Open Access Research article

Magnitude of discordance between registry data and death certificate when evaluating leading causes of death in dialysis patients

Jean-Philippe Lafrance123*, Elham Rahme45, Sameena Iqbal46, Martine Leblanc23, Vincent Pichette123, Naoual Elftouh1 and Michel Vallée23

Author Affiliations

1 Centre de recherche Hôpital Maisonneuve-Rosemont, Montréal, Canada

2 Département de Médecine, Université de Montréal, Montréal, Canada

3 Service de néphrologie, Hôpital Maisonneuve-Rosemont, 5415, boul. de l’Assomption, Montréal (Québec), H1T 2M4, Canada

4 Department of Medicine, McGill University, Montreal, Quebec, Canada

5 Research Institute, McGill University Health Centre, Montreal, Quebec, Canada

6 Division of Nephrology, McGill University Health Centre, Montréal, Québec, Canada

For all author emails, please log on.

BMC Medical Research Methodology 2013, 13:51  doi:10.1186/1471-2288-13-51

Published: 27 March 2013



Discordance between dialysis registry and death certificate reported death has been demonstrated. Since cause of death is measured using registry data in dialysis patients and death certificate data in the general population, comparisons of cause of death proportions between dialysis patients and the general population may be biased. Our aim was to compare the proportion of deaths attributed to cardiovascular disease (CVD), malignancy, and infections between patients receiving dialysis and the general population using death certificates for both, and to quantify the magnitude of discrepancy between registry and death certificate estimates in dialysis patients.


A retrospective cohort study of 5858 patients initiating maintenance dialysis between 2001 and 2007 was conducted. Cause of death was obtained from both registry and death certificate data for dialysis patients, and from death certificate data for the general population.


Compared to the general population, use of death certificate data in dialysis patients resulted in smaller differences in the proportion of deaths attributed to CVD or infection than that from the registry. In the general population, the proportion of deaths due to CVD is 29.3% for men and 28.2% for women, and the proportion of deaths due to infection is 3.3% for men and 3.6% for women. For men, the proportion of deaths in dialysis patients due to CVD using registry data is 41.5%, compared with a proportion of 32.1% using death certificate data. Similarly for women, the proportion of deaths due to CVD using registry data is 35.2% and that using death certificate data 24.3%. The proportion of deaths due to infection in dialysis patients follows the same pattern: for men, the proportion of deaths due to infection using registry data is 9.9% and that from death certificate data at 5.0%; while for women the proportions are 11.6% and 4.8%, respectively.


While absolute cause-specific mortality rates did differ, evaluation of causes of death using death certificate in dialysis patients in Quebec revealed that they do not have substantially different proportion of death due to CVD or infections than the general population. Infections appeared to be a frequent complication leading to death, suggesting that infections are an important target to consider for reducing mortality in dialysis populations.

Registries; Renal dialysis; Kidney failure; Chronic; Mortality; Cardiovascular diseases; Infections