Are fractures due to anticoagulants or the disease necessitating the treatment?

Abstract

Assessing potential adverse effects of drugs is not an easy task. Iyisoy et al. raises relevant methodological questions and advise clinician to be cautious on using data from non-randomised trials in clinical practice. Unfortunately, randomised controlled trails (RCT) performed prior to the acceptance of new drugs do not always reveal all adverse effects [ [1] Stricker B.H. Psaty B.M. Detection, verification, and quantification of adverse drug reactions. BMJ. 2004; 329: 44-47 Crossref PubMed Scopus (154) Google Scholar ]. Moreover, post-marketing's RCT are not always an option due to ethical concerns. A good example is whether treatment with vitamin K antagonists (VKA) increases the risk of fracture. Due to the well-documented beneficial effect of VKA in e.g. atrial fibrillation (AFLI), it is not an option to randomise patients with AFLI to treatment with either VKA or placebo. Neither is it feasible to randomise healthy subjects in order to study potential harmful effects. Thus, other methodological approaches are needed. In our pharmaco-epidemiological study, we took advantages of the uniformly organized Danish health care system [ [2] Rejnmark L. Vestergaard P. Mosekilde L. Fracture risk in users of oral anticoagulants: A nationwide case–control study. Int J Cardiol. 2007; 118: 338-344 Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar ]. In Denmark, it is not possible to redeem a prescription for a drug, without registration in the national prescription database. Moreover, the Danish National Hospital Discharge Register covers all contacts (on in- or out-patient basis) to the hospitals with an almost 100% completeness of recordings and a high precision of diagnoses, especially of fracture diagnoses [ [3] Vestergaard P. Mosekilde L. Fracture risk in patients with celiac Disease, Crohn's disease, and ulcerative colitis: a nationwide follow-up study of 16,416 patients in Denmark. Am. J. Epidemiol. 2002; 156: 1-10 Crossref PubMed Scopus (273) Google Scholar ]. Using these registers it is possible to identify, with an almost 100% certainty, all patients who have sustained a fracture (cases), as well as subject who have not fractured (controls). As Denmark has a Civil Registration System, which has electronic records on all citizens it is possible to make a valid link between these registers. Iyisoy et al. state that our study is a “retrospective” case–control study. It is true, that case–control studies by design are retrospective, meaning that they are comparing previous risk factors between cases and controls. They may therefore typically be at risk of i.e. recall bias. However, all data have been registered prospectively in our study i.e., data on exposure and confounders were collected before the date of hospitalization for fracture. Therefore, recall bias did not influence data collection. However, all study designs have drawbacks e.g., our analysis may be influenced by potential confounding factors not included in the analyses.