Background & Aims

The clinical course of chronic pancreatitis is unpredictable. There is no model to assess disease severity or progression or predict patient outcomes.

Methods

We performed a prospective study of 91 patients with chronic pancreatitis; data were collected from patients seen at academic centers in Europe from January 2011 through April 2014. We analyzed correlations between clinical, laboratory, and imaging data with number of hospital readmissions and in-hospital days over the next 12 months; the parameters with the highest degree of correlation were used to develop a 3-stage chronic pancreatitis prognosis score (COPPS). The predictive strength was validated in 129 independent subjects identified from 2 prospective databases.

Results

The mean number of hospital admissions was 1.9 (95% confidence interval [CI], 1.39–2.44) and 15.2 for hospital days (95% CI, 10.76–19.71) for the development cohort and 10.9 for the validation cohort (95% CI, 7.54–14.30) (P = .08). Based on bivariate correlations, pain (numeric rating scale), level of glycated hemoglobin A1c, level of C-reactive protein, body mass index, and platelet count were used to develop the COPPS system. The patients’ median COPPS was 8.9 points (range, 5–14). The system accurately discriminated stages of disease severity (low to high): A (5–6 points), B (7–9), and C (10–15). In Pearson correlation analysis of the development cohort, the COPPS correlated with hospital admissions (0.39; P < .01) and number of hospital days (0.33; P < .01). The correlation was validated in the validation set (Pearson correlation values of 0.36 and 0.44; P < .01). COPPS did not correlate with results from the Cambridge classification system.

Conclusions

We developed and validated an easy to use dynamic multivariate scoring system, similar to the Child-Pugh-Score for liver cirrhosis. The COPPS allows objective monitoring of patients with chronic pancreatitis, determining risk for readmission to hospital and potential length of hospital stay.

中文翻译：

---

2个独立队列的慢性胰腺炎预后评分的制定和验证

背景与目标

慢性胰腺炎的临床过程是不可预测的。没有模型可以评估疾病的严重程度或进展或预测患者的预后。

方法

我们对91例慢性胰腺炎患者进行了一项前瞻性研究。数据收集自2011年1月至2014年4月在欧洲学术中心就诊的患者。我们分析了临床，实验室和影像学数据与未来12个月住院再住院次数和住院天数之间的相关性；相关程度最高的参数用于制定3期慢性胰腺炎预后评分（COPPS）。预测强度在从2个前瞻性数据库中鉴定出的129个独立受试者中得到验证。

结果

发展队列的平均住院人数为1.9（95％置信区间[CI]，1.39–2.44），住院天的平均住院次数为15.2（95％CI，10.76-19.71），验证队列的平均住院人数为10.9（95％CI， 7.54–14.30）（P  = .08）。基于双变量相关性，使用疼痛（数字等级量表），糖化血红蛋白A1c水平，C反应蛋白水平，体重指数和血小板计数来开发COPPS系统。患者的COPPS中位数为8.9点（范围5-14）。该系统可以准确地区分疾病严重程度的各个阶段（从低到高）：A（5-6分），B（7-9）和C（10-15）。在对发展队列的皮尔逊相关分析中，COPPS与住院人数（0.39; P <.01）和住院天数（0.33; P <.01）。在验证集中验证了相关性（Pearson相关值分别为0.36和0.44；P <0.01）。COPPS与剑桥分类系统的结果不相关。

结论

我们开发并验证了一种易于使用的动态多元评分系统，类似于用于肝硬化的Child-Pugh-Score。COPPS可以对慢性胰腺炎患者进行客观监测，确定再次入院的风险和可能的住院时间。