Sensitivity to biases of case-control studies on medical procedures, particularly surgery and blood transfusion, and risk of Creutzfeldt-Jakob disease.

BACKGROUND Evidence of risk of Creutzfeldt-Jakob disease (CJD) associated with medical procedures, including surgery and blood transfusion, is limited by susceptibility to bias in epidemiological studies. METHODS Sensitivity to bias was explored using a central-birth-cohort model using data from 18 case-control studies obtained after a review of 494 reports on medical procedures and risk of CJD, systematic for the period January 1, 1989 to December 31, 2011. RESULTS The validity of the findings in these studies may have been undermined by: recall; control selection; exposure assessment in life-time periods of different duration, out of time-at-risk of effect, or asymmetry in case/control data; and confounding by concomitant blood transfusion at the time of surgery. For sporadic CJD (sCJD), a history of surgery or blood transfusion was associated with risk in some, but not all, recent studies at a ≥10 year lag time, when controls were longitudinally sampled. Space-time aggregation of surgical events was not seen. Surgery at early clinical onset might be overrepresented among cases. Neither surgical history nor blood transfusion unlabelled for donor status, dental treatments or endoscopic examinations were linked to variant CJD (vCJD). CONCLUSIONS These results indicate the need for further research. Common challenges within these studies include access to and content of past medical/dental treatment records for diseases with long incubation periods.