Cancer and bone fractures in observational follow-up of the RECORD study

Acta Diabetol. 2015 Jun;52(3):539-46. doi: 10.1007/s00592-014-0691-y. Epub 2014 Dec 19.


Aims: The RECORD study evaluated the effects of rosiglitazone on cardiovascular outcomes. A 4-year observational follow-up was added to the study to monitor the occurrence of cancer and bone fractures. We present the cancer and bone fracture data aggregated across the main study and its observational follow-up.

Methods: RECORD was a multicentre, open-label trial in people with type 2 diabetes on metformin or sulfonylurea monotherapy randomly assigned to addition of rosiglitazone (n = 2,220) or to a combination of metformin and sulfonylurea (n = 2,227). At the end of the main study, patients stopped study drug and were invited to enter the observational follow-up during which glucose-lowering treatment was selected by the patient's physician. Serious adverse events of cancer and serious and non-serious events of bone fracture were recorded. The study is registered with, number NCT00379769.

Results: Of the 4,447 patients comprising the intent-to-treat population, 2,546 entered the observational follow-up (1,288 rosiglitazone, 1,258 metformin/sulfonylurea) and added 9,336 patient-years experience to the main RECORD study, making an aggregate of 33,744 patient-years. Based on the totality of follow-up, malignancies were reported in 179 of 2,220 patients (8.1 %) in the group originally randomised to rosiglitazone and in 195 of 2,227 patients (8.8 %) in the group allocated metformin/sulfonylurea [relative risk, RR, 0.92 (95 % CI 0.76-1.12)]. More patients reported bone fractures in the rosiglitazone group (238, 10.7 %) than in the metformin/sulfonylurea control [151, 6.8 %; RR 1.58 (1.30-1.92)]. For women, the corresponding figures were rosiglitazone 156 (14.5 %), metformin/sulfonylurea 91 (8.5 %), RR 1.71 (1.34-2.18), and for men, the corresponding figures were rosiglitazone 82 (7.2 %), metformin/sulfonylurea 60 (5.2 %), RR 1.37 (0.99-1.90). Potentially high-morbidity fractures (hip, pelvis, femur, and spine) occurred in the same number of patients (31, 1.4 %) in the two treatment groups.

Conclusions: We conclude that data from a 4-year observational follow-up, combined with the main RECORD study data, do not suggest an increased risk of cancer in patients randomised to rosiglitazone combination use compared with those randomised to metformin/sulfonylurea. Consistent with the main study, rosiglitazone is associated with an increased risk of peripheral bone fracture in women, and probably in men, but the combined data do not suggest an increase in potentially high-morbidity (hip, pelvis, femur, and spine) fractures.

Publication types

  • Multicenter Study
  • Observational Study
  • Randomized Controlled Trial
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Aged, 80 and over
  • Diabetes Mellitus, Type 2 / complications
  • Diabetes Mellitus, Type 2 / drug therapy*
  • Diabetes Mellitus, Type 2 / metabolism
  • Drug Therapy, Combination
  • Female
  • Follow-Up Studies
  • Fractures, Bone / etiology*
  • Humans
  • Hypoglycemic Agents / administration & dosage
  • Hypoglycemic Agents / adverse effects*
  • Male
  • Metformin / administration & dosage
  • Metformin / adverse effects
  • Middle Aged
  • Neoplasms / etiology*
  • Rosiglitazone
  • Sulfonylurea Compounds / administration & dosage
  • Sulfonylurea Compounds / adverse effects
  • Thiazolidinediones / administration & dosage
  • Thiazolidinediones / adverse effects


  • Hypoglycemic Agents
  • Sulfonylurea Compounds
  • Thiazolidinediones
  • Rosiglitazone
  • Metformin

Associated data