Nine scoring models for short‐term mortality in alcoholic hepatitis: cross‐validation in a biopsy‐proven cohort

Abstract

BackgroundSeveral prognostic models have emerged in alcoholic hepatitis (AH), but lack of external validation precludes their universal use.AimTo validate the Maddrey Discriminant Function (DF); Glasgow Alcoholic Hepatitis Score (GAHS); Mayo End-stage Liver Disease (MELD); Age, Bilirubin, INR, Creatinine (ABIC); MELD-Na, UK End-stage Liver Disease (UKELD), and three scores of corticosteroid response at 1 week: an Early Change in Bilirubin Levels (ECBL), a 25% fall in bilirubin, and the Lille score.MethodsSeventy-one consecutive patients with biopsy-proven AH, admitted between November 2007-September 2011, were evaluated. The clinical and biochemical parameters were analysed to assess prognostic models with respect to 30- and 90-day mortality.ResultsThere were no significant differences in the areas under the receiver operating characteristics curve (AUROCs) relative to 30-day/90-day mortality: MELD 0.79/0.84, DF 0.71/0.74, GAHS 0.75/0.78, ABIC 0.71/0.78, MELD-Na 0.68/0.76, UKELD 0.56/0.68. One-week rescoring yielded a trend towards improved predictive accuracies (30-day/90-day AUROCs: 0.69–0.84/0.77–0.86). In patients with admission DF ≥32 (n = 31), response to corticosteroids according to ECBL, 25% fall in bilirubin and the Lille model yielded AUROCs of 0.73/0.73, 0.78/0.72 and 0.81/0.82 for a 30-day/90-day outcome respectively. All models showed excellent negative predictive values (NPVs; range: 86–100%), while the positive ones were low (range: 17–50%).ConclusionsMELD, DF, GAHS, ABIC and scores of corticosteroid response proved to be valid in an independent cohort of biopsy-proven alcoholic hepatitis. MELD modifications incorporating sodium did not confer any prognostic advantage over classical MELD. Based on excellent NPVs, the models are best to identify patients at low risk of death.

Letter: scoring models in alcoholic hepatitis – authors' reply

“…4 Therefore, it is not surprising that a significant effect of glucocorticosteroids on survival was not shown in our relatively small cohort. 2,5 Similarly, it is reasonable to speculate that treatment effects are negligible as a source of biasing prognostic performances in present series. The most striking finding by Wieten et al is that prognostic indices in AH may retain their prognostic ability for up to 3 years from presentation.…”

Model for end-stage liver disease-Na score or Maddrey discrimination function index, which score is best?

“…Many strategies have been used to predict morbidity and mortality in AH allowing a better medical support for those very ill patients. Such strategies include the search for single parameters (i.e., alkaline phosphatase) or the development of scoring systems like the Maddrey DFI, the GAHS, the ABIC, the Lille score and MELD [20][21][22][23][24] . According to our results we propose that MELD-Na is also a useful scoring system to predict severity in AH.…”

“…However, in this study the optimal Maddrey DFI cut off point corresponding to the optimal MELD score was higher than the conventional one and the authors propose that MELD score may be used as an alternative to DFI score for predicting short-term mortality in AH [17] . Three previous studies have compared the ability of MELD-NA to predict mortality compared to other scores [24][25][26] . The first study, a small sample size study from the Mayo Clinic, showed that MELD-Na was a better predictor of 180-d mortality than MELD in patients with ascites [26] .…”

“…Specifically for MELD-Na, the AUC was 0.83 to predict mortality at 90 d, similar to our findings. In a more recent study, nine scoring models were compared in 71 biopsy-proven patients with AH and all models showed excellent negative predictive values and MELD modifications incorporating sodium did not confer any prognostic advantage over classical MELD [24] . Interestingly, in this cohort the 30-d mortality and 90-d mortality rates were lower compared to other studies (14.1% and 19.7%, respectively).…”

“…MELD-Na is more efficient than MELD to identify subjects with poor outcome and significantly increase the efficacy of the score to predict waitlist mortality [22] . Several studies have examined the use of MELD in assessing the severity of AH [12,16,17,23,24] and sensitivity and specificity in predicting 30-d mortality ranges from 75% to 86%. Few studies, have evaluated the usefulness of the MELD-Na in AH [24][25][26] and results are controversial.…”