Measures are given per 1-SD increment. Cox proportional hazards ratios (HR) per 1-SD increment were adjusted for age, sex, race or ethnic origin, educational level, exercise, family history of T2DM, and smoking (model 1). HDL indicates high-density lipoprotein; IDL, intermediate-density lipoprotein; LDL, low-density lipoprotein; LPIR, lipoprotein insulin resistance; and VLDL, very low-density lipoprotein.
aScores range from 0 (most insulin sensitive) to 100 (most insulin resistant).
eMethods 1. Calculation of Lipoprotein Insulin Resistance (LPIR) Score
eMethods 2. Discrimination and Reclassification of Incident Diabetes With LPIR Score
eTable 1. Baseline Characteristics of Current Study Participants
eTable 2. Baseline Lipoprotein Measures of Current Study Participants According to LPIR Score
eTable 3. Baseline Characteristics of Current Study Participants According to Incident Diabetes
eTable 4. Baseline Lipoprotein Measures of Current Study Participants According to Incident Diabetes
eTable 5. Lipid and Lipoprotein Measures at Baseline and at 12 Months, According to Randomized Treatment Arm
eTable 6. Spearman Correlation Coefficients of LPIR Score With Lipoproteins
eTable 7. Baseline Lipid and Lipoprotein Measures in Relation to Incident Diabetes According to Randomized Treatment Arm
eTable 8. 12-Month Lipoprotein Measures in Relation to Incident Diabetes According to Randomized Treatment Arm
eTable 9. Baseline Lipoprotein Measures in Relation to Incident Diabetes in Placebo-Allocated Individuals
eTable 10. Baseline Lipoprotein Measures in Relation to Incident Diabetes in Rosuvastatin-Allocated Individuals
eTable 11. 12-Month Lipoprotein Measures in Relation to Incident Diabetes in Rosuvastatin-Allocated Individuals
eTable 12. Likelihood Ratio Tests, Discrimination, and Reclassification of Incident Diabetes With LPIR Score
Dugani SB, Akinkuolie AO, Paynter N, Glynn RJ, Ridker PM, Mora S. Association of Lipoproteins, Insulin Resistance, and Rosuvastatin With Incident Type 2 Diabetes Mellitus Secondary Analysis of a Randomized Clinical Trial. JAMA Cardiol. 2016;1(2):136-145. doi:10.1001/jamacardio.2016.0096
Statins decrease levels of low-density lipoprotein (LDL) and triglycerides as well as cardiovascular events but increase the risk for a diagnosis of type 2 diabetes mellitus (T2DM). The risk factors associated with incident T2DM are incompletely characterized.
To investigate the association of lipoprotein subclasses and size and a novel lipoprotein insulin resistance (LPIR) score (a composite of 6 lipoprotein measures) with incident T2DM among individuals randomized to a high-intensity statin or placebo.
Design, Setting, and Participants
This secondary analysis of the JUPITER trial (a placebo-controlled randomized clinical trial) was conducted at 1315 sites in 26 countries and enrolled 17 802 men 50 years or older and women 60 years or older with LDL cholesterol levels less than 130 mg/dL, high-sensitivity C-reactive protein levels of at least 2 mg/L, and triglyceride levels less than 500 mg/dL. Those with T2DM were excluded. A prespecified secondary aim was to assess the effect of rosuvastatin calcium on T2DM. Incident T2DM was monitored for a median of 2.0 years. Data were collected from February 4, 2003, to August 20, 2008, and analyzed (intention-to-treat) from December 1, 2013, to January 21, 2016.
Rosuvastatin calcium, 20 mg/d, or placebo.
Main Outcomes and Measures
Size and concentration of lipids, apolipoproteins, and lipoproteins at baseline (11 918 patients with evaluable plasma samples) and 12 months after randomization (9180 patients). The LPIR score, a correlate of insulin resistance, was calculated as a weighted combination of size and concentrations of LDL, very low-density lipoprotein (VLDL), and high-density lipoprotein (HDL) particles.
Among the 11 918 patients (4334 women [36.4%]; median [interquartile range] age, 66 [60-71] years), rosuvastatin lowered the levels of LDL particles (−39.6%; 95% CI, –49.4% to –24.6%), VLDL particles (−19.6%; 95% CI, −40.6% to 10.3%), and VLDL triglycerides (−15.2%; 95% CI, −35.9% to 11.3%) and shifted the lipoprotein subclass distribution toward smaller LDL size (−1.5%; 95% CI, −3.7% to 0.5%), larger VLDL size (2.8%; 95% CI, −5.8% to 12.7%), and lower LPIR score (−3.2%; 95% CI, −20.6% to 16.9%). In analyses adjusted for age, sex, race or ethnic origin, exercise, educational level, family history, and smoking, the hazard ratio (HR) for T2DM per SD of LPIR score in the placebo arm was 1.99 (95% CI, 1.64-2.42); in the rosuvastatin arm, 2.06 (95% CI, 1.74-2.43). After additional adjustment for systolic blood pressure, body mass index, high-sensitivity C-reactive protein, hemoglobin A1c, HDL cholesterol, LDL cholesterol, and triglycerides, the LPIR score remained associated with T2DM in the placebo arm (HR, 1.35; 95% CI, 1.03-1.76) and rosuvastatin arm (HR, 1.60; 95% CI, 1.27-2.03). Similar trends were seen at 12 months. The LPIR score improved the model likelihood ratio (χ2 = 18.23; P < .001) and categorical net reclassification index (0.039; 95% CI, 0.003-0.072).
Conclusions and Relevance
In apparently healthy people, LPIR score was positively associated with incident T2DM, including during rosuvastatin therapy.
clinicaltrials.gov Identifier: NCT00239681
Statins substantially reduce cardiovascular events1- 3 but are associated with an increased risk for being diagnosed with type 2 diabetes mellitus (T2DM).2- 7 Statin users who develop T2DM often have evidence of prior impaired fasting glucose levels, features of insulin resistance, or metabolic syndrome,8,9 factors that also predispose to the development of T2DM in statin-naive individuals.10 Identifying statin users at risk for T2DM has gained greater significance because recent guidelines to reduce cholesterol levels11 could increase the global prescription of statins.
Insulin resistance and T2DM are associated with lipoprotein profile changes12- 16 that precede the appearance of overt hyperglycemia. Lipoprotein particles are categorized according to density into low-density lipoproteins (LDL), high-density lipoproteins (HDL), and very low-density lipoproteins (VLDL), and these are further categorized on the basis of particle size and concentration (or number). Nonrandomized observational studies focusing predominantly on statin-naive populations have reported positive associations of T2DM with higher particle concentrations of small LDL, small HDL, and large VLDL particles and inverse associations of T2DM with large LDL and large HDL particles,12- 18 underscoring the complex and incompletely characterized association of lipoproteins with insulin resistance and T2DM. To date, no studies have examined the various lipoprotein characteristics that precede the onset of T2DM among individuals randomly allocated to statin therapy vs placebo.
To address these issues, we used nuclear magnetic resonance (NMR) spectroscopy, immunoassay-measured apolipoprotein values, and standard lipid measurements to characterize the lipoprotein profiles comprehensively at baseline and 12 months after randomization to rosuvastatin calcium, 20 mg/d, or placebo in the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER). JUPITER is a primary prevention trial of individuals without prior cardiovascular disease or T2DM but with elevated high-sensitivity C-reactive protein (hsCRP) and low LDL cholesterol levels who were followed up prospectively for incident cardiovascular events.2 A prespecified secondary aim of JUPITER was to assess the effect of rosuvastatin on incident T2DM.9
After the trial was completed but before obtaining NMR measurements, we prespecified the hypothesis that a lipoprotein insulin resistance (LPIR) score, which reflects lipoprotein derangements of insulin resistance, would be associated with incident T2DM in placebo- and rosuvastatin-allocated individuals. The LPIR score combines 6 measures of LDL, VLDL, and HDL particle size and concentration and incorporates lipoprotein characteristics that previously have been individually associated with T2DM and/or insulin resistance.12- 16,18 The LPIR score is more strongly correlated with T2DM19 and insulin resistance (measured by the homeostasis model assessment of insulin resistance [HOMA-IR]) than each of its 6 subclasses individually and has been proposed to better reflect the complex biology and regulation of lipoproteins.20 Herein, we describe the prospective association of individual lipoprotein measures and the LPIR score with incident T2DM according to randomized treatment allocation.
Question What risk factors are associated with developing type 2 diabetes mellitus for individuals receiving statin therapy?
Findings Among 11 918 participants in the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) clinical trial, in which apparently healthy adults were randomized to placebo or rosuvastatin, lipoprotein insulin resistance, a novel composite of 6 lipoprotein measures, was significantly associated with incident type 2 diabetes mellitus, including during rosuvastatin therapy.
Meaning Lipoprotein insulin resistance may identify apparently healthy adults at risk of developing diabetes, including during statin therapy.
JUPITER was a double-blind, placebo-controlled, randomized clinical trial conducted at 1315 sites in 26 countries.2,21,22 The effect of rosuvastatin on incident T2DM was a prespecified secondary aim of JUPITER. We used a randomized study design to elucidate possible unique associations between baseline lipoproteins and incident T2DM before and after randomization to rosuvastatin vs placebo. The study protocol was approved by the institutional review boards at Brigham and Women’s Hospital and at the participating centers. All patients provided written informed consent.
JUPITER was a primary prevention trial of 17 802 apparently healthy men and women who were eligible to participate if they had LDL cholesterol levels of less than 130 mg/dL (to convert to millimoles per liter, multiply by 0.0259), hsCRP levels of at least 2 mg/L (to convert to nanomoles per liter, multiply by 9.524) and triglyceride levels of less than 500 mg/dL (to convert to millimoles per liter, multiply by 0.0113). Because incident T2DM was a prespecified secondary aim of JUPITER, exclusion criteria of the trial included preexisting T2DM, defined as fasting glucose level of at least 126 mg/dL (to convert to millimoles per liter, multiply by 0.0555) at screening visit 2 or by the use of insulin and/or an oral hypoglycemic. Other trial requirements included a willingness to participate for the duration of the study and the ability to provide written consent.2 The placebo and rosuvastatin arms each included 8901 participants who were requested to provide a blood sample at baseline and 12 months after randomization; a total of 11 918 samples had plasma available to obtain complete NMR lipoprotein measurements at baseline, and of these, 9180 samples had plasma available for 12-month measurements. Incident T2DM cases were tracked throughout the study period (see below) and were physician reported, as described.2,9
Data were collected from February 4, 2003, to August 20, 2008. We measured plasma levels of lipids, lipoproteins, apolipoproteins, hsCRP, and hemoglobin A1c as described.2,23,24 We calculated LDL cholesterol level using the Friedewald equation (for a serum triglyceride level of <400 mg/dL) or ultracentrifugation (for a serum triglyceride level of ≥400 mg/dL).24,25 Lipoprotein particle concentration and mean size of LDL, HDL, and VLDL particles were determined by NMR spectroscopy at LipoScience (now LabCorp).13,26 The LPIR score is a weighted combination of 6 lipoprotein subclass measures and reflects the concentrations of large VLDL, large HDL, and small LDL particles and mean sizes of VLDL, LDL, and HDL particles. The LPIR score ranges from 0 (most insulin sensitive) to 100 (most insulin resistant).
The LPIR score was developed as described previously20 (eMethods 1 in the Supplement). Briefly, using HOMA-IR measurements for guidance, the 6 NMR-measured lipoprotein variables known to be associated with insulin resistance were combined to produce a multiplex LPIR score that ranges from 0 (most insulin sensitive) to 100 (most insulin resistant). The algorithm used to generate LPIR divides the 6 lipoprotein variables into several particle concentration or size categories, assigns each a numerical weighting score, and sums these to produce the LPIR score. The weighting scores were chosen empirically to reflect the strength and independence of each variable’s association with HOMA-IR in the Multi-Ethnic Study of Atherosclerosis population.20 Accordingly, mean VLDL particle size and concentration of large VLDL particles were assigned the greatest weighting scores (32 and 22, respectively), followed by mean HDL particle size (20), concentration of large HDL particles (12), concentration of small LDL particles (8), and mean LDL particle size (6). This combination of 6 lipoprotein variables was more strongly related to insulin resistance than any of the individual variables or the ratio of triglycerides to HDL cholesterol.
Data were analyzed from December 1, 2013, to January 21, 2016. Analyses were performed using SAS software (version 9.3; SAS Institute Inc). Spearman coefficients were used to quantify correlations. Incidence rates of T2DM were calculated per 100 person-years, and exposure time was calculated as the time from randomization to occurrence of the end point or to a participant’s last blinded follow-up visit, a process that concluded August 20, 2008, 6 months after the primary JUPITER trial was ended by the data and safety monitoring board.9 Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% CIs to compare the risk for T2DM according to tertiles and per 1 SD of lipoprotein measures using separate models for each measure. To allow for comparison across groups, HRs were calculated using the SD of baseline levels among all participants. Cox proportional hazards regression models were adjusted for age, sex, self-reported race or ethnic origin, educational level, exercise, family history of T2DM, and self-reported smoking in the prior month (model 1). Race or ethnic origin was assessed to explain possible heterogeneity in the risk for developing T2DM. To account for lipoprotein correlations with each other and with metabolic variables, model 1 was further adjusted for systolic blood pressure, body mass index (BMI), and levels of hemoglobin A1c, LDL cholesterol, HDL cholesterol, log-transformed triglycerides, and log-transformed hsCRP (model 2). Metabolic syndrome was defined according to the consensus criteria of the American Heart Association and the National Heart, Lung, and Blood Institute.2 The probability value for linear trend was obtained by using the median value for each tertile. Statistical tests for interaction between categorical LPIR score and treatment allocation in relation to outcomes were obtained by use of likelihood ratio tests. Statistical significance was established at 2-tailed P < .05. The contribution of the LPIR score to the prediction of T2DM in models with traditional risk factors was evaluated by likelihood ratio χ2 tests, discrimination (Harrell’s C statistic),27 relative integrated discrimination improvement index,28 and net reclassification index29 (eMethods 2 in the Supplement).
Among the 11 918 patients included in the analysis (7584 men [63.6%]; 4334 women [36.4%]; and median [interquartile range] age, 66 [60-71] years), a total of 370 individuals (3.1%) were diagnosed with T2DM during a median follow-up of 2.0 (interquartile range, 1.6-2.5) years. Among these individuals, 158 and 212 cases were randomized to the placebo and rosuvastatin arms, respectively. The incidence rate was higher among individuals allocated to rosuvastatin (1.64 per 100 person-years; 95% CI, 1.43-1.87) vs placebo (1.17 per 100 person-years; 95% CI, 1.00-1.36). Compared with the overall study population and with those excluded from this study, the present study population had a higher proportion of white individuals, whereas other characteristics were generally similar (eTable 1 in the Supplement). Increasing tertiles of LPIR score (ie, higher insulin resistance) were associated with a higher prevalence of clinical risk factors, including metabolic syndrome, higher triglyceride levels, hypertension, and lower HDL cholesterol levels, although LDL cholesterol levels were similar (Table 1 and eTable 2 in the Supplement). Compared with those who did not develop T2DM, individuals who developed T2DM had a higher prevalence of clinical risk factors, including impaired fasting glucose levels, elevated hemoglobin A1c levels, family history of T2DM, metabolic syndrome, and BMI; higher concentrations of LDL particles, VLDL particles, and triglycerides; and higher LPIR scores (eTables 3 and 4 in the Supplement).
Rosuvastatin lowered levels of LDL cholesterol (−49.0%; 95% CI, –58.2% to –32.7%), triglycerides (−15.4%; 95% CI, −32.8% to 5.3%), non-HDL cholesterol (−42.7%; 95% CI, −51.0% to −27.6%), and apolipoprotein B (−39.4%; 95% CI, −47.6% to −26.7%) and the LPIR score (−3.2%; 95% CI, −20.6% to 16.9%) and raised HDL cholesterol levels (6.0%; −3.0% to 17.0%) (eTable 5 in the Supplement). In addition, rosuvastatin lowered the concentrations of LDL and VLDL particles and increased the concentration of HDL particles; these effects differed by particle subclasses, resulting in overall smaller mean sizes for LDL particles and larger sizes for VLDL and HDL particles (P < .001 for all comparisons) (eTable 5 in the Supplement). The Spearman correlation coefficients of the LPIR score with lipoprotein particle subclasses were generally similar at baseline and after 12 months of rosuvastatin therapy (eTable 6 in the Supplement).
Among placebo- and rosuvastatin-allocated individuals, incident T2DM was inversely associated with baseline concentrations of LDL and HDL cholesterol and positively associated with concentrations of triglycerides and apolipoprotein B and LPIR score (Figure, A and eTable 7 in the Supplement). In analyses adjusted for age, sex, race or ethnic origin, educational level, exercise, family history of T2DM, and smoking (model 1), the HR per 1 SD of LPIR score was 1.99 (95% CI, 1.64-2.42; P < .001) in placebo-allocated individuals and 2.06 (95% CI, 1.74-2.43; P < .001) in rosuvastatin-allocated individuals (Figure, A and eTable 7 in the Supplement). Substituting fasting glucose level for hemoglobin A1c level somewhat attenuated the association between the LPIR score and T2DM in individuals allocated to placebo (HR, 1.55; 95% CI, 1.27-1.90; P < .001) and rosuvastatin (HR, 1.63; 95% CI, 1.38-1.93; P < .001). In addition, adjusting for systolic blood pressure, BMI, and levels of hsCRP, hemoglobin A1c, HDL and LDL cholesterol, and triglycerides (model 2) attenuated the HRs among individuals allocated to placebo (1.35; 95% CI, 1.03-1.76; P = .03) and rosuvastatin (1.60; 95% CI, 1.27-2.03; P < .001). Similar trends were obtained when assessed across increasing tertiles of the LPIR score (Table 2). Tests of interaction between randomized treatment and LPIR tertiles in relation to T2DM in models 1 and 2 were P = .99 and P = .79, respectively.
Overall (model 1), placebo- and rosuvastatin-allocated individuals showed generally similar associations for baseline lipoprotein subclass characteristics and incident T2DM (Figure, A and eTable 7 in the Supplement). However, after additionally adjusting for blood pressure, BMI, and levels of hsCRP, hemoglobin A1c, and lipids (model 2), we found some notable differences. In particular, in rosuvastatin-allocated individuals, baseline (but not 12-month) small LDL particles, large VLDL particles, and medium VLDL particles were positively associated and small VLDL particles were inversely associated with incident T2DM (eTable 8 in the Supplement). These associations were not seen in placebo-allocated individuals (eTables 9 and 10 in the Supplement).
Similar analyses were performed at 12 months, and model 1 associations seen at baseline were generally preserved (Figure, B). The HR per 1 SD of LPIR score was 2.03 (95% CI, 1.65-2.49; P < .001) in placebo-allocated individuals and 2.03 (95% CI, 1.68-2.45; P < .001) in rosuvastatin-allocated individuals (Figure, B). Substituting fasting glucose level for hemoglobin A1c level attenuated the association of LPIR score in placebo-allocated (HR, 1.62; 95% CI, 1.32-1.99; P < .001) and rosuvastatin-allocated (HR, 1. 63; 95% CI, 1.35-1.98; P < .001) individuals. Additional adjustment for model 2 variables also slightly attenuated the association between LPIR score with T2DM in individuals allocated to placebo (HR, 1.51; 95% CI, 1.15-1.96; P = .003) and rosuvastatin (HR, 1.58; 95% CI, 1.23-2.01; P < .001). Similar trends were obtained across increasing tertiles of LPIR score (Table 3 and eTable 11 in the Supplement). The tests of interaction between randomized treatment and LPIR tertiles in relation to T2DM in models 1 and 2 were P = .98 and P = .86, respectively.
In addition to the LPIR score, baseline and 12-month levels of apolipoprotein B were positively associated with T2DM. In mutually adjusted models, the LPIR score remained associated with T2DM, whereas the association between apolipoprotein B and T2DM was completely abrogated (Tables 2 and 3). Likewise, the LPIR score and the ratio of triglycerides to HDL cholesterol were significantly associated with T2DM (Tables 2 and 3), but in mutually adjusted models, the LPIR score remained significant in both arms, whereas the ratio of triglycerides to HDL cholesterol was no longer significant in the rosuvastatin arm.
To investigate whether the LPIR score would improve risk prediction metrics, we used model 2 and a modified model (model 2A) that included fasting glucose level instead of hemoglobin A1c level (eTable 12 in the Supplement). The likelihood ratio χ2 tests were significantly improved when the LPIR score was added to model 2 (χ2 = 18.23; P < .001) and model 2A (χ2 = 12.26; P < .001). The C statistic of model 2 (0.827; 95% CI, 0.804-0.851) was unchanged after adding the LPIR score (0.827; 95% CI, 0.804-0.851). Similarly, the C statistic of model 2A (0.853; 95% CI, 0.832-0.873) was unchanged after adding the LPIR score (0.855; 95% CI, 0.835-0.875). Further, the relative integrated discrimination improvement index did not improve significantly after addition of the LPIR score to model 2 (0.038; 95% CI, −0.0009 to 0.076) and model 2A (0.024; 95% CI, −0.003 to 0.050). However, the categorical net reclassification index was improved after adding the LPIR score to model 2 (0.039; 95% CI, 0.003-0.072), driven mainly by reclassification of nonevents (0.036) and events (0.002). The categorical net reclassification index of model 2A after adding the LPIR score was 0.012 (95% CI, −0.008 to 0.03). Although the improvement in model 2 net reclassification index was statistically significant, the gain in the model’s predictive ability was modest.
Results from this study support several conclusions. First, rosuvastatin was associated with differential effects on the size and concentration of LDL, HDL, and VLDL particles. Second, rosuvastatin substantially reduced LDL cholesterol and triglyceride levels but only slightly reduced the LPIR score, shifting the LDL and VLDL lipoprotein subclass distribution toward a smaller mean size for LDL particles and a larger mean size for VLDL particles. Third, the LPIR score, a lipoprotein correlate of insulin resistance, was positively associated with incident T2DM in placebo- and rosuvastatin-allocated individuals. Finally, the LPIR score was significantly associated with incident T2DM after adjusting for traditional risk factors, including family history of T2DM, exercise, BMI, and levels of hemoglobin A1c, HDL cholesterol, non-HDL cholesterol, triglycerides, and apolipoprotein B. This study, nested within a placebo-controlled randomized clinical trial, is to our knowledge the first to characterize the association between the LPIR score and T2DM among individuals randomly allocated to placebo or high-intensity statin therapy, and our results suggest that the LPIR score may identify additional individuals at risk for developing T2DM, including when taking rosuvastatin.
Recently, a similar positive association between LPIR score and incident T2DM was described in a multiethnic cohort19 in which a small proportion of participants were using medications to lower cholesterol levels. Statin users at risk for developing T2DM often have preexisting impairments in fasting glucose levels and features of the metabolic syndrome,8,9 factors that also predispose to T2DM in statin-naive individuals.10 To identify other risk factors that might predispose to the development of T2DM, we focused on lipoprotein characteristics, given reports of differing associations between T2DM and the size and concentration of LDL, HDL, and VLDL particles.12- 16,18 In the present study, the LPIR score remained associated with T2DM after adjusting for clinical risk factors that have been shown to be independent predictors of developing T2DM during statin treatment.8 Although rosuvastatin substantially reduced LDL cholesterol and triglyceride levels, it only slightly reduced the LPIR score (by 3.2%; eTable 5 in the Supplement) because rosuvastatin altered the LDL and VLDL lipoprotein subclass distribution, resulting in smaller mean sizes for LDL particles and larger mean sizes for VLDL particles.
Although statins are associated with a higher incidence of T2DM, the underlying mechanisms are not well understood30 and may involve inhibition of on-target 3-hydroxy-3-methylglutaryl-CoA reductase31 and its effects on glucose tolerance, insulin sensitivity, and insulin secretion32 and circulating levels of adipocytokines.33 Animal models have shown that statins alter the expression levels of glucose transporter 434 through isoprenoid synthesis,35 the main insulin-responsive glucose transporter that facilitates glucose uptake in muscle and adipose tissue. Conceivably, lipoproteins that constitute the LPIR score could be involved in these and/or additional pathways.35
Our study has several strengths, including that it was nested within a prospective randomized clinical trial and its large sample size and number of events, excellent follow-up, robust information on risk factors that could modify or confound our interpretation, and prespecification of incident T2DM as a secondary end point of interest. Our study also has potential limitations. JUPITER was stopped early by the independent data and safety monitoring board after a median follow-up of 1.9 years for cardiovascular and mortality efficacy, and the long-term effect of statins on incident T2DM could not be determined. JUPITER included individuals with elevated levels of hsCRP (which is also associated with incident T2DM) and excluded those with serum triglyceride levels of 500 mg/dL or greater. This study evaluated a fixed dose of one statin (rosuvastatin, 20 mg/d), and the association of the LPIR score and T2DM with other statins or among individuals who do not meet the study’s inclusion or exclusion criteria requires further evaluation. Future studies should also examine whether modifying or adding additional metabolic predictors (eg, branched-chain amino acids) to the LPIR score could result in improved T2DM risk prediction.
Among placebo- and rosuvastatin-allocated individuals in JUPITER, the LPIR score is positively associated with incident T2DM after adjusting for traditional risk factors. The LPIR score has the potential to serve as part of a broader clinical approach to identify additional cases at risk for T2DM.
Corresponding Author: Samia Mora, MD, MHS, Center for Lipid Metabolomics, Division of Preventive Medicine, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, 900 Commonwealth Ave E, 3rd Floor, Boston, MA 02215 (firstname.lastname@example.org).
Accepted for Publication: January 25, 2016.
Published Online: April 13, 2016. doi:10.1001/jamacardio.2016.0096.
Author Contributions: Dr Mora had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Dugani, Paynter, Glynn, Mora.
Acquisition, analysis, or interpretation of data: Dugani, Akinkuolie, Glynn, Ridker, Mora.
Drafting of the manuscript: Dugani.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Dugani, Akinkuolie, Paynter, Glynn.
Obtained funding: Ridker, Mora.
Administrative, technical, or material support: Ridker, Mora.
Study supervision: Ridker, Mora.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Glynn reports receiving grants from AstraZeneca during the conduct of the study. Dr Ridker reports receiving grants from AstraZeneca and nonfinancial support from LipoScience (now LabCorp) during the conduct of the study; receiving grants from AstraZeneca, Novartis, and Pfizer and nonfinancial support from Amgen outside the submitted work; and having a patent, “Use of inflammatory biomarkers in cardiovascular disease,” issued to AstraZeneca and Siemens. Dr Mora reports receiving grants from AstraZeneca and nonfinancial support from LipoScience during the conduct of the study as well as grants from Atherotech Diagnostics and personal fees from Genzyme, Quest Diagnostics, Eli Lilly, Pfizer, and Cerenis Therapeutics outside the submitted work. No other disclosures were reported.
Funding/Support: This study was supported by grant R01HL117861 from the National Heart, Lung, and Blood Institute of the National Institutes of Health (Dr Mora) and grant T32 HL007575 from the National Heart, Lung, and Blood Institute (Dr Akinkuolie). LipoScience Inc supplied the nuclear magnetic resonance information at no additional charge.
Role of the Funder/Sponsor: The Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) trial was financially supported by AstraZeneca, who collected trial data and monitored sites but had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Previous Presentation: This study was presented in part as an abstract at the American Heart Association Meeting; November 17, 2014; Chicago, Illinois.
Additional Contributions: Chunying Li, MD, PhD, Jean MacFadyen, BA, M. V. Moorthy, PhD, Latha Padmanabhan, MS, and Lynda Rose, MA, Division of Preventive Medicine, Brigham and Women’s Hospital, provided guidance and technical assistance during this study. None of these contributors received additional compensation for this role.