Development and validation of a Chronic Pancreatitis Prognosis Score (COPPS) in two independent cohorts

Introduction:

The clinical course of chronic pancreatitis is unpredictable and so far, no prediction model to assess the dynamic disease severity has been established.

Patients and Methods:

We prospectively recruited 111 patients with chronic pancreatitis. Clinical, laboratory and imaging data were correlated to number of readmissions and in-hospital days over the next 12 months, and parameters showing best correlation were used for development of a three stage chronic pancreatitis prognosis score (COPPS). The predictive value was validated in 129 independent subjects identified from two prospective databases.

Results:

The median number of hospital admissions was 2.3 [95%CI 1.8 – 2.8] and 16.5 for hospital days [95%CI 12.7 – 20.2] for the development cohort and 10.9 for the validation cohort [95% CI 7.5 – 14.3] (P =0.08). Based on bivariate correlations, pain (numeric rating scale), HbA1c, CRP, BMI and thrombocyte count were used for development of COPPS. Median COPPS was 9.1 points (range 5 to 14) and it successfully discriminated stages of disease severity (low to high): A (5 – 6 points), B (7 – 9) and C (10 – 15). Pearson correlation showed significant correlation to hospital admissions (0.52; P<0.01) and number of hospital days (0.40; P<0.01) in the development cohort. This correlation was validated in the validation set (0.44; P < 0.01). There was no correlation with the Cambridge classification.

Conclusion:

We developed and validated a novel, easy to use dynamic multivariate score similar to the Child-Pugh-Score for cirrhosis. COPPS allowed objective monitoring of chronic pancreatitis regarding readmission to hospital as well as length of hospital stay.