Papers

Mortality from liver disease in the West Midlands, 1993-2000: observational study

N C Fisher, consultant physician and gastroenterologist ^a, J Hanson, public health information specialist ^b, A Phillips, director of public health ^c, J N Rao, consultant in public health medicine ^d, E T Swarbrick, consultant physician and gastroenterologist ^e. 

^a Department of Gastroenterology, Dudley Group of Hospitals NHS Trust, Russells Hall Hospital, Dudley, West Midlands DY1 2HQ, ^b Dudley Health Authority, Dudley DY12 2DD, ^c Wolverhampton Health Authority, Wolverhampton WV13 0XE, ^d Sandwell Health Authority, West Bromwich B70 9LD, ^e University of Wolverhampton, Division of Clinical Sciences, New Cross Hospital, Wolverhampton WV10 0QP

Correspondence to: N Fisher Neil.Fisher{at}dudleygoh-tr.wmids.nhs.uk

Advanced liver failure carries a poor prognosis, and its prevalence may be reflected by mortality statistics in the form of death certifications for liver disease. In the United Kingdom mortality from cirrhosis and other liver diseases increased slowly from the 1970s to the early 1990s.¹ We aimed to ascertain the current mortality from liver disease in the West Midlands region of the United Kingdom.

	Methods and results

Top Methods and results Comment References

The study was set in three adjacent metropolitan boroughs in the West Midlands with a total population of 837 000. Around 8.4% of residents are of south Asian origin (Indian, Pakistani, or Bangladeshi; 1991 census). Deaths from liver disease were identified from public health mortality files supplied by the Office for National Statistics, which we searched using ICD-9 (international classification of diseases, 9th revision) reference codes 570-573, and from files supplied by the registrar of the local health authority. South Asian origin and religion were identified from subjects' names. In cases of deaths from liver disease of unspecified cause (ICD 571.5 and related codes) we analysed case notes to search for underlying causative factors.

Crude mortality from primary liver disease increased from 6.0 per 100 000 population in 1993 to 12.7/100 000 in 2000 (figure). The increase was almost exclusively the result of alcoholic liver disease (ICD codes 571.0-571.3), which increased almost threefold from 2.8/100 000 in 1993 to 8.0/100 000 in 2000 (regression coefficient +0.89/100 000/year, 95% confidence interval 0.57 to 1.21), although it seemed to have stabilised from 1998 onwards. Rates of increase in deaths from alcoholic liver disease were similar for white men, white women, and Asian men. Asian men had a standardised mortality ratio 3.79 times (3.21 to 4.26) that of white men (based on 46 observed deaths compared with 12.4 expected by extrapolation from the white male population).² Eighty per cent of the Asian men were judged to be of Sikh religion.

View larger version (19K): [in this window] [in a new window]   Crude mortality from primary liver disease in the West Midlands per 100 000 population, 1993-2000

After alcoholic liver disease, the largest cause of death was "unspecified" liver disease, with an annual incidence of 2.5/100 000 population (figure). Alcohol misuse was the presumed cause in 67% (44) of 66 such cases as judged by analysis of case notes. Annual mortality from other defined liver diseases was about 0.5/100 000.

	Comment

Top Methods and results Comment References

Deaths from alcoholic liver disease increased in the West Midlands in the past decade; this is also a nationwide trend.³ The increasing prevalence of alcoholic liver disease is corroborated by our own data and data on increasing admission rates for patients with alcoholic liver disease to hospital.⁴

This apparent increase might be the result of increasing alcohol consumption, but available evidence does not show any notable increase in the total national alcohol consumption in the past decade nor in the number of people drinking heavily.⁵ The type of alcoholic drink consumed and dietary or other (such as genetic or unidentified environmental) factors may therefore be implicated. The increasing prevalence of alcoholic liver disease among Asians may also be contributing; although the overall population of Asians in our study is small, their excess risk of mortality is worth further study. Our study cannot disprove the possibility of an artefactual increase in morbidity and mortality from alcoholic liver disease as a result of clinicians changing their certification practice, although we are unaware of any evidence to support such a change in certification in the past decade.

These data have important implications for public health and hospital physicians. The halting or reversal of the trend in deaths from alcoholic liver disease that we have described requires further public emphasis on the risk of fatal liver disease from excessive alcohol consumption.

	Acknowledgments

We thank Tim Marshall, Department of Public Health and Epidemiology, University of Birmingham, for informal statistical advice and Amtar Ali and Amrik Jheeta for help with assigning Asian religions.

Contributors: NCF conceived and coordinated the study, and wrote the manuscript. JH and JNR computed the standardised mortality ratio details. JH, AP, JNR, and ETS each provided practical help and critical input into the study methodology, and each critically reviewed the final manuscript. G Criddle (Dudley Health Authority), J Gwinnett (Wolverhampton Health Authority), and C Matthews (University of Wolverhampton, Division of Clinical Sciences) extracted and tabulated data for the study.

	Footnotes

Funding: Wolverhampton Digestive Foundation.

Competing interests: None declared.

	References

Top Methods and results Comment References

1.	Noble B. Deaths associated with the use of alcohol, drugs, and volatile substances. Population Trends 1994; 76: 1-11.
2.	Armitage P, Berry G. Statistical methods in medical research. 3rd ed. London: Blackwell, 1994.
3.	Medical Council on Alcohol. June 2001. www.medicouncilalcol.demon.co.uk (accessed 7 May 2002).
4.	Chick J. Evidence suggesting increasing health damage in Scotland related to alcohol. Health Bull 1997; 55: 134-139.
5.	Department of Health. Statistics on alcohol: 1976 onwards. London: Stationery Office, 1999. (Health Bulletin 1999/24.)

(Accepted 25 January 2002)