Resident Physician Children's Mercy Mission, Kansas, United States
Abstract: Introduction With the implementation of venous thromboembolism (VTE) prophylaxis in our interstage patients we have seen an increase in incidence of hematochezia. Bloody stools are also an indicator of cardiac necrotizing enterocolitis (cNEC), thought to be secondary to poor perfusion and ischemia. It has been difficult to differentiate between hematochezia solely from anticoagulation and hematochezia from a poorly perfused gastrointestinal tract, identified as pneumatosis on an abdominal x-ray. We aimed to detect clinical prognostic factors that might serve to predict which patients with bloody stools were more at risk of developing cNEC. Methods This was a single center, retrospective chart review study. Demographic, pre-operative and post-operative clinical data were obtained for all post-operative interstage patients from January 2022 to May 2022. Classification and regression tree (CART) analysis was used to identify variables associated with bloody stools and pneumatosis on abdominal x-ray. Results Of the 70 study patients, 53 (76%) had at least one episode of bloody stools, with 29 (41%) having multiple episodes. Being treated with Aspirin or Lovenox for was the strongest predictor of multiple bloody stools, experienced by 55% of the 53 patients who received Aspirin or Lovenox vs. 0% of the 17 patients who did not. Among patient on Aspirin or Lovenox, being off nasal cannula oxygen at time of transfer out of the intensive care unit (ICU) was associated with increased risk of multiple bloody stools (62% without oxygen vs. 27% with). The full CART model had 76% sensitivity and 83% specificity. Of 53 patients with bloody stools, 20 (38%) were diagnosed with cNEC by identifying pneumatosis on abdominal x-ray. In the 103 incidents of bloody stools, the strongest risk factor for pneumatosis was patient birthweight < 3.1 kg (38% of incidents resulting in pneumatosis vs. 16% for birthweight >3.1 kg). In bloody stool incidents for the lower-birthweight group, the strongest predictor of pneumatosis was bloody stool occurring in the ICU. The full model had 20% misclassification rate, 48% sensitivity, and 84% specificity. White blood cell count (WBC) and c-reactive protein (CRP) were comparable for bloody stool incidents without vs. with pneumatosis (9.6, 95% IC 7.5-12.2 vs 9.8, IC 8.4-12.4; P=0.7 and 0.8, IC 0.5-1.4 vs 0.9, IC 0.5-1.6; P=0.4, respectively). Conclusions Treatment with Aspirin or Lovenox was the strongest predictor for multiple bloody stools during the interstage period. We were unable to identify predictive variables with reliable sensitivity for determining which bloody stool incidents would lead to cNEC diagnosis. WBC and CRP appear to play no role in identifying patients with cNEC in our cohort. More research is needed to address the possible protective value against bloody stools of nasal cannula oxygen therapy at time of transfer out of the ICU.
