Nomogram for predicting fulminant necrotizing enterocolitis

Necrotizing enterocolitis (NEC) is a devastating disease that occurs during the neonatal period and is an important contributor to newborn mortality. Approximately 30% of very-low-birth-weight preterm infants die from NEC, but the mortality rates from NEC in extremely low-birth-weight preterm infants are higher, ranging from 30% to 50.9% [1‐3]. Surviving infants may develop digestive tract and neurological sequelae.

Fulminant NEC (FNEC) is the most serious NEC subtype, accounting for approximately 10% of NEC cases and causing precipitous decline and death. Additionally, a subset of these infants with FNEC have a particularly virulent form of NEC known as NEC-totalis. Definitions of NEC-totalis vary in the literature but generally refer to patients with massive, near-total bowel ischemia and necrosis [4‐7]. The lack of typical imaging findings and abdominal signs makes identification and diagnosis difficult. Counseling parents regarding decision-making in these difficult cases is challenging given the lack of clinical data for these patients. Previous studies have shown that early recognition of the clinical situation, timely diagnosis, and surgical therapy for FNEC may be important in reducing mortality [4]. Although it is known that the pathophysiology of FNEC is shaped by the combined action of feeding, immunological, infectious, drug-induced, and hemodynamic factors [6‐9], most studies independently assessed the risk factors predicting FNEC, such as lower lymphocyte counts, lower platelet counts, and full-volume feeding [6, 7]. However, few studies have comprehensively predicted FNEC onset. We hypothesize that the nomogram can predict the occurrence of FNEC. In this study, we synthetically analyzed perinatal risk factors, feeding strategies, clinical characteristics, and hematological counts in infants with FNEC. Importantly, we established a nomogram prediction model with common clinical indicators to determine the predictors of FNEC with the hope of providing further direction for the early diagnosis and timing of intervention.

Previous studies have determined that the main risk factors for FNEC are hematological abnormalities and rapid feeding escalation using multivariate regression analysis. Our findings identified seven factors related to the risk of FNEC through single factor and multivariate regression analyses, including assisted ventilation after onset; shock at onset; feeding volume before onset; neutrophil counts on the day of onset; and neutrophil, lymphocyte, and monocyte counts on day 1 after NEC onset, all of which were used to establish a nomogram for predicting FNEC. In recent years, prognosis models based on statistical methods have rapidly developed; however, to the best of our knowledge, this is the first nomogram study to predict FNEC. Routine blood examination is easily analyzed, often available around the clock in all kinds of medical institutions. To know that a larger decrease in blood counts corresponds to increased odds of severe NEC at NEC onset can be helpful in the clinical management of NEC infants. Based on the area under the curve and calibration curve evaluation, this prognostic model showed good discriminatory ability, calibration, and clinical usefulness.

Garg et al. reported that a fulminant disease course was associated with a more extensive clinical disease with prominent radiological signs and a greater need for assisted ventilation and inotropes before and after surgery [6]. Lin et al. reported that FNEC is characterized by urgent onset and prompt deterioration [15]. In the current study, we found that assisted ventilation and shock at FNEC onset were predictors of FNEC. However, these characteristics also exist at the onset of neonatal sepsis. Therefore, although both are predictors of FNEC in our nomogram model, they still need to be combined with other factors to predict the onset of FNEC. In previous studies, higher lactate levels were associated with NEC, and statistically higher blood lactate levels were found in non-survivors of NEC compared to those of survivors [16, 17]. Lactate level was found to be a valuable prognostic factor for NEC in preterm neonates with NEC [18]. However, in our study, lactate level could not predict FNEC, which may be related to the treatment of tissue perfusion, such as rectification of acid intoxication, use of vascular active medicine, and increased blood capacity.

In Garg’s predictive model, for a 5% increase in neutrophils on the day of NEC diagnosis relative to pre-NEC neutrophil percentages, the odds of FNEC decreased by 16% [6]. Lambert et al. reported lower lymphocyte counts in neonates with FNEC [19]. However, they did not reveal other relationships with blood values. Our findings suggest that neutrophil counts at NEC onset and neutrophil, lymphocyte, and monocyte counts at 24 h after NEC onset were both predictors of FNEC, which suggests the need to monitor blood count changes over time. Systemic inflammation during NEC has been associated with altered platelets and leukocytes [20]. At the onset of NEC, neutrophils reach the intestinal lamina propria as first responders to sites of inflammation to perform a variety of antimicrobial functions, such as phagocytosis and formation of massive amounts of reactive oxygen species and other toxic molecules, which not only effectively destroy pathogens but also cause mucosal injury to a certain extent [21]. However, in the course of severe NEC, numerous neutrophils in the peripheral blood are rapidly recruited to the intestine and peritoneum or attached to the wall of small blood vessels, resulting in a large reduction of neutrophils in the peripheral blood circulation. As part of the normal gut inflammatory response, neutrophils are recruited to sites of infection or inflammatory stimuli within minutes, and the response peaks at 24–48 h [22], which is similar to the change in neutrophil counts after FNEC onset in our data. Previous studies have provided evidence that NEC can be considered as a lymphocyte-mediated disease. NEC development requires an influx of lymphocytes into the lamina propria intestine via toll-like receptor 4 signaling in the intestinal epithelium [23]. Monocyte-derived intestinal macrophages participate in the gut wall infiltration classically observed in NEC, and the rapid efflux of monocytes to NEC lesions is likely to deplete the limited circulating pool of monocytes in premature neonates [24]. This may explain why blood count changes can predict FNEC. However, the mechanism underlying NEC-related thrombocytopenia remains unclear. This mechanism may be linked to platelet activation and the consumption of platelets in microthrombi formed in the intestinal microvasculature during NEC-like injury [25, 26]. Thrombocytopenia is a common clinical finding in NEC, typically observed within 24–72 h after NEC development [27]. This may explain why the platelet count was not a sensitive predictor.

In our nomogram model, enteral feed volume was a predictor of the occurrence of FNEC. Hartman et al. reported that full-volume feeding was associated with FNEC on multivariate analysis [7], and Lambert et al. identified that more cases in the fulminant group had a faster escalation of feeding preceding the development of NEC [16]. This suggests that feeding practice affects the development of this more virulent form of NEC.

Our study also found lower neutrophil counts in neonates with FNEC on the first day after birth, and neutrophil, lymphocyte, and monocyte counts were still lower in the FNEC group on the second day after birth compared with those in the NEC group. In addition, the leukocyte counts in the FNEC group were lower than the average values reported in the existing literature [28‐30]. Tröger et al. reported that the numbers of white blood cells and neutrophils were diminished in small for gestational age infants at birth, and on day 3, these preterm infants had a higher risk for combined adverse outcomes [31]. Nguyen et al. reported that preterm pigs had very low blood neutrophil and lymphocyte counts and commonly showed poor weight gain and diarrhea [32, 33]. This suggests that there may be a connection between intestinal diseases and systemic immunity; in other words, neonates with immature systemic immune development who also have an immature digestive system may have an increased risk of gastrointestinal diseases, such as NEC. Therefore, individual feeding and therapy are necessary for premature babies with lower leukocyte counts. This interesting clinical phenomenon may provide clues for identifying high-risk infants with severe NEC; however, the causes need to be further explored.

This study has a few limitations. First, it was conducted retrospectively in a single institution. The nomogram must be confirmed in multicenter studies with large sample sizes. Second, this was a retrospective study based on a review of medical records, and patients with incomplete medical records were excluded. Third, the mechanisms of hematological counts and FNEC need to be further explored.

We developed a nomogram with seven factors, including assisted ventilation after NEC onset; shock at NEC onset; feeding volume before NEC onset; neutrophil counts on the day of NEC onset; and neutrophil, lymphocyte, and monocyte counts on day 1 after NEC onset. This nomogram may help in the individualized prediction and treatment of FNEC.