CARDIOVASCULAR DISEASE AND DIABETES Diabetes Care Volume 39, May 2016 709 Cardiovascular and Other ORIGIN Trial Investigators* Outcomes Postintervention With Insulin Glargine and Omega-3 Fatty Acids (ORIGINALE) Diabetes Care 2016;39:709–716 | DOI: 10.2337/dc15-1676 OBJECTIVE The Outcome Reduction With Initial Glargine Intervention (ORIGIN) trial reported neutral effects of insulin glargine on cardiovascular outcomes and cancers and reduced incident diabetes in high–cardiovascular risk adults with dysglycemia after 6.2 years of active treatment. Omega-3 fatty acids had neutral effects on cardiovascular outcomes. The ORIGIN and Legacy Effects (ORIGINALE) study mea- sured posttrial effects of these interventions during an additional 2.7 years. RESEARCH DESIGN AND METHODS Surviving ORIGIN participants attended up to two additional visits. The hazard of clinical outcomes during the entire follow-up period from randomization was calculated. RESULTS Of 12,537 participants randomized, posttrial data were analyzed for 4,718 originally allocated to insulin glargine (2,351) versus standard care (2,367), and 4,771 originally allocated to omega-3 fatty acid supplements (2,368) versus placebo (2,403). Posttrial, small differences in median HbA1c persisted (glargine 6.6% [49 mmol/mol], standard care 6.7% [50 mmol/mol], P = 0.025). From randomization to the end of posttrial follow-up, no differences were found between the glargine and standard care groups in myocardial infarction, stroke, or cardiovascular death (1,185 vs. 1,165 Corresponding author: Zubin Punthakee, zubin. events; hazard ratio 1.01 [95% CI 0.94–1.10]; P = 0.72); myocardial infarction, stroke, [email protected]. cardiovascular death, revascularization, or hospitalization for heart failure (1,958 vs. Received 30 July 2015 and accepted 9 November 2015. 1,910 events; 1.03 [0.97–1.10]; P = 0.38); or any cancer (524 vs. 529 events; 0.99 – P Clinical trial reg. no. NCT00069784, clinicaltrials [0.88 1.12]; = 0.91) or between omega-3 and placebo groups in cardiovascular .gov. – P death (688 vs. 700; 0.98 [0.88 1.09]; = 0.68) or other outcomes. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ CONCLUSIONS suppl/doi:10.2337/dc15-1676/-/DC1. During >6 years of treatment followed by >2.5 years of observation, insulin glar- *The ORIGIN Trial Investigators are listed in the gine had neutral effects on health outcomes and salutary effects on metabolic Supplementary Data, and the Writing Group control, whereas omega-3 fatty acid supplementation had no effect. members are listed in the APPENDIX. © 2016 by the American Diabetes Association. Readers may use this article as long as the work is Type 2 diabetes is a strong, independent risk factor for myocardial infarction, stroke, properly cited, the use is educational and not for premature death, a variety of cancers, and other serious health outcomes. Possible profit, and the work is not altered. explanations for this are prolonged exposure to elevated glucose levels, insulin See accompanying articles, pp. 664, resistance, endogenous hyperinsulinemia, and obesity. Observational studies 668, 677, 686, 694, 701, 717, 726, 735, have also suggested that exogenous insulin use may provoke these outcomes. and 738. 710 ORIGIN and Legacy Effects Diabetes Care Volume 39, May 2016 However, the large, prospective Out- diabetes at the end of the trial. Funding glucose tolerance tests or other criteria come Reduction With Initial Glargine In- was available to continue follow-up until (1). A broader definition included pos- tervention (ORIGIN) trial showed that May 2014. All research sites were invited sible but unconfirmed diabetes (i.e., the addition of exogenous basal insulin to continue noninterventional posttrial diabetes diagnosed or treated by a non- (as insulin glargine) sufficient to normal- follow-up of participants, and those study physician during the trial without ize fasting glucose levels had a neutral that agreed obtained local ethics ap- documented confirmation) (1). Oral glu- effect on cardiovascular events, cancers, proval. ORIGIN participants (or family cose tolerance tests were not performed and other serious outcomes compared members in the case of death since the during posttrial follow-up. Instead, a di- with standard care without insulin in pa- last study visit) were contacted between agnosis of diabetes during this phase tients with dysglycemia at high risk for July 2012 and May 2014 to ascertain was based on either HbA1c $6.5% cardiovascular disease and reduced the their clinical status. Participants providing (48 mmol/mol) or the new use of any incidence of diabetes in participants informed consent were invited to attend glucose-lowering drugs. All events were without diabetes at randomization (1). up to two visits during that period: the verified by ensuring that they were con- Omega-3 fatty acid supplements are first at the time of initial contact and the sistent with the same event definition cri- widely taken for their putative cardio- second between February and May teria used during the ORIGIN trial (1,3). protective effects (2). However, the find- 2014. Participants whose first visit was ing that random allocation to 1 g omega-3 completed after November 2013 did Statistical Analysis fatty acid supplements daily versus pla- not need a second visit. At study visits, SAS software was used for analyses ac- cebo had a neutral effect on cardiovascular clinical outcomes, episodes of severe cording to a prespecified plan finalized death and other serious health outcomes hypoglycemia requiring third-party before the review of any data by original in the ORIGIN trial did not support such assistance or of any symptomatic hypo- treatment group. The nominal level of supplementation (3). glycemia (1), medication use, blood pres- significance was P , 0.05, with no ad- Several type 2 diabetes trials have sure, anthropometry, HbA1c, and serum justments made for multiple testing. All reported that the health effects of in- creatinine were assessed. If the partici- data were analyzed from the time of terventions often persist for several pant could not attend in person, as much randomization until the first occurrence years after the active intervention pe- data as possible were gathered for that of the relevant event, whether it oc- riod (4–7). Analyses of such legacy visit by telephone or from the medical curred during the active treatment effects can provide further information record. phase or the posttrial follow-up phase. about the long-term health effects of a Data for each participant were censored time-limited intervention. The aim of the Outcomes at the time of the relevant event, death, ORIGIN and Legacy Effects (ORIGINALE) Apart from the composite microvascular or last contact. Kaplan-Meier curves study was to determine the effects of a outcome and incident diabetes, the were plotted and comparisons made us- median 6.2 years of exposure to insulin ORIGINALE outcomes were identical to ing stratified log-rank tests. Hazard ra- glargine versus standard care or omega-3 those of the ORIGIN trial. The two copri- tios (HRs) were calculated based on Cox fatty acids versus placebo on cardiovascu- mary composite outcomes for the glar- regression stratified by allocation in the lar and other outcomes during an addi- gine versus standard care comparison other factorial arm, baseline diabe- tional median follow-up of 2.7 years. were death from cardiovascular causes tes status, and baseline history of car- or myocardial infarction or stroke and diovascular events. As in the original RESEARCH DESIGN AND METHODS any of these three outcomes or hospi- ORIGIN analyses, for outcomes with The design and main results of ORIGIN talization for heart failure or carotid, fewer than five events in a stratum, have been published previously (1,3,8). coronary, or peripheral revasculariza- Cox regressions treating these three Briefly, 12,537 people aged $50 years tion. The primary outcome for the factors as covariates were used (1). with impaired fasting glucose, impaired omega-3 versus placebo comparison For incident diabetes and hypoglyce- glucose tolerance, or type 2 diabetes was death from cardiovascular causes. mia, odds ratios (ORs) were calculated, were included if they had additional Because urine specimens were not col- and comparisons were made using risk factors for cardiovascular disease. lected during the ORIGINALE study, Cochran-Mantel-Haenszel tests strati- Through a 2 32 factorial design, participants the microvascular composite outcome fied by allocation to omega-3 or placebo were randomized to the addition of measured during the ORIGIN trial (death and baseline history of cardiovascular one daily injection of insulin glargine due to renal failure, dialysis, doubling events. Participants were analyzed ac- targeting a fasting plasma glucose of of serum creatinine, treated diabetic cording to the groups to which they #5.3 mmol/L (95 mg/dL) versus standard retinopathy, or progression of albumin- were randomized. care and to n-3 polyunsaturated fatty uria) (1) could not be measured. Some sites were unable to include all acid supplement 1 g daily (omega-3) ver- Instead, a clinical microvascular outcome their surviving participants in ORIGINALE. sus placebo. that excluded albuminuria progression Participants who agreed to ORIGINALE The trial finished in December 2011, but retained all of the other elements follow-up may have differed from those and subsequent use of all medications of the composite outcome was used.