Are Inflammatory Markers Predictive of Length of Hospital Stay in Patients with Inflammatory Bowel Disease?

Abstract

Purpose: Formulas such as the Crohn's Disease Activity Index and Harvey Bradshaw Index rely on subjective criteria and have had poorly documented relationships to inflammation in the gut and extraintestinal tissues. C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) levels, however, have served as objective measures of IBD activity. If CRP and ESR forecast severity of disease, then a prolonged hospital stay might be expected. We sought to determine if CRP and ESR were correlated with length of hospital stay (LOS) in IBD patients and if there was a difference in those with Crohn's disease (CD) compared to ulcerative colitis (UC). Methods: IRB-approved retrospective study of 35 IBD patients admitted for an IBD flare in the past 10 years. Inclusion criteria were age >18, diagnosis of UC or CD,ESR and/or CRP taken within the first 2 days of admission, and an admission for an IBD flare. Statistical analysis done using SAS version 9.2. Results: All patients were of AfroCaribbean descent. Seventeen patients had CD and 18 patients had UC. Fourteen patients had IBD for more than 5 years. Mean age was 38 years. Fifteen patients had extraintestinal manifestations (EIM). For all IBD patients, mean CRP level was 62.0 mg/L (normal value <0.4) and mean ESR was 58.7 mm/hr (normal value 0-20). Average LOS was 8.1 days. Correlation of CRP with LOS was r=0.21 (p=0.29) and ESR with LOS r=0.271 (p=0.13). Stronger correlations were found with males (CRP r=0.40 and ESR r=0.48 in males; CRP r=0.22 and ESR r=0.23 in females). No difference was found in those with >5 years of IBD compared to <5 years (p=0.43). No difference was found in those with EIM compared to those without (p=0.79). In UC patients, mean CRP was 56.7 mg/L and ESR was 58 mm/hr, with a mean LOS of 8.9 days. Correlation of CRP with LOS was r=0.36 (p=0.18) and ESR with LOS being r=0.30 (p=0.22). In CD patients, mean CRP was 69 mg/L and ESR was 59 mm/hr, with the mean LOS being 7.2 days. Correlation of CRP with LOS was r=0.02 (p=0.94) and with ESR and LOS r= 0.26(p=0.38). Conclusion: We found no correlation between inflammatory markers and LOS. Differences may not have been seen due to our small sample size. In addition, our results come solely from AfroCaribbeans, an often under-reported group of IBD patients. Whether our findings can be applied to other ethnic IBD populations remain to be seen. No data have been available correlating inflammatory markers and LOS in ethnic populations. We question if stronger correlations could found among other ethnicities given that racial differences in phenotypic expression of IBD have been previously reported. In the meantime, though CRP and ESR have been correlated with disease activity, we do not encourage the use of these markers in forecasting LOS.