Liver cirrhosis in England--an observational study: are we measuring its burden occurrence correctly?

Objectives: Mortality due to liver disease (of which cirrhosis is the end‐stage) is increasing more than any other chronic condition in the UK. This study aims to demonstrate that (i) exclusive reliance on mortality rates may not reveal the true burden of liver cirrhosis, and (ii) diverse use of diagnostic coding may produce misleading estimates. Design: Observational study Setting: The Office for National Statistics death registry was interrogated to investigate liver cirrhosis mortality trends in England and Wales, from 1968 to 2011. Main outcome: Standardised mortality trends according to three different definitions of liver cirrhosis based on the specificity of diagnostic codes were calculated: 1(chronic liver diseases), 2 (alcoholic and unspecified cirrhosis only) and 3 (cirrhosis as end‐stage liver disease). The mortality trends were compared to incidence rates established in a previous population‐based study (based on definition 3), from 1998 to 2009, to investigate discrepancies between these two measures. Results: Over the study period, the overall standardised liver cirrhosis mortality rates were 8·8, 5∙1 and 5∙4 per 100,000 person‐years for definitions 1, 2 and 3 of respectively. The mortality rates for definition 3 in 1998 and 2009 were 6∙2 and 5∙9 per 100,000 person‐years respectively; whilst the equivalent incidence rates were at least three‐ and six‐fold higher: 23∙4 and 35∙9 per 100,000 person-years respectively. This discrepancy between incidence and mortality rates was also at least three‐fold in men and women separately, and across age‐groups. Conclusion: Mortality rates underestimated the incidence of liver cirrhosis by at least three‐fold between 1998 and 2009 and varied with differing definitions of disease. Mortality data should not be used exclusively as an indicator for the occurrence of liver cirrhosis in the population. Routinely collected healthcare data are available to measure occurrence of this disease. Careful consideration should be taken when selecting diagnostic codes for cirrhosis