Enteral Nutrition Safety and Outcomes of Patients with COVID-19 on Continuous Infusion of Neuromuscular Blockers: A Retrospective Study

Background Intravenous infusions of neuromuscular blocking agents (NMBAs) and prone positioning are recommended for acute respiratory distress syndrome (ARDS) due to COVID-19. The safety of enteral nutrition (EN) during these treatments is unclear. This study assessed EN tolerance and safety during NMBA infusion in proned and nonproned patients with ARDS due to COVID-19. Methods This retrospective study evaluated patients who were admitted to a tertiary-care ICU between March and December 2020, had ARDS due to COVID-19, and received NMBA infusion. We assessed their EN data, gastrointestinal events, and clinical outcomes. The primary outcome was gastrointestinal intolerance, defined as a gastric residual volume (GRV) ≥500 ml or 200–500 ml with vomiting. We compared proned and nonproned patients. Results We studied 181 patients (mean age 61.2 ± 13.7 years, males 71.1%, and median body mass index 31.4 kg/m2). Most (63.5%) patients were proned, and 94.3% received EN in the first 48 hours of NMBA infusion at a median dose <10 kcal/kg/day. GRV was mostly below 100 ml. Gastrointestinal intolerance occurred in 6.1% of patients during NMBA infusion and 10.5% after NMBA discontinuation (similar rates in proned and nonproned patients). Patients who had gastrointestinal intolerance during NMBA infusion had a higher hospital mortality (90.9% versus 60.0%; p=0.05) and longer mechanical ventilation duration and ICU and hospital stays compared with those who did not. Conclusion In COVID-19 patients on NMBA infusion for ARDS, EN was provided early at low doses for most patients, and gastrointestinal intolerance was uncommon in proned and nonproned patients, occurred at a higher rate after discontinuing NMBAs and was associated with worse outcomes. Our study suggests that EN was tolerated and safe in this patient population.

Prolonged fasting has been associated with prolonged ICU stays, increased infectious complications, and higher mortality rate [13]. Hence, early enteral nutrition (EN), within 24-48 hours of ICU admission, has been recommended [14,15], may prevent malnutrition, preserve muscle mass, and reduce mortality [16,17]. Adequate EN is frequently hampered by gastrointestinal intolerance that includes vomiting, diarrhea, abdominal distension, constipation, and increased gastric residual volumes (GRVs) [18]. Tere are also safety concerns about EN during NMBA infusion and prone positioning [19,20]. NMBA infusion, which is often accompanied by the use of intravenous sedatives and vasopressors, may increase the risk of vomiting and ileus and hence gastrointestinal intolerance [19]. Prone positioning may also increase the risk of regurgitation and aspiration [19,20].
Te ideal approach when treating patients with COVID-19 receiving intravenous NMBA infusion while on mechanical ventilation is the concomitant use of EN. Whereas the Society of Critical Medicine made no recommendation regarding nutritional requirements specifc to patients receiving NMBA infusion [21], the European Society of Intensive Care Medicine suggested that EN should not be delayed merely because of NMBA infusion [22]. Due to the lack of quality evidence, the safety and tolerance of EN during NMBA infusion with and without prone positioning need further study [19]. We aimed to evaluate whether EN was tolerated and safe during NMBA infusion for critical COVID-19 with acute hypoxemic respiratory failure/ARDS.

Materials and Methods
Te STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines [23] were followed in reporting this study.

Study Design, Setting, and Participants.
Te study was a retrospective study that was approved by the Institutional Review Board of the Ministry of National Guard Health Afairs and was performed in accordance with its ethical standards. Te study was conducted in the adult ICUs of King Abdulaziz Medical City in Riyadh, Saudi Arabia. Te hospital was a tertiary-care center with >1000 beds. During the frst wave of the COVID-19 pandemic in the spring of 2020, four ICUs (ICU-A with 28 beds, ICU-B with 10 beds, ICU-C with 16 beds, and ICU-D with 16 beds) were designated to care for patients with severe COVID-19 [24]. Tese units functioned as closed ICUs with board-certifed ICU consultants leading the provision of medical care 24 hours/7 days a week [24]. We included all consecutive patients with severe COVID-19, confrmed by reverse transcription polymerase chain reaction, who were admitted to the ICUs between March 1, 2020, and December 31, 2020, received invasive mechanical ventilation, and had continuous NMBA infusion (cisatracurium besylate) for at least 6 hours as part of the management of moderate-severe ARDS. We excluded patients who did not receive any EN during the NMBA infusion.
In this study, treatments with NMBA infusion and/or prone positioning in our ICU were at the discretion of the treating ICU team. Patients with ARDS and persistent hypoxemia (ratio of arterial oxygen partial pressure to fractional inspired oxygen <150) after intubation were usually given NMBA infusion (cisatracurium besylate infusion at 0.5-3 mcg/kg/min with dose titrated to achieve a train-of-four of 2/4 on peripheral nerve stimulation) in addition to intravenous sedative (propofol and/or midazolam) and narcotic (fentanyl) infusion. If hypoxemia persisted, prone positioning was performed, according to an ICU protocol. Te head of the patient was kept elevated at 20-30 degrees in the prone position. Te treating ICU team decided on starting EN, including the initial hourly volume. Te ICU dietician selected the EN formula and determined the caloric goal and hourly EN volume, usually after discussion with the ICU team, especially for proned patients. Te ICU nurses used an evidence-based EN protocol [15,25] and routinely checked GRV every 4 hours by aspirating the feeding tube using a syringe and recorded vomiting episodes and bowel movements.

Data Collection.
Te collected data included demographics, comorbid conditions, date of intubation, date of initiation and discontinuation of NMBA infusion, date of initiation of EN, EN formula (energy-dense versus regular formula), total volume of delivered EN per day for up to the frst 7 days, and use of parenteral nutrition. We defned early EN as EN that started within the frst 24 to 48 hours of NMBA infusion. We calculated the hourly and daily caloric intake based on the delivered volume and the EN formula.
To assess EN safety and tolerance, we also collected data on frank aspiration incidents, diagnosis of ICU-acquired pneumonia based on the presence of a bacterial culture from deep tracheal aspirate with clinical features pneumonia (new infltrates on chest X-ray) during ICU stay, vomiting episodes, bowel movements, and GRVs (per check and per 24 hours) during NMBA infusion and in the 48 hours after stopping NMBAs and use of prokinetics (metoclopramide and erythromycin) for gastrointestinal intolerance. In this study, the primary outcome was gastrointestinal intolerance. As there is no agreement on its defnition and GRV is frequently considered its surrogate [18], we defned gastrointestinal intolerance as having a GRV ≥500 ml on a single check or 200-500 ml with vomiting [15,18]. Diarrhea was defned as having ≥3 loose or liquid bowel movements per day [26]. We also noted tracheostomy events, duration of mechanical ventilation, length of stay in the ICU and hospital, and vital status at ICU and hospital discharge (secondary outcomes).

Statistical Analysis.
Te study patients were categorized into two groups depending on whether they received prone positioning. Te results were presented as frequency and percentage for categorical data and mean and standard deviation or median and interquartile range (IQR) for continuous data. Te Chi square or Fisher's exact test was used to compare categorical variables and Student's t-test or Mann-Whitney U test to compare continuous variables. As GRV might be proportional to the delivered EN volume, we correlated GVR per day with the delivered EN volume per day using the Pearson correlation and reported Pearson's r. We used Statistical Package for the Social Sciences version 21 (SPSS) for all statistical analyses. All statistical tests were considered signifcant at p value <0.05.  Table 1. Te baseline characteristics of proned and nonproned patients were similar, except for proned patients having a lower body mass index and white blood cell count and a higher hemoglobin level. Most (76.8%) patients received systemic corticosteroids as part of COVID-19 management without signifcant diferences between proned and nonproned patients (p � 0.33). Tere was no signifcant diference in the use of energydense EN formulas between proned and nonproned patients.

Enteral Nutrition
Te caloric intake was lowest on the frst day and gradually increased but remained moderate for both proned and nonproned patients ( Table 2 and Figure 1). For nonproned patients, the median caloric intake per kg per day was 0 (IQR: 0-9.0) on the day before NMBA infusion, 1.  Figure 1). Tere were no signifcant diferences in the caloric intake between proned and nonproned patients on any of these days. None of the patients received parenteral nutrition.
Gastrointestinal intolerance while receiving NMBA infusion occurred in only 11/181 (6.1%, 95% confdence interval: 3.1-10.6%) patients: 2.9% of patients on the day before, 0.7% on day 1, 3.6% on day 2, 0% on day 3, and 3.3% on day 4 of NMBAs ( Figure 2). Tere were no signifcant diferences in the rates of gastrointestinal intolerance between proned and nonproned patients on these days. Gastrointestinal intolerance was more common after discontinuing NMBA and occurred in 19/181 (10.5%) patients: 6.1% of patients on day 1 and 6.7% on day 2 after discontinuation of NMBAs ( Figure 2). Patients who were proned had higher rates of gastrointestinal intolerance on day 1 post-NMB(p � 0.06). Te correlation between total GRV per day and EN volume per day was signifcant only on the frst day (r � 0.435, p < 0.001) and fourth day (r � 0.472, p < 0.001), but not on the second and third days.
Vomiting episodes were rare during NMBA infusion, occurring in 7 patients (3 proned patients and 3 unproned patients on day 3 and 1 proned patient on day 4). Metoclopramide was used in 68 (37.6%) patients: 33.0% of proned and 46.2% of nonproned patients (p � 0.08). Only one patient received the combination of metoclopramide and erythromycin as prokinetic agents. None of the patients in the study had a documented aspiration event. Diarrhea was more common, occurring in 3.9-10.5% of patients per day during NMBA infusion (Table 2). Table 3. Out of the 181 patients, 58 (32.0%) developed ICUacquired pneumonia with no signifcant diference among proned and nonproned patients. ICU-acquired pneumonia occurred in 5/11 (45.5%) patients who had any gastrointestinal intolerance while on NMBA infusion compared with 53/170 (31.2%) patients who did not have any gastrointestinal intolerance (p � 0.33). Only 35 (19.4%) of the patients underwent tracheostomy with no diference between proned and nonproned patients. Also, the tracheostomy rate was similar between those who had any gastrointestinal intolerance and those who did not while on NMBA infusion (18.2% versus 20.0%, p � 1.0). As for the duration of mechanical ventilation and ICU and hospital lengths of stay, they were similar in proned and nonproned patients ( Table  Te ICU and hospital mortality rates were similar in proned and nonproned patients (

Discussion
In this study, we evaluated the safety and tolerance of EN during NMBA infusion with and without prone positioning in moderate-severe ARDS due to COVID-19. We found that EN was provided early at low doses for most patients; gastrointestinal intolerance was uncommon during NMBA infusion in proned and nonproned patients; gastrointestinal intolerance occurred at a higher rate after discontinuing NMBAs; and gastrointestinal intolerance was associated with a worse outcome.
In the current study, early EN was provided at low doses (<10 kcal/kg/day during NMBA infusion and around 15 kcal/kg/day after NMBA discontinuation) for most patients while on NMBA infusion with and without prone positioning and with frequent use of energy-dense EN formulas. Using propensity score matching, a retrospective study showed that early EN in patients receiving NMBA infusion was safe [27]. Early EN has been recommended in critical COVID-19 patients at trophic or hypocaloric doses with advancing EN dose slowly as tolerated over the frst week of critical illness to meet the energy goal of 15-20 kcal/kg actual body weight [28]. Tis recommendation included proned patients [28]. Te EN dose in our study was in line with the clinical practice guidelines [28]. However, a multicenter study from Singapore observed that the minimum caloric of 15 kcal/kg was achieved in only 39/ 74 (54%) patients with critical COVID-19 [29]. Energydense EN formulas have been suggested to safely deliver more calories, especially during prone positioning [19]. However, recent data suggested that energy-dense feeding may increase GRV and delay gastric emptying [30].
In our study, gastrointestinal intolerance was uncommon during NMBA infusion occurring in only 6.1% of patients. Tere was also no recorded aspirational event. Gastrointestinal intolerance is common in critically ill patients, with a systematic review of 72 studies estimating its prevalence at 38% (95% confdence interval: 31-46%) [31].   Figure 1: Boxplots of the caloric intake in proned and nonproned patients with COVID-19 on intravenous infusion of neuromuscular blockers for acute respiratory distress syndrome. Te boxplots show the 25th, 50th, and 75th percentiles (box), 10th and 90th percentiles (whiskers), and outliers.  In a retrospective study of 52 critically ill patients with COVID-19 not receiving NMBA, gastrointestinal intolerance occurred in 18 patients (32.4%) within the frst 7 days [32]. Whether NMBA infusion increases the risk of gastrointestinal intolerance in ARDS is unclear. Gastrointestinal intolerance was not studied in the randomized controlled trials that evaluated NMBA infusion in ARDS [33,34]. One prospective study found similar GRVs in 20 patients while receiving opioid infusion versus opioid and cisatracurium infusion (95 ± 76 ml and 105 ± 90 ml two hours after 200 ml of tube feeding, respectively) [35]. Another study found that 10/47 patients (21%) admitted to a trauma ICU had gastrointestinal intolerance to EN during NMBA infusion for ≥48 hours [36]. Besides being tolerated, our fndings suggested that EN during NMBA infusion was safe. Te clinical practice guidelines of the European Society of Intensive Care Medicine suggested that EN should not be delayed merely because of NMBA infusion [22]. Whether prone positioning, which is a common concomitant treatment for patients with ARDS, increases the risk of gastrointestinal intolerance is also unclear. We found no diference in gastrointestinal intolerance between proned and nonproned patients receiving NMBA infusion. A systematic review of studies that assessed early EN during prone positioning in patients receiving invasive mechanical ventilation found that fve out of six studies reported no diferences in GRV between supine and prone positions [37]. One study reported a higher rate of the need to stop EN while in the prone position [38], while another did not [37,39]. Vomiting episodes were more common in the prone position when EN was administered over 18 hours rather than 24 hours [37,40]. Te rates of ventilator-associated pneumonia, lengths of stay, and mortality were similar between supine and prone positions [37]. One prospective comparative study showed lower mortality in patients receiving a 24-hour versus 18-hour administration protocol [37,40]. Additional studies were recently reported on COVID-19 patients. In a retrospective study in patients with ARDS due to COVID-19, gastrointestinal intolerance was observed in 30.8% of 57 patients in the prone position and 23.2% of 69 patients in the supine position (p � 0.81) [41]. In another multicenter retrospective study conducted in 83 critically ill patients with COVID-19, the prone position was not associated with a higher rate of high GRV (≥250 mL) [29]. Our study and the other studies indicate EN is well tolerated in patients with ARDS while in the prone position. Te lower rate of gastrointestinal intolerance in our study compared to other studies could be related to the duration of assessment (up to 4 days on NMBA infusion) and the moderate hourly EN volume, which may correlate with GRV. Additionally, head-of-bed tilting to 20-30 degrees, administering EN over 24-hour rather than shorter durations, employing an EN protocol, using prokinetics, and close monitoring may have contributed to the low rate of intolerance in our study and are recommended when providing EN while in the prone position [19,28]. Whether postpyloric EN during the prone position and intravenous NMBA infusion would reduce gastrointestinal intolerance is unclear and needs further studies [37]. In our study, neither early versus late EN nor prone versus supine position was associated with mortality. A meta-analysis of four studies in critically ill patients with COVID-19 found that early EN was signifcantly associated with lower mortality risk (risk ratio: 0.89, 95% confdence interval: 0.79-1.00, p � 0.05) [42]. Prone positioning reduces mortality in moderate-severe ARDS in non-COVID-19 patients [43], but its efectiveness in intubated patients with COVID-19 is not clear [44]. In the current study, gastrointestinal intolerance was associated with worse outcomes (longer duration of mechanical ventilation, longer stay in the ICU and hospital, and higher mortality). Tere was no documented aspiration event associated with gastrointestinal intolerance events to explain worse outcomes. Unwitnessed or microscopic aspiration may have occurred. Besides, gastrointestinal intolerance may be a manifestation of acute gastrointestinal injury, which is a manifestation of multiorgan dysfunction [45].
Te study's fndings should be interpreted taking into consideration its strengths and limitations. Te sample size is larger than that of most published studies on this topic, and we collected detailed data on EN and gastrointestinal intolerance. On the other hand, our study is a singlecentered retrospective study, which limits the generalizability of our fndings and questions causality in the reported associations. Moreover, our fndings may be the result of unmeasured confounders. Te low rate of gastrointestinal intolerance prevented us from performing a multivariable regression analysis to evaluate its predictors. We also did not obtain additional data on protein intake and whether the feeding tube was intragastric or postpyloric.

Conclusions
In this study of COVID-19 patients on NMBA infusion for ARDS, EN was provided early at low doses for most patients, gastrointestinal intolerance was uncommon in proned and nonproned patients, occurred at a higher rate after discontinuing NMBAs, and was associated with worse outcome. Our study suggests that EN given at low doses with the use of a 24-hour administration protocol was tolerated and safe in this patient population. Tese fndings were similar in proned and nonproned patients. Te optimal dose and volume of EN in patients treated with NMBA infusion and/ or prone positioning for ARDS need further evaluation.

Data Availability
Te data presented in this study are available upon request from the corresponding author. Te data are not publicly available due to institutional policies of maintaining the confdentiality of patient data.

Ethical Approval
Tis study was approved by the Institutional Review Board of King Abdullah International Medical Research Center, Riyadh, Saudi Arabia (IRBC/01003/21, 31 May 2021).

Consent
Informed consent was waived by the Institutional Review Board as the study used only electronic medical record review for data collection.

Conflicts of Interest
Te authors declare that there are no conficts of interest.