P195 Predictors of colectomy in patients with acute severe ulcerative colitis: single-center retrospective analysis.
Mingazov, A.(1);Achkasov, S.(1);Vardanyan, A.(1);Baranova, T.(1);
(1)National medical research center of Coloproctology, Oncology and Colon Surgery Department, Moscow, Russian Federation
Background
Acute severe ulcerative colitis is observed in 25% of patients1. Near 40% of this severe patients might undergo surgery2 and the mortality rate in this group of patients reaches 6.3%3. Untimely surgery leads to severe metabolic disorders, infectious postoperative complications, and increases the risk of death4,5. Eventually, the purpose of this study is identify predictors of colectomy in patients with acute severe ulcerative colitis.
Methods
A retrospective study included 74 patients with acute severe ulcerative colitis, who were treated at the clinic in 2017(Fig. 1). The patients were divided into the groups: group of colectomy - 54/74 (73%) and group of conservative treatment 20/74 (27%) (Fig 1). The predictors, like serum albumin, C-reactive protein, hemoglobin, endoscopic picture of «extensive ulcers», and clinical data were analyzed. The statistical analysis was performed using the software«Statistica 13.3» and «RStatistica».
Figure 1.
Results
There were no difference by gender, age and duration of the disease in the groups. Mean of albumin and hemoglobin levels were significantly lower in the colectomy group. The endoscopy «extensive ulcers» was significantly more common in the operated patients (Tab 1). Also univariate, multivariate and ROC analyses were performed (Tab 2), (Fig 2).
The risk of colectomy when endoscopy «extensive ulcers» combined with an albumin level<31 g/l and hemoglobin<107 g/l, was 100%. A nomogram for predicting the probability of colectomy was constructed (Fig 3). This logistic model has a statistically significantly high predictive value (AUC=0.93, p=0.006).
Table 1.
| Factors | Colectomy, n=54 | No colectomy, n=20 | P |
| Age, median | 36 (18-75) | 32 (23-66) | 0.33 |
| Males, % Females, % | 28 (52%)26 (48%) | 12 (60%)8 (40%) | 0.6 |
| Time of disease, median, months | 12 (0-200) | 31 (1-200) | 0.14 |
| Acute colitis, % | 16 (29%) | 5 (25%) | 0.77 |
| Biological therapy, % | 10 (18%) | 10 (50%) | 0.001 |
| Albumin, mean, g/l | 28 (±5,2) | 34 (±4,8) | 0.0002 |
| CRP, median, mg/l | 67 (1-200) | 35 (1-120) | 0.01 |
| Hemoglobin, mean, g/l | 96 (±19) | 118 (±18) | <0.0001 |
| Endoscopy "extensive ulcers", % | 42 (78%) | 1 (5%) | <0.0001 |
Table 2.
| Factor | OR (univariate) | p | OR (multivariate) | p |
| Albumin | 0.8 (0.71-0.91) | 0.001 | 0.67 (0.48-0.9) | 0.017 |
| CRP | 1.01 (0.99-1.02) | 0.053 | ||
| Hemoglobin | 0.94 (0.91-0.97) | 0.0005 | 0.88 (0.79-1.0) | 0.018 |
| Endoscopic picture | 83.6 (9.9-699.8) | 0.000045 | 2926 (17.1-499021.5) | 0.0023 |
Figure 3.
Conclusion
The endoscopic picture of «extensive ulcers» in combination with an albumin level of less than 31 g/l and hemoglobin less than 107 g/l are predictors of colectomy with high predictive value.