Article Text

Download PDFPDF

O06 A cumulative liver damage index (CLDI) identifies patients at risk of significant liver disease
  1. Emma Wesley1,
  2. Wolf-Rudiger Matull1,
  3. Angus Kitchin1,
  4. Katharine Hutchison1,
  5. Sam Madge2,
  6. Timothy Jobson1
  1. 1Somerset NHS Foundation Trust, Taunton, UK
  2. 2Environmental Futures and Big Data Lab, Exeter, UK


Background Late diagnosis of liver disease is a major factor in poor outcomes, with many missed opportunities for effective interventions. Identifying at risk patients is hampered by the high prevalence of abnormal liver tests in the population. Although it is intuitive that persistent elevation of liver enzymes is associated with liver damage, current risk scores don’t incorporate a measure of disease activity over time. The purpose of this study was to investigate the Cumulative Liver Damage Index (CLDI), a novel tool we developed, approximating the integral of enzyme elevation over time for an individual (ALT or ALP; calculated by the trapezium rule) as a predictor of liver disease, to be used for case finding.

Methods NHS laboratory data in Somerset, UK (441,785 patients, c140 million results) were imported into our case finding tool, linked to 4258 patients already diagnosed with liver disease over 15 years (39% F4 fibrosis) and anonymised. The CLDI (excluding inpatient tests) was calculated for ALT and ALP (I.U. x days). Results for patients known to have liver disease were compared to the general population. Trend lines for platelets (PLT) and bilirubin (BILI) were calculated for each patient (>4 data points).

Results The mean CLDI-ALT (integral up to 10 yrs) was 51.4k and was significantly higher in patients with any liver disease at 113k unit.days (p<0.000001) – Figure 1. A cut-off of CLDI-ALT of 250k was a strong predictor of any liver disease (O.R. 18.9; 95% CI 15.8–22.4) and F4 fibrosis (O.R. 19.0; 95% CI 14.7–24.5) and was associated with positive BILI trends (p<0.001; mean population BILI trend was near zero). Similarly a CLDI-ALP > 600k unit.days strongly predicted a rise in bilirubin and advanced liver disease. Population PLT counts showed a negative trend of -1.57x 109/yr with wide variation (s.d. 130x109/yr). F4 fibrosis was associated with a significant increase in the negative trend of PLT (mean -3.88±17.9x109/yr). Combining CLDI-ALT>250k with PLT trend less than -10.95x109/yr gave an even stronger predictor of F4 liver disease (O.R. 24.6; 95% CI 13.9–43.5).

Abstract O06 Figure 1

Comparison of CLDI-ALT for population and known liver disease

Conclusion These data demonstrate the utility of time trend analysis of laboratory data to identify patients at risk of advanced liver disease. Further work will define optimum combinations and gender specific thresholds for these indicators for use within our case-finding database, which we are using to recall patients with potential liver disease in Somerset.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.