PDF
Citation
Print
INTRODUCTION
Enteral nutrition (EN) helps preserve mucosal structure and the immune system within the gastrointestinal (GI) tract, thereby reducing the risk of bacterial translocation, systemic infection, and the development of multiple organ dysfunction syndrome in critically ill patients [1-3]. Therefore, it is recommended to initiate enteral feeding (EF) within 24–48 hours of intensive care unit (ICU) admission for critically ill patients [4,5]. However, the initiation of EF is often hindered by various GI symptoms and signs, such as gastric retention, abdominal distension, diarrhea, nausea, and vomiting. Clinicians use the term enteral feeding intolerance (EFI) to describe this phenomenon [6]. Although broad and inconsistent definitions of EFI lead to variability in incidence across studies, it appears to be common, with reported rates ranging from 2% to 75% [7,8].
Previous studies have shown that EFI is closely associated with worse clinical outcomes, such as inadequate nutrition, fewer ventilator-free days, longer ICU and hospital stays, and higher mortality, emphasizing the importance of carefully monitoring EFI during EN feeding in critically ill patients [9-11]. Patients with septic shock are particularly vulnerable due to the need for high-dose vasopressor support, the risk of low GI tract perfusion, and the potential for complications such as mesenteric ischemia or, in severe cases, non-occlusive bowel necrosis upon initiation of EF [3,12]. Consequently, current guidelines have not established a consensus on the optimal timing for initiating EN in patients with septic shock [13].
This study aimed to determine quantify the number of days with EFI events in patients with septic shock and examine the association between frequency and clinical outcomes. We hypothesized that EFI would be highly prevalent among septic shock patients and that a higher frequency of EFI events would be associated with worse clinical outcomes.
MATERIALS AND METHODS
Ethics Statement
This study was conducted in compliance with the amended Declaration of Helsinki and was approved by the Institutional Review Board of Seoul National University Hospital (No. H-2202-049-1298). The requirement for written informed consent was waived due to the retrospective nature of the study.
Study Population
We included adult patients aged 18 years and older admitted to the medical ICU for septic shock at a tertiary referral center in South Korea from January 1, 2023 to November 30, 2023. Septic shock was defined in patients with a Sequential Organ Failure Assessment (SOFA) score of 2 or above who required vasopressor support to maintain a mean arterial pressure of 65 mm Hg or higher despite receiving adequate fluid resuscitation with 30 ml/kg of crystalloid within 3 hours of diagnosis [14]. The exclusion criteria were as follows: patients who were not receiving vasopressors at the time of EF initiation, those in moribund status, those readmitted to the ICU during the same hospitalization period, patients who did not receive EF during their ICU stay, and those with GI conditions contraindicating EF, such as GI obstruction, GI bleeding, ischemic colitis, esophageal rupture, peritonitis, or acute pancreatitis.
EF Protocol
Our institution implements a structured, multidisciplinary approach to EN, managed collaboratively by a dedicated Nutrition Support Team and ICU staff. Twice-weekly multidisciplinary rounds are conducted to optimize nutrition strategies. In hemodynamically stable patients without contraindications, EN is typically initiated at 300–600 kcal/day and advanced by 200–300 kcal/day based on clinical tolerance. However, for hemodynamically unstable patients, an individualized feeding protocol is established for each patient, with decisions on initiation of nutrition, caloric advancement, and formula selection made based on the risks and benefits.
Definitions and Assessment of EFI Events
We defined an EFI event as the presence of any of the following GI complications: vomiting, abdominal distension, abdominal pain, diarrhea, or ileus on abdominal x-ray. Diarrhea was defined by a stool volume of 250 ml or more per day or the occurrence of three or more stools per day [15]. Vomiting referred to any visible regurgitation of gastric contents. Abdominal distension was identified either through clinical suspicion or an increase in abdominal girth measured by the clinician [11]. Abdominal pain was defined as subjective discomfort in the abdominal area or the presence of tenderness or rebound tenderness observed during a physical examination. Ileus was diagnosed based on the radiologist’s report of the abdominal x-ray. The presence of EFI was assessed daily, and a day was considered an EFI day if any of these GI complications were observed. We then calculated the total number of EFI days during the first 72 hours of EF for analysis.
Clinical Variables
Demographic, laboratory, and clinical data at baseline and through the study period was collected through retrospective review of medical records. For norepinephrine equivalent (NEE) dose, we assessed the dose at the time of EF initiation and the maximum dose administered during each 24-hour period within the first 72 hours following EF initiation. To measure the NEE dose, we standardized all vasopressor and inotropic medications administered to patients into a single equivalent dose expressed as micrograms per kilogram per minute of norepinephrine. The NEE dose was calculated using the formula outlined in a previously published study [16]. To ensure consistency and relevance, doses that were administered continuously for more than 2 hours were included in the calculations.
Outcomes
The primary outcomes of this study were 30-day and 90-day mortality. Secondary outcomes were length of stay (LOS) in the ICU, LOS in the hospital after ICU admission, mechanical ventilation (MV)-free days within 28 days from the day of intubation, and number of EFI-related GI complications.
Statistical Analysis
For continuous variables, data were summarized using either the mean and standard deviation or the median and interquartile range. Categorical variables were summarized using frequencies and proportions. To assess the trend in the association between increasing total days of EFI and clinical characteristics/outcomes, the P-value for the trend was calculated using regression models. Linear regression was applied for continuous outcomes, while logistic regression was used for binary outcomes. Total days of EFI was treated as an ordinal variable, and the P-value for the trend was derived from the regression coefficient. Variables with P-values below 0.05 were considered to demonstrate significant trends.
Cox proportional hazards analysis was utilized to assess the impact of EFI on mortality outcomes, with hazard ratios (HRs) and 95% CIs calculated. Multivariable analysis was adjusted for variables associated with mortality including relevant clinical confounders. All statistical analyses were conducted using the R statistical software version 4.2.0 (R Foundation for Statistical Computing).
RESULTS
Study Population and Clinical Characteristics
Among a total of 451 patients with septic shock, aged 18 years or older, 357 were excluded based on the exclusion criteria. Figure 1 presents a flowchart outlining the patient selection process. A total of 94 patients were included in analysis. While 17 patients did not experience EFI during the first 3 days of EF, most patients (77 patients, 81.9%) had at least one episode of EFI. Among the patients who experienced EFI, 25 patients (26.6%) experienced EFI for 1 day, 22 patients (23.4%) experienced EFI for 2 days, and 30 patients (31.9%) experienced EFI for all 3 days.
Age, sex, body mass index (BMI), clinical frailty scale, high risk of refeeding syndrome, immunocompromised status, and comorbidities did vary significantly by number of EFI days. In contrast, blood urea nitrogen and creatinine levels increased significantly with higher total EFI days. The proportion of patients receiving renal replacement therapy (RRT) also rose progressively with increasing EFI days. Severity scores, including the Acute Physiology and Chronic Health Evaluation (APACHE) II score and the SOFA score, increased in parallel with total EFI days. mNUTRIC score was also significantly associated with longer duration of EFI. The most common site of infection was pulmonary (n=66), followed by abdominal (n=18), unknown origin (n=8), catheter-related (n=4), urinary tract (n=3), and skin/soft tissue (n=3) (Table 1).
Association between EFI Days and Mortality
In the analysis of mortality variables, an increase in total EFI days was significantly associated with increased ICU mortality (P=0.036), in-hospital mortality (P=0.007), 30-day mortality (P=0.004), and 90-day mortality (P=0.006) (Table 1). Multivariable Cox proportional hazards regression analysis demonstrated that number of EFI days was independently associated with 90-day mortality (adjusted HR [aHR], 1.400; 95% CI, 1.021–1.919; P=0.037) but not 30-day mortality (aHR, 1.496; 95% CI, 0.992–2.258, P=0.055) (Table 2).
Secondary Outcomes
LOS in both ICU and hospital showed a decreasing trend as the number of EFI days increased (P for trend=0.027 and 0.013, respectively). However, MV-free days within 28 days from the day of intubation did not exhibit a significant increase or decrease with an increasing number of EFI days. Similarly, no significant differences across groups or according to total EFI days were observed for the time from ICU admission to EF initiation, the time from septic shock to EF initiation, the amount of EF intake, the percentage ratio of delivered to target kcal in EF, average protein intake, or the concomitant administration of parenteral nutrition. There were no significant differences in NEE dose at the time of EF initiation or in the maximum dose on day 1 following EF initiation across groups. However, on days 2 and 3 following EF initiation, patients with a greater number of EFI days demonstrated a tendency toward higher maximum NEE doses (Table 3).
The most frequently observed GI complication was diarrhea, reported in 41 patients (43.6%) with a total of 102 occurrences. This was followed by abdominal distension, observed in 35 patients (37.2%) with 79 occurrences; ileus detected on abdominal x-ray, noted in 7 patients (7.4%) with 13 occurrences; vomiting, seen in 6 patients (6.4%) with 6 occurrences; and abdominal pain, reported in 2 patients (2.1%) with 3 occurrences (Figure 2).
DISCUSSION
This study demonstrated a relatively high prevalence of EFI (81.9%) among septic shock patients admitted to the ICU compared to the previously reported range of 2% to 75% in critically ill patients [6]. This discrepancy may be attributable to the study’s specific focus on patients with septic shock. EN increases splanchnic metabolic demands, which can result in oxygen and/or energy mismatch in the hypo-perfused GI tract during septic shock. This mismatch may cause the absorptive processes of the GI tract to exceed the limits of cellular metabolism, leading to cellular ischemia [17]. For these reasons, current guidelines provide conservative and vague recommendations regarding early EN support in patients with septic shock [18]. However, initiating EN as early as possible, even in the context of septic shock, can help preserve the mechanical and immunological functions of the gut barrier, stimulate intestinal trophism, reduce bacterial translocation, and lower the incidence of sepsis and multiple organ failure [19,20]. A meta-analysis of five randomized controlled trials and 10 non-randomized studies suggested that early EN in septic shock patients may reduce the duration of MV and SOFA scores in critically ill septic patients, albeit with an apparent increase in the frequency of non-serious adverse events such as diarrhea [21].
In our study, an increase in the total number of EFI days was significantly associated with higher 90-day mortality even after adjusting for relevant clinical confounders. As the total number of EFI days increased, there was a significant increase in both ICU mortality and in-hospital mortality. EFI is conceptually and clinically aligned with acute intestinal failure (AIF), a state of insufficient gut function to maintain health and homeostasis [22,23]. AIF reflects a continuum of GI dysfunction, which encompasses the features of EFI, especially in critically ill patients. AIF has increasingly been recognized as a central contributor to poor outcomes in the ICU due to its impact on nutrient absorption, inflammation, and systemic deterioration [24]. Incorporating the concept of AIF provides a stronger physiological rationale for our findings and highlights the importance of recognizing persistent EFI in septic shock. While SOFA score is widely used to evaluate organ dysfunction and predict outcomes in sepsis [25], it does not account for GI function. Given the growing evidence that GI dysfunction, including EFI, contributes to poor outcomes, complementing existing severity scores with direct GI assessment may offer a more comprehensive measure of organ failure in this population.
Unfortunately, a standardized definition of EFI has not yet been established, and the definitions employed in previous studies vary widely. These definitions are largely categorized into delayed gastric emptying, GI symptoms, and inadequate EN delivery [8]. Concerns have been raised regarding the reliability of gastric residual volume (GRV) monitoring as an indicator of delayed gastric emptying due to the lack of standardized measurement techniques, its poor predictive value for regurgitation or aspiration, the absence of established normal values, and the lack of a validated single cutoff value to define "large" GRV [26,27]. Furthermore, a systematic review concluded that GRV thresholds are not significantly associated with the prevalence of EFI, given the weak correlation between GRV and gastric emptying [28]. Inadequate EN delivery may be influenced by factors unrelated to GI function, such as interruptions for procedures, delayed initiation of EN, or clinician decisions to withhold EN. Also, there is no universally accepted threshold for adequate EN delivery.
Regarding mortality outcomes, total EFI days was identified as an independent risk factor for 90-day mortality, but not 30-day mortality after adjusting for confounding factors. This is likely because 30-day mortality is primarily determined by acute factors such as initial hemodynamic instability and multi-organ failure [29], whereas 90-day mortality is influenced by cumulative effects of EFI over time, particularly through its impact on nutritional status. Reduced volumes of EN delivered due to EFI can lead to malnutrition [30]. Malnutrition, in turn, can result in a wide range of adverse outcomes, including increased risk of pressure ulcers and impaired wound healing, immune suppression and higher infection rate, muscle wasting and functional loss increasing the risk of falls, and longer length of hospital stay [31]. These factors are likely linked to increased 90-day mortality.
Despite significant differences in clinical outcomes, total caloric intake across groups remained comparable. The presence and persistence of EFI may reflect more than just reduced energy delivery. It may be associated with impaired gut motility, increased mucosal permeability, and compromised nutrient absorption, all of which can promote systemic inflammation and contribute to multiple organ failure [32]. Through these mechanisms, EFI appears to influence clinical outcomes independently of the total volume of nutrition delivered. To accurately determine the isolated effect of enteral caloric intake on mortality outcomes, a fully standardized EF protocol is essential. However, such standardization was not feasible in our cohort due to the clinical characteristics of the population. Many patients were hemodynamically unstable, requiring individualized nutrition decisions based on ongoing clinical changes, with feeding initiation, rate of advancement, and formula selection determined through multidisciplinary discussion in real-world ICU practice.
This study has several limitations. First, it was conducted in a tertiary medical ICU with a high prevalence of complex comorbidities and immunocompromised patients, limiting generalizability. Second, the retrospective design introduced potential confounding and information bias. Third, GRV was not included in the definition of EFI due to growing evidence that GRV is a poor surrogate marker of gastric emptying and not consistently associated with feeding intolerance or clinical outcomes. This study focused on GI symptoms as the definition of EFI, as they are more easily measurable and applicable in routine clinical practice. This approach was chosen to enhance the practicality and consistency of EFI assessment in real-world settings. Fourth, the relatively small sample size limited this study’s statistical power and the ability to conduct robust subgroup analyses. Finally, as an observational study, patients were not managed under a uniform EN protocol, resulting in potential variation in formula composition, feeding methods and approach to advancing EN volume. Future large, prospective, multicenter studies with standardized EN protocols are warranted to validate these findings and further explore the impact of EFI on clinical outcomes.
In conclusion, 81.9% of septic shock patients admitted to the medical ICU experienced EFI. An increase in the total number of EFI days was significantly associated with higher 90-day mortality, even after adjusting for relevant clinical confounders. These findings highlight the critical role of GI dysfunction in septic shock and suggest that EFI may serve as a valuable prognostic marker. Further studies are needed to explore the role of EFI monitoring and to investigate its link to clinical outcomes among critically ill patients.
KEY MESSAGES
▪ A relatively high prevalence of enteral feeding intolerance (EFI) can be occurred among patients with septic shock admitted to the intensive care unit, compared to the other critically ill patients.
▪ An increase in the total number of EFI days was significantly associated with higher 90-day mortality, even after adjusting for relevant clinical confounders.
▪ These findings highlight the critical role of gastrointestinal dysfunction in septic shock and suggest that EFI may serve as a valuable prognostic marker.
XML Download