Abstract
For mechanically ventilated patients undergoing surgery, interrupting enteral feeding to prevent pulmonary aspiration is common; however, there are no published preoperative fasting guidelines for these patients, resulting in fasting practices that often vary greatly between hospitals. This retrospective study described fasting practices and surgical outcomes of mechanically ventilated patients across five trauma centres. The primary exposure was hours nil per os before surgery and was stratified into short (<6h) and moderate (≥6h) fasting duration. Shared frailty models assessed the relationship between time to perioperative complication and nil per os category. Three of the five hospitals had preoperative fasting guidelines, and those most compliant required patients to be fed up until surgery. Most patients were fasted ≥6h prior to surgery and no increased risk of complication was found for patients who were fasted <6h. Future studies are needed to establish appropriate preoperative fasting thresholds for mechanically ventilated patients.
Introduction
Evidence of enteral nutrition (EN) support for surgical patients has been well established with several randomised trials demonstrating the benefit for decreasing in-hospital complications, morbidity and overall hospital length of stay (Charoenkwan & Matovinovic 2014, Dag et al 2011, Drover et al 2011, Lewis et al 2009, Smedley et al 2004). Still, patients in the ICU often receive less calories than needed due to underordering of nutritional therapy by clinicians, inadequate delivery, or interruptions in feeding (Naber et al 1997, Segaran et al 2016, Yip et al 2014). Interruptions to feeding in the surgical critical care setting can arise due to a perceived risk for pulmonary aspiration and related perioperative complications that tend to arise during surgical procedures or diagnostic tests (Czapran et al 2015, Heyland et al 2011, Kozeniecki & Fritzshall 2015, Lee et al 2016, McClave et al 2016, Passier et al 2013, Peev et al 2016). Although previous research has demonstrated that pulmonary aspiration is a rare complication, surgical patients are frequently kept nil per os (NPO) ≥6h over the perceived risk of aspiration (Crenshaw, 2011, McClave et al 1999, 2016, Smith et al 2011, Taylor et al 2016).
Published guidelines by the American Society for Parenteral and Enteral Nutrition and the Society of Critical Care Medicine for nutritional support of critically ill adult patients recommend assessing the risk of aspiration based on specific risk factors (mechanical ventilation, age >70 years, supine position, neurologic deficits), as well as improving management through implementation of standardised protocols (Davidson et al 2007, McClave et al 2016). In 2017, the American Society of Anesthesiologists (ASA) published guidelines that reported fasting for healthy patients up to 6h prior to surgery is safe (The American Society of Anesthesiologists 2017); however, there still remains a wide range of preoperative fasting practices across studies (6.5–13.5h) (Passier et al 2013, Peev et al 2016) with no specific threshold for mechanically ventilated (trauma) ICU patients.
Accordingly, the primary aims of this study were to (1) describe practices in preoperative fasting for mechanically ventilated, trauma SICU patients receiving EN across five hospitals; (2) stratify patients into preoperative fasting times; and (3) describe patient characteristics and outcomes by preoperative fasting categories.
Methods
This retrospective study included adult trauma patients (≥18 years) admitted to the SICU between 1 June 2016 and 31 December 2017 at five Level I trauma centres. Patients were included if they had (1) a cuffed endotracheal tube or cuffed tracheostomy; (2) received EN via an orogastric tube, nasogastric tube, feeding gastrostomy or jejunostomy prior to operative procedure; and (3) had a traumatic injury diagnosis that required operative management during their SICU stay. To reduce variability in surgical practices among hospitals, patients were excluded if they had any abdominal surgery. They were also excluded if they had a surgery that required the patient to be in the prone position, had a manipulation of the airway, their EN was started after their operative procedure, or their only procedure was a tracheostomy or percutaneous endoscopic gastrostomy placed at the bedside. There was no sample size calculated because this was a preliminary exploratory study. The corresponding Institutional Review Board at each hospital approved the study. Consent was waived due to the retrospective nature of the study. The manuscript adheres to the applicable STROBE guidelines for observational studies.
Primary data originated from each centre’s trauma registry and included patient demographics, clinical characteristics and study outcomes. Variables of interest included age, sex, comorbidities, injury cause, procedure codes, and mechanical ventilator use and duration. Study coordinators at each centre retrospectively reviewed patient charts to ensure study inclusion or exclusion criteria were met and to abstract additional data. Abstracted data included Acute Physiology and Chronic Health Evaluation (APACHE) II score, nutrition status, and characteristics of the EN delivery. Nutrition status was quantified using the Nutrition Risk in the Critically Ill (NUTRIC) score which serves as a validated measure of nutritional insufficiency for ICU patients (Koekkoek & Van Zanten 2018). Characteristics of EN delivery included type of feeding tube, and if applicable, compliance with the hospital’s preoperative fasting guideline. Guidelines on EN exist at three of the five hospitals and serve as recommendations to assist with the management of feeding in critically ill patients that do not supersede clinical judgement.
The primary exposure of interest was hours NPO and was calculated for each interval spent in the SICU before or after patients were taken to the operating room. Patients were then stratified into two categories based on current ASA guidelines for elective surgery patients (The American Society of Anesthesiologists 2017). Patients with less than 6h of NPO time were defined as having a short fasting time, while fasting ≥6h was considered moderate fasting time.
Perioperative complications included the following: aspiration, infection, wound dehiscence, pressure ulcer and in-hospital mortality. Aspiration was identified in the progress notes and the discharge summary. Infection was also identified from progress notes and the discharge summary, and included sepsis, pneumonia (ventilator-associated, community-acquired, hospital-acquired) (Taylor et al 2013), urinary tract infection and other (Staphylococcus aureus, Escherichia coli, Candida, abscesses, Clostridioides difficile, skin infection, etc). Wound dehiscence and pressure ulcers were identified using wound care team documentation. In-hospital mortality was identified from the trauma registry.
Descriptive statistics were analysed using Fisher’s exact tests for categorical variables, and Student’s t-tests, one-way ANOVAs, Wilcoxon two-sample tests, or Kruskal–Wallis tests for continuous variables. Data are displayed as means and standard deviations (SDs), medians and interquartile ranges (IQRs), or proportions, as appropriate. Shared frailty models were constructed to assess the relationship between time to perioperative complications and NPO time category, accounting for patient clustering by facility. Follow-up time was equal to the sum of all complication-free days spent in the SICU. Patients who did not experience a complication were statistically censored at the end of their SICU stay. For these analyses, the moderate fasting category was used as the reference variable. Variables selected for adjustment had an entry criterion of ≤0.1 with the primary outcome and were each manually entered into the model. SAS 9.4 (Cary, NC) was used for all analyses. Two-tailed tests with an alpha of 0.05 were used for all tests.
Results
Seventy-five mechanically ventilated SICU patients had their data collected across the study period. Eleven patients were excluded due to inadequate documentation of fasting information, leaving 64 patients for analysis. On average, patients were 48 years old (SD: 17.8) and primarily male (81%). The mean APACHE II score was 20 (SD: 8.7) and the mean NUTRIC score was 3.4 (SD: 1.5), representing low risk for malnutrition.
Table 1 provides details on the preoperative fasting guidelines, average fasting time, and compliance to fasting guidelines prior to surgery at each of the five participating centres. Guidelines at Hospitals 1 and 3 permit patients to continue their feeds until being called to the OR. Hospital 2 requires patients to be NPO for a minimum of 4h prior to their operative procedures. Hospital 4 does not have formal protocol, but patients are usually NPO 8h prior to surgery, while Hospital 5 does not have a formal protocol. Interestingly, even with a ‘continued’ protocol, Hospital 1 had a longer median fasting time compared to Hospital 4 (no official protocol) (9.6 (2.4–13.8) versus 8.8 (6.1–9.5) h, respectively), yet Hospitals 1 (57%) and 3 (54%) also had some of the highest compliances in the study, followed by Hospital 4 (43%), and then Hospital 2 (2%).
Preoperative fasting guidelines across trauma centres
IQR: interquartile range; N/A: not applicable; NPO: nil per os; OR: operating room.
aCompliance calculated by the total number of compliant procedures over the total number of procedures performed per site.
Overall, patients were fasting for a median of 10.0 (IQR: 7.5–12.5) h prior to surgery. After stratifying by short and moderate fasting times, 13 (20%) were in the short fasting category, and 51 (80%) were in the moderate fasting category. Patients in the short fasting category fasted for a median of 1.6 (0.3–4.7) h, while patients in the moderate category fasted for 10.5 (9.5–13.5) h.
There were no differences in demographics and the majority of injury characteristics between groups (Table 2). However, compared to patients in the moderate fasting category, patients in the short fasting category had a significantly lower median (IQR) admission GCS (3 (3–7.5) versus 7 (3–15), p = 0.03), fewer EN interruptions (1 (IQR: 1–3) versus 4 (IQR: 2–6), p = 0.001), and also had more neurosurgical procedures (46% versus 8%, p = 0.003).
Patient characteristics and outcomes by fasting category
APACHE: Acute Physiology and Chronic Health Evaluation; EN: enteral nutrition; GCS: Glasgow coma scale; IQR: interquartile range; LOS: length of stay; NPO: nil per os; NUTRIC: Nutrition Risk in Critically Ill; OR: operating room; SD: standard deviation; SICU: surgical intensive care unit.
aIncluded in the shared frailty model.Bold p-values indicate statistical significance.
Perioperative complications were similar between groups. Patients in the short fasting category showed the lowest rate of delirium, compared to the moderate category (8% versus 52%, p = 0.004) but there were no differences observed in infection, aspiration, hospital LOS, SICU LOS, or mortality between categories (Table 2).
Table 3 shows unadjusted patient risk factors between those experiencing a perioperative complication and those without. Compared to those without a perioperative complication, a significantly greater proportion of patients with a perioperative complication had more time on mechanical ventilation (64% versus 36%, p = 0.03), spent more time in the SICU (20 (15–25) versus 15 (11.5–18.5), p = 0.01), and had more EN interruptions (4 (2–7) versus 2 (1–4), p = 0.02). Patients with >1 surgical procedure and mean APACHE II score were trending towards but did not reach statistical significance.
Unadjusted modelling of patient risk factors by complication status
APACHE: Acute Physiology and Chronic Health Evaluation; EN: enteral nutrition; GCS: Glasgow coma scale; HLOS: hospital length of stay; IQR: interquartile range; LOS: length of stay; NPO: nil per os; NUTRIC: Nutrition Risk in Critically Ill; Preop: preoperative; SD: standard deviation; SICU: surgical intensive care unit.
aPatients who were statistically censored in analysis.
bRepresents median cut-off for days on mechanical ventilation.Bold p-values indicate statistical significance.
Table 4 outlines results from the multivariable frailty model, examining the relationship between time to the development of a perioperative complication in the SICU and fasting category. Other covariates that were included for adjustment were number of surgical procedures, APACHE II score, and days on mechanical ventilation. Fewer surgical procedures (1 versus >1) (HR 0.43 (0.19–0.97), p = 0.04) and fewer days (≤12 versus >12 days) on mechanical ventilation (HR 0.38 (0.16–0.89), p = 0.03) were independently associated with fewer perioperative complications. Compared to patients in the moderate fasting category, those in the short fasting category had a non-significant, 27% reduced risk of perioperative complications (HR 0.73 (95% 0.26–2.0), p = 0.54) during their SICU stay.
Time to perioperative complication by fasting categories
NPO: nil per os.
aClustered by facility and adjusted for number of surgical procedures (1 versus >1), APACHE II score, and days on mechanical ventilation (≤12 versus >12 days).
bCI, 95% Wald confidence limits.
Discussion
Because no preoperative fasting guidelines exist for mechanically ventilated patients, there are wide variations in practices across facilities and further, it can be challenging to identify patients at risk of perioperative complications. Our multicentre exploratory study suggests that overall, hospitals did not adhere to their established guidelines and the majority of patients (80%) fell into the moderate (≥6h) preoperative fasting category. We found that patients in the short fasting category were more likely to arrive with a poor GCS score, while patients in the moderate fasting category were more likely to develop delirium. Further, earlier timing of NPO status (<6h prior to surgery) did not appear to have an increased risk of perioperative complication, although later timing of NPO did not appear to affect nutrition. Both malnutrition and perioperative complications should be considered together to ensure that there is a balance between the perceived risks of aspiration and malnutrition. Larger studies examining this topic might have the ability to generate important data that can help guide perioperative management strategies in mechanically ventilated patients.
To our knowledge, this is one of the first studies to stratify patients by NPO time and describe characteristics and outcomes between these categories. We found that significantly more patients in the short fasting category received a neurosurgical procedure compared to the moderate category, who primarily had orthopaedic procedures. Patients in the short fasting category arrived with poor admission GCS scores and may have required rapid neurosurgical intervention. As such, this category may have benefited not necessarily from a short fasting time, but rather, an expedited time to surgery. Furthermore, patients in the short fasting category also had a significantly lower proportion of delirium, compared to the moderate category. Delirium frequently occurs in older adult patients, or those going through drug or alcohol withdrawal (Inouye 1994). Delirium was significantly higher in the moderate fasting category, which may have introduced challenges in EN management; the data suggest that EN was also interrupted more often, potentially leading to less caloric intake. Though not significant, these patients also developed more infections.
These findings are consistent with some of those reported in the published literature (Angotti et al 2018, Douglas & Ciraulo 2017, Falconer et al 2014, Pham et al 2018, Segaran et al 2016). Douglas and Ciraulo (2017) explored variability in fasting practices among intubated patients and found that for non-emergent operations, the majority (82%) of patients fasted for >6h, with a large concentration (50%) falling into the 9–13h fasting range. Similarly to our study, emergent patients were typically (91%) fasted for <6h.
There are significant discrepancies in the fasting practices for SICU patients. Even with established hospital guidelines, it appears that clinicians continue to overvalue the perceived risk of aspiration and undervalue the risk of malnutrition. Research on preoperative fasting protocols suggests continued EN (up until the patient is transferred to OR) as a means to reduce prolonged fasting, yet the reported time for NPO in the ‘continued’ group can reach up to 9.5h (Angotti et al 2018, Pham et al 2018). We also observed variation in fasting time among both hospitals with guidelines that approved a ‘continued feeding’ protocol; however, these protocols may have at least helped the hospitals maintain a higher compliance on average, across multiple patient surgeries, as well as to reduce interruptions to EN. Despite the fact that a dietician is consulted on each EN case, if the OR has an unexpected opening, services may stop the EN even earlier than recommended. Some of these discrepancies could potentially be due to lack of adequate education and communication among clinical providers. In hospitals hoping to reduce their preoperative fasting times, implementing a continued feeding protocol may help them move the needle forward.
The first and primary limitation of this study is the small sample size; however, this was an exploratory study and the results reported here regarding how preoperative fasting characteristics affect surgical outcomes, should be taken as suggestive only and will serve to direct future studies capable of attaining larger samples. Nonetheless, the inclusion of patients across five Level I trauma centres in the United States allowed for an examination of differences in preoperative fasting practices across a diversity of centres and patients by region. Second, findings were dependent on accurate EN documentation which was limited in a number of cases. Thus, we were unable to determine if there were any caloric deficits that resulted from one fasting cut-off over another. However, we were able to calculate the NUTRIC score across all patients, which on average, showed a low risk of malnutrition. Third, due to the retrospective nature of the study, we were unable to determine causation and were significantly underpowered to detect any causality between (perioperative) complications and specific NPO fasting duration.
Conclusions
In a population largely underrepresented in the published literature, we found that fasting mechanically ventilated trauma patients <6h prior to surgery did not introduce additional risk. Additionally, most of the patients continue to be fasted ≥6h despite guidelines to the contrary. Stratifying patients by fasting category may help hospitals better understand how fasting practices are managed and which patients are at highest risk of perioperative complications. Larger prospective studies are needed to determine if an appropriate fasting threshold exists for mechanically ventilated trauma SICU patients.
Key phrases
Preoperative fasting practices of mechanically ventilated trauma patients are underreported in the published literature and are needed to improve understanding and management of these patients.
Although shorter fasting (<6h) periods have been shown to be effective without introducing additional risk, most patients in the study were still kept NPO ≥6h, because clinicians continue to overvalue the perceived risk of aspiration and undervalue the risk of malnutrition.
The current study did not identify an increased risk of perioperative complication for patients fasted <6h.
Across the surgeries examined, compliance with fasting guidelines was highest for hospitals with a continued feeding protocol, though on average, was generally low across facilities.
Declarations
Availability of data and materials
Data are available from the corresponding author upon reasonable request.
Competing Interests
None declared.
Funding
None declared. Internal funding was provided by all affiliated institutions.
Ethics approval
Ethical approval in accordance with the Declaration of Helsinki was obtained from the institutional review board at all participating sites before data collection (HCA-IRBs # 1064115 and 18-006, CHI-IRB # 1076488, Western-IRB # 20180181). Consent was waived due to the retrospective nature of the study.
Guarantor
DBO is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Contributorship
All authors provided final approval of the submitted manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. RB is responsible for study conception, interpretation of the data, and critical revisions. CM is responsible for literature review, design, data acquisition, data analysis and interpretation, and drafting the manuscript. JL, EM, KB, RM, AT, ML, PH, and DBO are responsible for interpretation of the data and critical revisions.
Footnotes
Acknowledgements
The authors wish to thank Kimberly Aumann; Carolyn Blue, BSN, CNS, MPH-C; Brenda Kuiphoff, RN, BSN; Jennifer Pekarek, RN, BSN, CCRP; Shenequa Deas, MPH, COMR, CCRC; Diane Redmond, MSN, RN, CCRP; Kathy Rodkey, RCIS, CCRC; and Jamie Shaddix, RN for performing the chart reviews at each institution.
