among statin users than among nonusers. Prior studies⁵⁴ have shown that adjusting to number of encounters that took place during follow-up period did not affect outcomes if baseline characteristics were adequately described.

It could be argued that the physicians who prescribed statins may have also attempted to aggressively control diabetes by adding more glucose-lowering agents; however, this would not explain the increased persistent hyperglycemia nor the ketoacidosis, which would be expected to decrease. It may also be argued that confounding by indication may have played a part in the study findings, specifically in patients with diabetic complications who might have been treated with statins and more aggressive antidiabetes therapy. However, the sensitivity analysis, which excluded patients who experienced any diabetic complications, ketoacidosis, uncontrolled diabetes, or CVD within fewer than 60 days from the index date, demonstrated an association of statin use with diabetes progression, with odds similar those of the primary analysis. We also could not ascertain whether the association of statin use with diabetes progression was because of the statin use or the lower LDL cholesterol—that is, statins are inseparable from their cholesterol-lowering effect. The definition of the study’s composite outcome also had some limitations; an increase in the number of glucose-lowering agents assumed that all agents had similar potency. Also, selecting 5 or more episodes of blood glucose seems arbitrary; nevertheless, the outcome was applied equally to both statin users and nonusers and all components of the composite outcome were directionally the same. The study data could not reliably differentiate between diabetes types because there is no reliable algorithm to accomplish this in a data set such as this,⁵⁵ and regardless, there is no reason to believe that it would have differentially affected statin users vs nonusers. We expect that the extensive matching of the 2 groups on baseline characteristics, including microvascular and macrovascular diabetes complications, may have acted as a surrogate for diabetes severity and chronicity. Patients covered by the VA were predominantly men, which may limit generalization; however, research has shown that men covered by the VA have characteristics similar to those of individuals covered by other insurance.⁵⁶ Lastly, the study protocol was not prepublished.

Conclusions

This large retrospective matched-cohort study found that statin use was associated with a higher risk of diabetes treatment escalation and an increased risk of hyperglycemic complications. This metabolic cost was not considered in RCTs of statins. Further research is needed to form a risk-tailored approach to balancing the cardiovascular benefits of statin therapy with its risk of diabetes progression.

Supplement.

eTable 1. Definition of glucose-lowering medication classes

eTable 2. Definition of administrative codes that defined outcomes

eTable 3. Administrative codes used in definitions of baseline characteristics

eTable 4. Laboratory investigation and vital signs handling

eMethods. Propensity score − matching details

eFigure. Study design and cohort assembly

eTable 5. Comparisons of changes in laboratory values from baseline to

follow-up periods

eDiscussion. Summary of some proposed mechanisms for increased insulin

resistance in association of statin use

Click here for additional data file.^{(567KB, pdf)}

References

1.Stone NJ, Robinson JG, Lichtenstein AH, et al. ; American College of Cardiology/American Heart Association Task Force on Practice Guidelines . 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63(25 Pt B):2889-2934. doi: 10.1016/j.jacc.2013.11.002 [DOI] [PubMed] [Google Scholar]
2.American Diabetes Association . Cardiovascular disease and risk management: standards of medical care in diabetes-2018. Diabetes Care. 2018;41(suppl 1):S86-S104. doi: 10.2337/dc18-S009 [DOI] [PubMed] [Google Scholar]
3.Holman RR, Paul S, Farmer A, Tucker L, Stratton IM, Neil HA; Atorvastatin in Factorial with Omega-3 EE90 Risk Reduction in Diabetes Study Group . Atorvastatin in Factorial with Omega-3 EE90 Risk Reduction in Diabetes (AFORRD): a randomised controlled trial. Diabetologia. 2009;52(1):50-59. doi: 10.1007/s00125-008-1179-5 [DOI] [PubMed] [Google Scholar]
4.Erlandson KM, Jiang Y, Debanne SM, McComsey GA. Rosuvastatin worsens insulin resistance in HIV-infected adults on antiretroviral therapy. Clin Infect Dis. 2015;61(10):1566-1572. doi: 10.1093/cid/civ554 [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Koh KK, Quon MJ, Han SH, Lee Y, Kim SJ, Shin EK. Atorvastatin causes insulin resistance and increases ambient glycemia in hypercholesterolemic patients. J Am Coll Cardiol. 2010;55(12):1209-1216. doi: 10.1016/j.jacc.2009.10.053 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Liew SM, Lee PY, Hanafi NS, et al. Statins use is associated with poorer glycaemic control in a cohort of hypertensive patients with diabetes and without diabetes. Diabetol Metab Syndr. 2014;6:53. doi: 10.1186/1758-5996-6-53 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Sukhija R, Prayaga S, Marashdeh M, et al. Effect of statins on fasting plasma glucose in diabetic and nondiabetic patients. J Investig Med. 2009;57(3):495-499. doi: 10.2310/JIM.0b013e318197ec8b [DOI] [PubMed] [Google Scholar]
8.Cederberg H, Stančáková A, Yaluri N, Modi S, Kuusisto J, Laakso M. Increased risk of diabetes with statin treatment is associated with impaired insulin sensitivity and insulin secretion: a 6 year follow-up study of the METSIM cohort. Diabetologia. 2015;58(5):1109-1117. doi: 10.1007/s00125-015-3528-5 [DOI] [PubMed] [Google Scholar]
9.Ridker PM, MacFadyen J, Cressman M, Glynn RJ. Efficacy of rosuvastatin among men and women with moderate chronic kidney disease and elevated high-sensitivity C-reactive protein: a secondary analysis from the JUPITER (Justification for the Use of Statins in Prevention-an Intervention Trial Evaluating Rosuvastatin) trial. J Am Coll Cardiol. 2010;55(12):1266-1273. doi: 10.1016/j.jacc.2010.01.020 [DOI] [PubMed] [Google Scholar]
10.Kilpatrick ES, Rigby AS, Atkin SL. Insulin resistance, the metabolic syndrome, and complication risk in type 1 diabetes: “double diabetes” in the Diabetes Control and Complications Trial. Diabetes Care. 2007;30(3):707-712. doi: 10.2337/dc06-1982 [DOI] [PubMed] [Google Scholar]
11.Brownlee M. The pathobiology of diabetic complications: a unifying mechanism. Diabetes. 2005;54(6):1615-1625. doi: 10.2337/diabetes.54.6.1615 [DOI] [PubMed] [Google Scholar]
12.Maki KC, Ridker PM, Brown WV, Grundy SM, Sattar N; The Diabetes Subpanel of the National Lipid Association Expert Panel . An assessment by the Statin Diabetes Safety Task Force: 2014 update. J Clin Lipidol. 2014;8(3)(suppl):S17-S29. doi: 10.1016/j.jacl.2014.02.012 [DOI] [PubMed] [Google Scholar]
13.Health Services Research and Development. Accessed October 22, 2020. https://www.hsrd.research.va.gov/for_researchers/vinci/cdw.cfm
14.Miller DR, Safford MM, Pogach LM. Who has diabetes? Best estimates of diabetes prevalence in the Department of Veterans Affairs based on computerized patient data. Diabetes Care. 2004;27(suppl 2):B10-B21. doi: 10.2337/diacare.27.suppl_2.B10 [DOI] [PubMed] [Google Scholar]
15.Alvarez CA, Halm EA, Pugh MJV, et al. Lactic acidosis incidence with metformin in patients with type 2 diabetes and chronic kidney disease: A retrospective nested case-control study. Endocrinol Diabetes Metab. 2020;4(1):e00170. doi: 10.1002/edm2.170 [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Lund JL, Richardson DB, Stürmer T. The active comparator, new user study design in pharmacoepidemiology: historical foundations and contemporary application. Curr Epidemiol Rep. 2015;2(4):221-228. doi: 10.1007/s40471-015-0053-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Yoshida K, Solomon DH, Kim SC. Active-comparator design and new-user design in observational studies. Nat Rev Rheumatol. 2015;11(7):437-441. doi: 10.1038/nrrheum.2015.30 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Danaei G, Rodríguez LA, Cantero OF, Logan R, Hernán MA. Observational data for comparative effectiveness research: an emulation of randomised trials of statins and primary prevention of coronary heart disease. Stat Methods Med Res. 2013;22(1):70-96. doi: 10.1177/0962280211403603 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Jee SH, Sull JW, Park J, et al. Body-mass index and mortality in Korean men and women. N Engl J Med. 2006;355(8):779-787. doi: 10.1056/NEJMoa054017 [DOI] [PubMed] [Google Scholar]
20.Allison DB, Faith MS, Heo M, Townsend-Butterworth D, Williamson DF. Meta-analysis of the effect of excluding early deaths on the estimated relationship between body mass index and mortality. Obes Res. 1999;7(4):342-354. doi: 10.1002/j.1550-8528.1999.tb00417.x [DOI] [PubMed] [Google Scholar]
21.Elixhauser A, Steiner C, Palmer L. Clinical Classifications Software (CCS) for ICD-9-CM. Databases and related tools from the Healthcare Cost and Utilization Project (HCUP): appendix a. 2012. Accessed August 30, 2021. https://www.hcup-us.ahrq.gov/toolssoftware/ccs/AppendixASingleDX.txt
22.Henderson DC, Cagliero E, Copeland PM, et al. Elevated hemoglobin A1c as a possible indicator of diabetes mellitus and diabetic ketoacidosis in schizophrenia patients receiving atypical antipsychotics. J Clin Psychiatry. 2007;68(4):533-541. doi: 10.4088/JCP.v68n0407 [DOI] [PubMed] [Google Scholar]
23.Wetterhall SF, Olson DR, DeStefano F, et al. Trends in diabetes and diabetic complications, 1980-1987. Diabetes Care. 1992;15(8):960-967. doi: 10.2337/diacare.15.8.960 [DOI] [PubMed] [Google Scholar]
24.Christakis DA, Feudtner C, Pihoker C, Connell FA. Continuity and quality of care for children with diabetes who are covered by medicaid. Ambul Pediatr. 2001;1(2):99-103. doi: [DOI] [PubMed] [Google Scholar]
25.Khokhar B, Jette N, Metcalfe A, et al. Systematic review of validated case definitions for diabetes in ICD-9-coded and ICD-10-coded data in adult populations. BMJ Open. 2016;6(8):e009952. doi: 10.1136/bmjopen-2015-009952 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Saydah SH, Geiss LS, Tierney E, Benjamin SM, Engelgau M, Brancati F. Review of the performance of methods to identify diabetes cases among vital statistics, administrative, and survey data. Ann Epidemiol. 2004;14(7):507-516. doi: 10.1016/j.annepidem.2003.09.016 [DOI] [PubMed] [Google Scholar]
27.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. doi: 10.1016/0895-4356(92)90133-8 [DOI] [PubMed] [Google Scholar]
28.D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117(6):743-753. doi: 10.1161/CIRCULATIONAHA.107.699579 [DOI] [PubMed] [Google Scholar]
29.Quan H, Li B, Saunders LD, et al. ; IMECCHI Investigators . Assessing validity of ICD-9-CM and ICD-10 administrative data in recording clinical conditions in a unique dually coded database. Health Serv Res. 2008;43(4):1424-1441. doi: 10.1111/j.1475-6773.2007.00822.x [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Goff SL, Pekow PS, Markenson G, Knee A, Chasan-Taber L, Lindenauer PK. Validity of using ICD-9-CM codes to identify selected categories of obstetric complications, procedures and co-morbidities. Paediatr Perinat Epidemiol. 2012;26(5):421-429. doi: 10.1111/j.1365-3016.2012.01303.x [DOI] [PubMed] [Google Scholar]
31.Stevens LA, Coresh J, Feldman HI, et al. Evaluation of the modification of diet in renal disease study equation in a large diverse population. J Am Soc Nephrol. 2007;18(10):2749-2757. doi: 10.1681/ASN.2007020199 [DOI] [PubMed] [Google Scholar]
32.Becker SO, Ichino A. Estimation of average treatment effects based on propensity scores. Stata J. 2002;2(4):358-377. doi: 10.1177/1536867X0200200403 [DOI] [Google Scholar]
33.Leuven E, Sianesi B. PSMATCH2: Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing, v 4.0.5. 2003. Accessed August 27, 2021. https://ideas.repec.org/c/boc/bocode/s432001.html
34.Bender R, Blettner M. Calculating the “number needed to be exposed” with adjustment for confounding variables in epidemiological studies. J Clin Epidemiol. 2002;55(5):525-530. doi: 10.1016/S0895-4356(01)00510-8 [DOI] [PubMed] [Google Scholar]
35.Mansi IA. Statins in primary prevention: uncertainties and gaps in randomized trial data. Am J Cardiovasc Drugs. 2016;16(6):407-418. doi: 10.1007/s40256-016-0190-3 [DOI] [PubMed] [Google Scholar]
36.US Centers for Disease Control and Prevention . US diabetes surveillance system and diabetes atlas. Assessed December 9, 2020. https://www.cdc.gov/diabetes/data
37.Gregg EW, Hora I, Benoit SR. Resurgence in diabetes-related complications. JAMA. 2019;321(19):1867-1868. doi: 10.1001/jama.2019.3471 [DOI] [PubMed] [Google Scholar]
38.Rawshani A, Rawshani A, Franzén S, et al. Risk factors, mortality, and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2018;379(7):633-644. doi: 10.1056/NEJMoa1800256 [DOI] [PubMed] [Google Scholar]
39.Rana JS, Liu JY, Moffet HH, Jaffe M, Karter AJ. Diabetes and prior coronary heart disease are not necessarily risk equivalent for future coronary heart disease events. J Gen Intern Med. 2016;31(4):387-393. doi: 10.1007/s11606-015-3556-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Hu FB, Stampfer MJ, Solomon CG, et al. The impact of diabetes mellitus on mortality from all causes and coronary heart disease in women: 20 years of follow-up. Arch Intern Med. 2001;161(14):1717-1723. doi: 10.1001/archinte.161.14.1717 [DOI] [PubMed] [Google Scholar]
41.Wannamethee SG, Shaper AG, Whincup PH, Lennon L, Sattar N. Impact of diabetes on cardiovascular disease risk and all-cause mortality in older men: influence of age at onset, diabetes duration, and established and novel risk factors. Arch Intern Med. 2011;171(5):404-410. doi: 10.1001/archinternmed.2011.2 [DOI] [PubMed] [Google Scholar]
42.Idris I. Diabetes and cardiovascular risk equivalency: do age at diagnosis and disease duration affect risk stratification?: comment on “impact of diabetes on cardiovascular disease risk and all-cause mortality in older men”. Arch Intern Med. 2011;171(5):410-411. doi: 10.1001/archinternmed.2010.524 [DOI] [PubMed] [Google Scholar]
43.Colhoun HM, Betteridge DJ, Durrington PN, et al. ; CARDS investigators . Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): multicentre randomised placebo-controlled trial. Lancet. 2004;364(9435):685-696. doi: 10.1016/S0140-6736(04)16895-5 [DOI] [PubMed] [Google Scholar]
44.Swerdlow DI, Preiss D, Kuchenbaecker KB, et al. ; DIAGRAM Consortium; MAGIC Consortium; InterAct Consortium . HMG-coenzyme A reductase inhibition, type 2 diabetes, and bodyweight: evidence from genetic analysis and randomised trials. Lancet. 2015;385(9965):351-361. doi: 10.1016/S0140-6736(14)61183-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Henriksbo BD, Lau TC, Cavallari JF, et al. Fluvastatin causes NLRP3 inflammasome-mediated adipose insulin resistance. Diabetes. 2014;63(11):3742-3747. doi: 10.2337/db13-1398 [DOI] [PubMed] [Google Scholar]
46.Mitchell P, Marette A. Statin-induced insulin resistance through inflammasome activation: sailing between Scylla and Charybdis. Diabetes. 2014;63(11):3569-3571. doi: 10.2337/db14-1059 [DOI] [PubMed] [Google Scholar]
47.Buse MG. Hexosamines, insulin resistance, and the complications of diabetes: current status. Am J Physiol Endocrinol Metab. 2006;290(1):E1-E8. doi: 10.1152/ajpendo.00329.2005 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Kernan WN, Viscoli CM, Furie KL, et al. ; IRIS Trial Investigators . Pioglitazone after ischemic stroke or transient ischemic attack. N Engl J Med. 2016;374(14):1321-1331. doi: 10.1056/NEJMoa1506930 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Mather KJ, Steinberg HO, Baron AD. Insulin resistance in the vasculature. J Clin Invest. 2013;123(3):1003-1004. doi: 10.1172/JCI67166 [DOI] [PMC free article] [PubMed] [Google Scholar]
50.Semenkovich CF. Insulin resistance and atherosclerosis. J Clin Invest. 2006;116(7):1813-1822. doi: 10.1172/JCI29024 [DOI] [PMC free article] [PubMed] [Google Scholar]
51.Smith AG, Singleton JR. Obesity and hyperlipidemia are risk factors for early diabetic neuropathy. J Diabetes Complications. 2013;27(5):436-442. doi: 10.1016/j.jdiacomp.2013.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Kissel JT, Smith AG. Understanding small fiber neuropathy: the long and short of it. JAMA Neurol. 2016;73(6):635-637. doi: 10.1001/jamaneurol.2016.0256 [DOI] [PubMed] [Google Scholar]
53.Verma A, Visintainer P, Elarabi M, Wartak S, Rothberg MB. Overtreatment and undertreatment of hyperlipidemia in the outpatient setting. South Med J. 2012;105(7):329-333. doi: 10.1097/SMJ.0b013e318259bad3 [DOI] [PubMed] [Google Scholar]
54.Mansi IA, Frei CR, Halm EA, Mortensen EM. Association of statins with diabetes mellitus and diabetic complications: role of confounders during follow-up. J Investig Med. 2017;65(1):32-42. doi: 10.1136/jim-2016-000218 [DOI] [PubMed] [Google Scholar]
55.Raebel MA, Schroeder E, Goodrich G, et al. Validating type 1 and type 2 diabetes mellitus in the Mini-Sentinel Distributed Database using the Surveillance, Prevention, and Management of Diabetes Mellitus (SUPREME-DM) DataLink. Accessed on March 4, 2021. https://www.sentinelsystem.org/sentinel/methods/validating-type-1-and-type-2-diabetes-mellitus-mini-sentinel-distributed-database
56.Wong ES, Wang V, Liu CF, Hebert PL, Maciejewski ML. Do Veterans Health Administration enrollees generalize to other populations? Med Care Res Rev. 2016;73(4):493-507. doi: 10.1177/1077558715617382 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement.

eTable 1. Definition of glucose-lowering medication classes

eTable 2. Definition of administrative codes that defined outcomes

eTable 3. Administrative codes used in definitions of baseline characteristics

eTable 4. Laboratory investigation and vital signs handling

eMethods. Propensity score − matching details

eFigure. Study design and cohort assembly

eTable 5. Comparisons of changes in laboratory values from baseline to

follow-up periods

eDiscussion. Summary of some proposed mechanisms for increased insulin

resistance in association of statin use

Click here for additional data file.^{(567KB, pdf)}

[ioi210057r1] 1.Stone NJ, Robinson JG, Lichtenstein AH, et al. ; American College of Cardiology/American Heart Association Task Force on Practice Guidelines . 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63(25 Pt B):2889-2934. doi: 10.1016/j.jacc.2013.11.002 [DOI] [PubMed] [Google Scholar]

[ioi210057r2] 2.American Diabetes Association . Cardiovascular disease and risk management: standards of medical care in diabetes-2018. Diabetes Care. 2018;41(suppl 1):S86-S104. doi: 10.2337/dc18-S009 [DOI] [PubMed] [Google Scholar]

[ioi210057r3] 3.Holman RR, Paul S, Farmer A, Tucker L, Stratton IM, Neil HA; Atorvastatin in Factorial with Omega-3 EE90 Risk Reduction in Diabetes Study Group . Atorvastatin in Factorial with Omega-3 EE90 Risk Reduction in Diabetes (AFORRD): a randomised controlled trial. Diabetologia. 2009;52(1):50-59. doi: 10.1007/s00125-008-1179-5 [DOI] [PubMed] [Google Scholar]

[ioi210057r4] 4.Erlandson KM, Jiang Y, Debanne SM, McComsey GA. Rosuvastatin worsens insulin resistance in HIV-infected adults on antiretroviral therapy. Clin Infect Dis. 2015;61(10):1566-1572. doi: 10.1093/cid/civ554 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r5] 5.Koh KK, Quon MJ, Han SH, Lee Y, Kim SJ, Shin EK. Atorvastatin causes insulin resistance and increases ambient glycemia in hypercholesterolemic patients. J Am Coll Cardiol. 2010;55(12):1209-1216. doi: 10.1016/j.jacc.2009.10.053 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r6] 6.Liew SM, Lee PY, Hanafi NS, et al. Statins use is associated with poorer glycaemic control in a cohort of hypertensive patients with diabetes and without diabetes. Diabetol Metab Syndr. 2014;6:53. doi: 10.1186/1758-5996-6-53 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r7] 7.Sukhija R, Prayaga S, Marashdeh M, et al. Effect of statins on fasting plasma glucose in diabetic and nondiabetic patients. J Investig Med. 2009;57(3):495-499. doi: 10.2310/JIM.0b013e318197ec8b [DOI] [PubMed] [Google Scholar]

[ioi210057r8] 8.Cederberg H, Stančáková A, Yaluri N, Modi S, Kuusisto J, Laakso M. Increased risk of diabetes with statin treatment is associated with impaired insulin sensitivity and insulin secretion: a 6 year follow-up study of the METSIM cohort. Diabetologia. 2015;58(5):1109-1117. doi: 10.1007/s00125-015-3528-5 [DOI] [PubMed] [Google Scholar]

[ioi210057r9] 9.Ridker PM, MacFadyen J, Cressman M, Glynn RJ. Efficacy of rosuvastatin among men and women with moderate chronic kidney disease and elevated high-sensitivity C-reactive protein: a secondary analysis from the JUPITER (Justification for the Use of Statins in Prevention-an Intervention Trial Evaluating Rosuvastatin) trial. J Am Coll Cardiol. 2010;55(12):1266-1273. doi: 10.1016/j.jacc.2010.01.020 [DOI] [PubMed] [Google Scholar]

[ioi210057r10] 10.Kilpatrick ES, Rigby AS, Atkin SL. Insulin resistance, the metabolic syndrome, and complication risk in type 1 diabetes: “double diabetes” in the Diabetes Control and Complications Trial. Diabetes Care. 2007;30(3):707-712. doi: 10.2337/dc06-1982 [DOI] [PubMed] [Google Scholar]

[ioi210057r11] 11.Brownlee M. The pathobiology of diabetic complications: a unifying mechanism. Diabetes. 2005;54(6):1615-1625. doi: 10.2337/diabetes.54.6.1615 [DOI] [PubMed] [Google Scholar]

[ioi210057r12] 12.Maki KC, Ridker PM, Brown WV, Grundy SM, Sattar N; The Diabetes Subpanel of the National Lipid Association Expert Panel . An assessment by the Statin Diabetes Safety Task Force: 2014 update. J Clin Lipidol. 2014;8(3)(suppl):S17-S29. doi: 10.1016/j.jacl.2014.02.012 [DOI] [PubMed] [Google Scholar]

[ioi210057r13] 13.Health Services Research and Development. Accessed October 22, 2020. https://www.hsrd.research.va.gov/for_researchers/vinci/cdw.cfm

[ioi210057r14] 14.Miller DR, Safford MM, Pogach LM. Who has diabetes? Best estimates of diabetes prevalence in the Department of Veterans Affairs based on computerized patient data. Diabetes Care. 2004;27(suppl 2):B10-B21. doi: 10.2337/diacare.27.suppl_2.B10 [DOI] [PubMed] [Google Scholar]

[ioi210057r15] 15.Alvarez CA, Halm EA, Pugh MJV, et al. Lactic acidosis incidence with metformin in patients with type 2 diabetes and chronic kidney disease: A retrospective nested case-control study. Endocrinol Diabetes Metab. 2020;4(1):e00170. doi: 10.1002/edm2.170 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r16] 16.Lund JL, Richardson DB, Stürmer T. The active comparator, new user study design in pharmacoepidemiology: historical foundations and contemporary application. Curr Epidemiol Rep. 2015;2(4):221-228. doi: 10.1007/s40471-015-0053-5 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r17] 17.Yoshida K, Solomon DH, Kim SC. Active-comparator design and new-user design in observational studies. Nat Rev Rheumatol. 2015;11(7):437-441. doi: 10.1038/nrrheum.2015.30 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r18] 18.Danaei G, Rodríguez LA, Cantero OF, Logan R, Hernán MA. Observational data for comparative effectiveness research: an emulation of randomised trials of statins and primary prevention of coronary heart disease. Stat Methods Med Res. 2013;22(1):70-96. doi: 10.1177/0962280211403603 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r19] 19.Jee SH, Sull JW, Park J, et al. Body-mass index and mortality in Korean men and women. N Engl J Med. 2006;355(8):779-787. doi: 10.1056/NEJMoa054017 [DOI] [PubMed] [Google Scholar]

[ioi210057r20] 20.Allison DB, Faith MS, Heo M, Townsend-Butterworth D, Williamson DF. Meta-analysis of the effect of excluding early deaths on the estimated relationship between body mass index and mortality. Obes Res. 1999;7(4):342-354. doi: 10.1002/j.1550-8528.1999.tb00417.x [DOI] [PubMed] [Google Scholar]

[ioi210057r21] 21.Elixhauser A, Steiner C, Palmer L. Clinical Classifications Software (CCS) for ICD-9-CM. Databases and related tools from the Healthcare Cost and Utilization Project (HCUP): appendix a. 2012. Accessed August 30, 2021. https://www.hcup-us.ahrq.gov/toolssoftware/ccs/AppendixASingleDX.txt

[ioi210057r22] 22.Henderson DC, Cagliero E, Copeland PM, et al. Elevated hemoglobin A1c as a possible indicator of diabetes mellitus and diabetic ketoacidosis in schizophrenia patients receiving atypical antipsychotics. J Clin Psychiatry. 2007;68(4):533-541. doi: 10.4088/JCP.v68n0407 [DOI] [PubMed] [Google Scholar]

[ioi210057r23] 23.Wetterhall SF, Olson DR, DeStefano F, et al. Trends in diabetes and diabetic complications, 1980-1987. Diabetes Care. 1992;15(8):960-967. doi: 10.2337/diacare.15.8.960 [DOI] [PubMed] [Google Scholar]

[ioi210057r24] 24.Christakis DA, Feudtner C, Pihoker C, Connell FA. Continuity and quality of care for children with diabetes who are covered by medicaid. Ambul Pediatr. 2001;1(2):99-103. doi: [DOI] [PubMed] [Google Scholar]

[ioi210057r25] 25.Khokhar B, Jette N, Metcalfe A, et al. Systematic review of validated case definitions for diabetes in ICD-9-coded and ICD-10-coded data in adult populations. BMJ Open. 2016;6(8):e009952. doi: 10.1136/bmjopen-2015-009952 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r26] 26.Saydah SH, Geiss LS, Tierney E, Benjamin SM, Engelgau M, Brancati F. Review of the performance of methods to identify diabetes cases among vital statistics, administrative, and survey data. Ann Epidemiol. 2004;14(7):507-516. doi: 10.1016/j.annepidem.2003.09.016 [DOI] [PubMed] [Google Scholar]

[ioi210057r27] 27.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613-619. doi: 10.1016/0895-4356(92)90133-8 [DOI] [PubMed] [Google Scholar]

[ioi210057r28] 28.D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117(6):743-753. doi: 10.1161/CIRCULATIONAHA.107.699579 [DOI] [PubMed] [Google Scholar]

[ioi210057r29] 29.Quan H, Li B, Saunders LD, et al. ; IMECCHI Investigators . Assessing validity of ICD-9-CM and ICD-10 administrative data in recording clinical conditions in a unique dually coded database. Health Serv Res. 2008;43(4):1424-1441. doi: 10.1111/j.1475-6773.2007.00822.x [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r30] 30.Goff SL, Pekow PS, Markenson G, Knee A, Chasan-Taber L, Lindenauer PK. Validity of using ICD-9-CM codes to identify selected categories of obstetric complications, procedures and co-morbidities. Paediatr Perinat Epidemiol. 2012;26(5):421-429. doi: 10.1111/j.1365-3016.2012.01303.x [DOI] [PubMed] [Google Scholar]

[ioi210057r31] 31.Stevens LA, Coresh J, Feldman HI, et al. Evaluation of the modification of diet in renal disease study equation in a large diverse population. J Am Soc Nephrol. 2007;18(10):2749-2757. doi: 10.1681/ASN.2007020199 [DOI] [PubMed] [Google Scholar]

[ioi210057r32] 32.Becker SO, Ichino A. Estimation of average treatment effects based on propensity scores. Stata J. 2002;2(4):358-377. doi: 10.1177/1536867X0200200403 [DOI] [Google Scholar]

[ioi210057r33] 33.Leuven E, Sianesi B. PSMATCH2: Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing, v 4.0.5. 2003. Accessed August 27, 2021. https://ideas.repec.org/c/boc/bocode/s432001.html

[ioi210057r34] 34.Bender R, Blettner M. Calculating the “number needed to be exposed” with adjustment for confounding variables in epidemiological studies. J Clin Epidemiol. 2002;55(5):525-530. doi: 10.1016/S0895-4356(01)00510-8 [DOI] [PubMed] [Google Scholar]

[ioi210057r35] 35.Mansi IA. Statins in primary prevention: uncertainties and gaps in randomized trial data. Am J Cardiovasc Drugs. 2016;16(6):407-418. doi: 10.1007/s40256-016-0190-3 [DOI] [PubMed] [Google Scholar]

[ioi210057r36] 36.US Centers for Disease Control and Prevention . US diabetes surveillance system and diabetes atlas. Assessed December 9, 2020. https://www.cdc.gov/diabetes/data

[ioi210057r37] 37.Gregg EW, Hora I, Benoit SR. Resurgence in diabetes-related complications. JAMA. 2019;321(19):1867-1868. doi: 10.1001/jama.2019.3471 [DOI] [PubMed] [Google Scholar]

[ioi210057r38] 38.Rawshani A, Rawshani A, Franzén S, et al. Risk factors, mortality, and cardiovascular outcomes in patients with type 2 diabetes. N Engl J Med. 2018;379(7):633-644. doi: 10.1056/NEJMoa1800256 [DOI] [PubMed] [Google Scholar]

[ioi210057r39] 39.Rana JS, Liu JY, Moffet HH, Jaffe M, Karter AJ. Diabetes and prior coronary heart disease are not necessarily risk equivalent for future coronary heart disease events. J Gen Intern Med. 2016;31(4):387-393. doi: 10.1007/s11606-015-3556-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r40] 40.Hu FB, Stampfer MJ, Solomon CG, et al. The impact of diabetes mellitus on mortality from all causes and coronary heart disease in women: 20 years of follow-up. Arch Intern Med. 2001;161(14):1717-1723. doi: 10.1001/archinte.161.14.1717 [DOI] [PubMed] [Google Scholar]

[ioi210057r41] 41.Wannamethee SG, Shaper AG, Whincup PH, Lennon L, Sattar N. Impact of diabetes on cardiovascular disease risk and all-cause mortality in older men: influence of age at onset, diabetes duration, and established and novel risk factors. Arch Intern Med. 2011;171(5):404-410. doi: 10.1001/archinternmed.2011.2 [DOI] [PubMed] [Google Scholar]

[ioi210057r42] 42.Idris I. Diabetes and cardiovascular risk equivalency: do age at diagnosis and disease duration affect risk stratification?: comment on “impact of diabetes on cardiovascular disease risk and all-cause mortality in older men”. Arch Intern Med. 2011;171(5):410-411. doi: 10.1001/archinternmed.2010.524 [DOI] [PubMed] [Google Scholar]

[ioi210057r43] 43.Colhoun HM, Betteridge DJ, Durrington PN, et al. ; CARDS investigators . Primary prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): multicentre randomised placebo-controlled trial. Lancet. 2004;364(9435):685-696. doi: 10.1016/S0140-6736(04)16895-5 [DOI] [PubMed] [Google Scholar]

[ioi210057r44] 44.Swerdlow DI, Preiss D, Kuchenbaecker KB, et al. ; DIAGRAM Consortium; MAGIC Consortium; InterAct Consortium . HMG-coenzyme A reductase inhibition, type 2 diabetes, and bodyweight: evidence from genetic analysis and randomised trials. Lancet. 2015;385(9965):351-361. doi: 10.1016/S0140-6736(14)61183-1 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r45] 45.Henriksbo BD, Lau TC, Cavallari JF, et al. Fluvastatin causes NLRP3 inflammasome-mediated adipose insulin resistance. Diabetes. 2014;63(11):3742-3747. doi: 10.2337/db13-1398 [DOI] [PubMed] [Google Scholar]

[ioi210057r46] 46.Mitchell P, Marette A. Statin-induced insulin resistance through inflammasome activation: sailing between Scylla and Charybdis. Diabetes. 2014;63(11):3569-3571. doi: 10.2337/db14-1059 [DOI] [PubMed] [Google Scholar]

[ioi210057r47] 47.Buse MG. Hexosamines, insulin resistance, and the complications of diabetes: current status. Am J Physiol Endocrinol Metab. 2006;290(1):E1-E8. doi: 10.1152/ajpendo.00329.2005 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r48] 48.Kernan WN, Viscoli CM, Furie KL, et al. ; IRIS Trial Investigators . Pioglitazone after ischemic stroke or transient ischemic attack. N Engl J Med. 2016;374(14):1321-1331. doi: 10.1056/NEJMoa1506930 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r49] 49.Mather KJ, Steinberg HO, Baron AD. Insulin resistance in the vasculature. J Clin Invest. 2013;123(3):1003-1004. doi: 10.1172/JCI67166 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r50] 50.Semenkovich CF. Insulin resistance and atherosclerosis. J Clin Invest. 2006;116(7):1813-1822. doi: 10.1172/JCI29024 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r51] 51.Smith AG, Singleton JR. Obesity and hyperlipidemia are risk factors for early diabetic neuropathy. J Diabetes Complications. 2013;27(5):436-442. doi: 10.1016/j.jdiacomp.2013.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]

[ioi210057r52] 52.Kissel JT, Smith AG. Understanding small fiber neuropathy: the long and short of it. JAMA Neurol. 2016;73(6):635-637. doi: 10.1001/jamaneurol.2016.0256 [DOI] [PubMed] [Google Scholar]

[ioi210057r53] 53.Verma A, Visintainer P, Elarabi M, Wartak S, Rothberg MB. Overtreatment and undertreatment of hyperlipidemia in the outpatient setting. South Med J. 2012;105(7):329-333. doi: 10.1097/SMJ.0b013e318259bad3 [DOI] [PubMed] [Google Scholar]

[ioi210057r54] 54.Mansi IA, Frei CR, Halm EA, Mortensen EM. Association of statins with diabetes mellitus and diabetic complications: role of confounders during follow-up. J Investig Med. 2017;65(1):32-42. doi: 10.1136/jim-2016-000218 [DOI] [PubMed] [Google Scholar]

[ioi210057r55] 55.Raebel MA, Schroeder E, Goodrich G, et al. Validating type 1 and type 2 diabetes mellitus in the Mini-Sentinel Distributed Database using the Surveillance, Prevention, and Management of Diabetes Mellitus (SUPREME-DM) DataLink. Accessed on March 4, 2021. https://www.sentinelsystem.org/sentinel/methods/validating-type-1-and-type-2-diabetes-mellitus-mini-sentinel-distributed-database

[ioi210057r56] 56.Wong ES, Wang V, Liu CF, Hebert PL, Maciejewski ML. Do Veterans Health Administration enrollees generalize to other populations? Med Care Res Rev. 2016;73(4):493-507. doi: 10.1177/1077558715617382 [DOI] [PubMed] [Google Scholar]

PERMALINK

Association of Statin Therapy Initiation With Diabetes Progression

Ishak A Mansi, MD

Matthieu Chansard

Ildiko Lingvay, MD, MPH, MSCS

Song Zhang, PhD

Ethan A Halm, MD, MPH, MBA

Carlos A Alvarez, PharmD, MSc, MSCS

Key Points

Question

Findings

Meaning

Abstract

Importance

Objective

Design, Setting, and Participants

Interventions

Main Outcomes and Measures

Results

Conclusions and Relevance

Introduction

Methods

Study Design

Study Groups

Study Intervals

Primary Outcome—Diabetes Progression

Secondary Outcomes

Cohort Characterization

Table 1. Baseline Characteristics of Propensity Score–Matched Statin Users and Active Comparators in the Overall and Diabetes-Prevalent Cohorts.

Primary and Secondary Analyses

Sensitivity Analysis

Post Hoc Analysis

Statistical Analysis

Results

Primary Analysis

Table 2. Odds of Outcomes During Follow-up Period Between Statin Users and Active Comparators in Propensity Score–Matched Cohorts.

Secondary Analysis

Table 3. Secondary Analysis and Sensitivity Analysis Comparing the Diabetes Progression Composite Outcome During Follow-up Between Statin Users and Active Comparators.

Post Hoc Analysis

Discussion

Strengths and Limitations

Conclusions

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases