Risks in Management of Enteral Nutrition in Intensive Care Units: A Literature Review and Narrative Synthesis

Critically ill patients in the intensive care unit (ICU) have a high risk of developing malnutrition, and this is associated with poorer clinical outcomes. In clinical practice, nutrition, including enteral nutrition (EN), is often not prioritized. Resulting from this, risks and safety issues for patients and healthcare professionals can emerge. The aim of this literature review, inspired by the Rapid Review Guidebook by Dobbins, 2017, was to identify risks and safety issues for patient safety in the management of EN in critically ill patients in the ICU. Three databases were used to identify studies between 2009 and 2020. We assessed 3495 studies for eligibility and included 62 in our narrative synthesis. Several risks and problems were identified: No use of clinical assessment or screening nutrition assessment, inadequate tube management, missing energy target, missing a nutritionist, bad hygiene and handling, wrong time management and speed, nutritional interruptions, wrong body position, gastrointestinal complication and infections, missing or not using guidelines, understaffing, and lack of education. Raising awareness of these risks is a central aspect in patient safety in ICU. Clinical experts can use a checklist with 12 identified top risks and the recommendations drawn up to carry out their own risk analysis in clinical practice.


Introduction
In intensive care units (ICUs) critically ill patients have a high risk of developing malnutrition which is associated with a poorer clinical outcome [1]. Therefore, enteral nutrition (EN) has become an increasingly important research topic in recent years. In 2017, Arabi et al. (2017) reported that the intensive care medicine research agenda in nutrition and metabolism includes topics like optimal protein dose combined with standardized active and passive mobilization during the acute and post-acute phases of critical illness, nutritional assessment and nutritional strategies in critically obese patients. Moreover, the effects of continuous versus intermittent EN were classified as a hot topic [2]. The European Society for Clinical Nutrition and Metabolism (ESPEN), established standard operating procedures (SOPs) and guidelines for the provision of the best nutritional therapy for critically ill patients. They are regularly updated and cover various aspects of medical Nutrients 2021, 13, 82 2 of 31 nutritional therapy such as duration, timing, vulnerable patient populations with, for example, dysphagia or frailty and provide clinicians with practical procedures [3]. However, in clinical practice, EN is often not given highest priority due to other symptomatic problems such as cardiovascular status or the need of ventilation. Moreover, ICU patients are mostly very heterogeneous in terms of their illness, resulting in multiple risks and safety issues for patients. The literature offers possible solutions to this problem, but many unresolved questions still cannot be answered conclusively. However, there is some degree of consensus. For ventilated patients in ICUs, if possible, it is crucial to prefer EN over parenteral nutrition (PN) [3]. Other topics like the microbiome or recommendations on additives such as micronutrients and vitamins are still being discussed, for example, optimal vitamin D levels. Many critically ill patients suffer from vitamin D deficiency (serum 25-hydroxyvitamin D [25(OH)D] < 20 ng/mL), with levels lower than 12 ng/dL [4].
In addition to the clinical impact there are further uncertainties that affect the outcome of patients. Overall, human errors are the third leading cause why patients die in a hospital [5]. ICUs are in general a high-risk area, where critical ill patients receive a highly sophisticated care. Patients receive a lot of drugs; medical devices are used for administration of drugs and for ventilation. To increase patient safety, clinical risk management focuses on improving the quality and safety of health care services by identifying the circumstances and opportunities that expose patients to risk of harm, and on acting to prevent or control those risks [6]. In fact, clinical risk management is a key element of clinical governance and management [7]. Risks must be handled appropriately with a bundle of measures and they need to be audited regularly to ascertain whether risk-reducing measures are being used sufficiently [8]. If risks are unknown, a risk management audit is needed to identify potential risks. Especially in EN, there are several risks such as a failure in reaching the nutrition target (CIRSmedical ® no. 26 (V1)). The aforementioned error can lead to hypoglycemia and malnutrition [9]. Furthermore, EN and aspiration [10], which can occur when a patient was not placed in a head-up position or when post-pyloric feeding (nasojejunal tube) was not administered to patients with a high risk of aspiration [11]. On top, the worldwide introduction of the DIN EN ISO 80369-3 and the use of ENFit™-technology was a decisive step to increase patient safety in EN and can be seen as an important contribution in terms of risk management. However, the implementation of this standard still has shortcomings as not all security gaps were eliminated. It is still possible that enteral medication in Luer injections or tube feeding can be administered intravenously [12]. In addition to ICU-related risks there are further general risks such as an incorrect patient identification, nosocomial infections, medication errors, overlooked allergies, insufficient pain management, and failures in communication and documentation or failures in handling medical devices [6]. A consequent risk management in ICUs can help to increase patient safety; therefore, the aim of this rapid literature review was to identify risks for patient safety in the management and handling of EN in critically ill patients in ICU.

Materials and Methods
We conducted a literature review guided by the Rapid Review Guidebook by Dobbins (2017) [13] and the PRISMA Checklist [14]. As the first step in the EN risk identification process, we (MH, CMS, SF, CS, EL) used feedback from ICU staff (physicians, nurses, dieticians), reports from the Critical Incident Reporting System (CIRS) of the University Hospital of Graz, Austria, CIRSmedical ® and, online, the "berrypicking" method [15]. We structured these risks in a process: admission, prescription, verification, preparation, administration, monitoring, discharge, and general risks in EN in the ICU. We then systematically identified and screened the literature for these risks.

Search Strategy
Keywords, related MeSH terms and Boolean operators (AND, OR) were used. The main terms in the first search were: enteral nutrition, risks and intensive care and the identified Nutrients 2021, 13, 82 3 of 31 risks. For example, "enteral nutrition" AND "risks" OR safety AND "intensive care unit" AND "aspiration". At least two different terms were used for each risk. The primary search was limited to systematic reviews, meta-analyses, and guidelines. Next, studies, including randomized controlled trials (RCTs), retrospective studies and prospective observational studies were included to detect possible risks in EN in the ICU. The search was done with a language restriction to German and English and available full-text articles. Three different scientific databases, PubMed, Cochrane Library and Web of Science, were used to identify studies between 2009 and 2020.
Four independent reviewers (CMS, SF, CS, EL) searched in scientific databases between June and August 2020. The search was always carried out by two independent researchers in order to reduce a possible selection bias. The risks and the related literature were reviewed by independent researchers (CMS, SF, CS, EL and senior researcher (MMJ) with research experience >10 years). After the decision for inclusion of studies, the results were viewed and discussed by two researchers (MH, CMS). All identified risks were thematically summarized in a narrative synthesis by one senior researcher (MMJ) and checked by another researcher (MH). Data were prepared for a summary in a narrative synthesis and a checklist.

Results
In total we found 20 risks. We included 12 risks in our top risk list, which we could verify in the literature. We had 3495 hits in scientific databases (PubMed n = 1301, Web of Science n = 2102, Cochrane n = 92) and found 146 relevant studies and included 62 in our narrative synthesis. A detailed presentation of the included literature can be found in the following Table 1. Prevalence of malnutrition ranged from 38% to 78%. Malnutrition diagnosed by nutrition assessments was independently associated with increased ICU LOS, ICU readmission, IOI, and the risk of hospital mortality. The SGA had a better predictive validity than the MNA. Compared with nutrition assessment tools the predictive validity of nutrition screening tools was less consistent. [2] Canada Post hoc analysis of an existing database derived from a RCT.
To externally validate a modified version of the NUTRIC score.
1199 ICU patients with multi-organ failure, mechanically ventilated, with expected length of stay >5 days with a primary outcome of 28-day mortality.
Increased nutritional adequacy is associated with increased survival in patients with higher NUTRIC scores (>6) but not in patients with lower NUTRIC scores (<5). There is a strong positive association between nutritional adequacy and 28-day survival in patients with a high NUTRIC score but this association diminishes with decreasing NUTRIC score. Higher NUTRIC scores are significantly associated with higher 6-month mortality (p < 0.0001) and the positive association between nutritional adequacy and 6 months survival was significantly stronger in patients with higher NUTRIC score (p = 0.038). [3] Australia Retrospective audit of patients

Australia
Retrospective observational study, single-center To assess the feasibility of using the mNUTRIC tool to screen for patients at increased nutrition risk and to determine the proportion of those high-risk patients who were reviewed by a dietitian.
The median time required to complete a full mNUTRIC screen was 4 min and 54 s. During the study period, 160 patients admitted to the ICU were screened as being at low nutrition risk (mNUTRIC < 5). Of these patients, 63% (n = 101) were not reviewed formally by a dietitian. Eighty-one patients were flagged as high nutrition risk (mNUTRIC < 6); of these, 45 (56%) were formally reviewed by a dietitian and 36 (44%) did not have a dietetic consultation during their ICU stay. Enteral nutrition support to critically ill patients via the nasogastrojejunal approach guided by semi-automated ultrasound.

critically ill patients
The application of saline can be taken as sound window and the metal wire as the tracking target, the bedside nasogastrojejunal tube guided by semi-automated ultrasound is an effective feeding tube placement method. The total nursing service satisfaction of patients was 90.24%, and the total incidence of adverse reactions was 17.07%. [10] China Single-center, randomized controlled trial To compare the effectiveness of EM (electromagnetic)-guided and endoscopic nasoenteral feeding tube placement among critically ill patients. The primary end point was the total success rate of correct jejunal placement.
161 adult patients admitted to ICUs requiring nasoenteral feeding. Patients were randomly assigned to EM-guided or endoscopic nasoenteral feeding tube placement (1:1) Success rate was achieved in 74/81 and 76/80 patients who underwent EM-guided and endoscopic jejunal tube placements, respectively (91.4% vs. 95%; relative risk, 0.556; (CI), 0.156-1.980; p = 0.360). The EM-guided group had more placement attempts, longer placement time, and shorter inserted nasal intestinal tube length. They had shorter total placement procedure duration and physician's order-tube placement and order-start of feeding intervals. The EM-guided group had higher discomfort level and recommendation scores and lesser patient costs. [12] Italy Prospective, multicenter observational study To evaluate the validity of predictive formulas and equations for the calculation of energy expenditure and protein needs, by using indirect calorimetry (IC) and the protein catabolite rate; and to compare prescribed and actually received nutrients with estimated and measured needs.
42 critically ill adult patients hospitalized with acute kidney injury.
There were 654 days of artificial nutrition. Average energy and protein prescribed were respectively 1551 ± 644 kcal and 70.5 ± 38.2 g, while energy and protein actually delivered were 1408 ± 651 kcal and 63.4 ± 35.3 g (p < 0.0001 for both comparisons). In general, average energy needs measured by IC were significantly higher than both the prescribed and delivered nutrient amounts (IC 1724 kcal ± 431; prescribed 1575 ± 672; received 1439 ± 680, p < 0.0001). No predictive formula was precise enough, and Bland-Altman plots wide limits of agreement for all equations highlight the potential to under-or overfeed individual patients. 24.6% of the 350 nutritional prescriptions correctly estimated the need. In 40.0% of cases, nutritional needs were insufficiently covered. Overestimation occurred in the remaining 35.4%. Caloric prescription resulted in accurate delivery in 56.0% of cases. Effective feeding was not met in 32.6% of prescriptions, and in 9.14% actual feeding surpassed the prescribed amount by more than 10%. This study demonstrated a dissimilarity between the amount of calories prescribed according to current nutritional guidelines and the caloric need calculated by a stress-corrected Harris-Benedict equation. [14] USA Single center, comparative, longitudinal predictive study To quantify estimation errors against indirect calorimetry measurements indirect calorimetry was used to measure resting metabolic rate for 7 days. Three estimation methods were compared with the cumulative measurement. Cumulative energy expenditure was the primary end point.
The actual mean 7-day cumulative difference of the sample was −618 kcal with a standard deviation of 774 kcal. This difference was equivalent to −4.7% ± 6.2% of the cumulative measured value over 7 days. The difference between measured and estimated cumulative resting metabolic rate was not statistically significant (p = 0.079). Among the 7 patients in group 1 (standard Penn State equation), the cumulative error of the extrapolated value compared with the measured value was −1423 ± 1524 kcal (p = 0.049 vs. measurement) representing −8.8% ± 9.1% of the cumulative measurement. On average, the Penn State equations predict resting metabolic rate over time within 5% of the measured value. This performance is similar to the practice of making 1 measurement and extrapolating it over 1 week. Median measured energy expenditure was 2439 (1806-2703) kcal, the study estimate was 2247 (1986-2502) kcal (or −156 (−328 to 18) kcal lower than the measured expenditure), and the guideline estimate of 11-14 kcal/kg was 1444 (1259-1500) kcal (or −950 (−1254 to −595) kcal lower than the measured expenditure). Bland-Altman bias and 95% limits of agreement between the study estimate and measured expenditure was −8% (±46%) and between the guideline estimate, −49%. Poor clinical utility; and that, furthermore, measured energy expenditure increased over time with large individual variation [16] Country

Switzerland
Prospective interventional study over three periods (A, baseline; B and C, intervention periods).

Single-center
To measure the clinical impact of a two-step interdisciplinary quality nutrition program.
The daily energy balance difference was significant between periods A and C with a dietitian (p = 0.0012), whereas it was not significant between periods A and B. The normalized daily energy delivery (kcal day ≥ 1 or kcal kg ≥ 1 day ≥ 1) improved significantly in both periods B and C. The cumulated energy balance on day 7 improved progressively over the three periods, becoming significantly less negative. The cumulated ICU stay energy balances also improved significantly. The dietitian interventions significantly improved the day 7 energy balances. The BOB award ranking ranged from 1 for the best site to 81 for the worst site. There were significant correlations between the overall BOB score and nutrition adequacy (r = 0.94).
Regression analysis of the categorical variables suggested that the presence of a dietitian in the ICU was associated with a high BOB award ranking. After controlling for region, hospital size, and ICU structure, compared with ICUs without dietitians, the overall rank of ICUs with dietitians was 23.6 better. [18] USA Retrospective, single-center, performance improvement project The purpose was to evaluate the effect of registered dietitian nutritionist (RDN) order-writing privileges on enteral nutrition (EN) order compliance and nutrition delivery in ICUs.
50 critically ill patients, 150 EN days, data collection retrospectively via electronic health record.
Nonsignificant increase in EN order compliance occurred after implementation of RDN order-writing privileges, as measured by cumulative and component EN order parameters. Compliance increased by 17% for the cumulative EN order and 15% for the tube feed infusion rate order post-RDN order-writing privileges. RDN order-writing privileges improved EN order compliance and significantly improved protein delivery in selected ICUs The percent of protein needs delivered significantly increased from a mean (±SD) of 72.1% (±28.6) to 89.1% (±24.8) after implementation of RDN orderwriting privileges p < 0.001). and modular, hospital-build tube feeding (MTF).
In the hyperthermal group the quantity of microbial colonization soon exceeded FDA recommendations in the MTF-group. [21] USA Systematic review and meta-analysis To assess the potential effect of methodologic bias on nutrition trials. 15 RCT, Primary (mortality, morbidity) and secondary (time on ventilator or in intensive care unit/hospital, cost) outcomes were abstracted from each trial comparing early enteral nutrition (EEN) to no/delayed enteral nutrition.
EEN had a favorable effect on mortality (RR 0.61, 95% CI 0.41, 0.89) and infectious morbidity (RR 0.80, 95% CI 0.72, 0.89), but not on non-infectious morbidity or any secondary outcome. Mortality benefit was observed only in trials with more risks of bias; infectious morbidity benefit was observed in some analyses of trials with fewer bias risks. [22] China Systematic review and meta-analysis in the "no EN" group (p = 0.038). No patient had bowel obstruction, or ileus. One patient in the "no EN" group had VAP, compared with 0 in the early trophic EN group. Candida was isolated in subsequent urine or respiratory culture in 6/16 (38%) patients in the "no EN" group and in 1/15 (7%) patient from the early EN group (p = 0.083). To examine differences between prescribed and actual enteral nutrition (EN) delivery and to identify the specific causes of EN interruption and to quantify these.

Adult regional American
Burn Association-verified burn center, total of 90 subjects were studied. On postburn days 0 to 14 the daily volume of EN prescribed by the dietitian was compared with the actual volume received by the patient.
[ There were 62 interruptions to enteral nutrition delivery with the first data collection and 64 in the second. Prolonged fasting before and after surgery and airway procedures were initially identified as the two most important causes of delays.  GRVs in the SR and RL positions, although significantly and respectively different from the supine position, were not significantly different from each other (p > 0.05). [34] Brazil Systematic review To evaluate the effect of enteral feeding of critically ill adult and pediatric patients in the prone position on gastric residual volume and other clinical outcomes.
Four studies with adult patients and one with preterm patients were included. Main outcome = gastric residual volume Three studies did not show differences in the gastric residual volume between the prone and supine positions (p > 0.05), while one study showed a higher gastric residual volume during enteral feeding in the prone position and another group observed a greater gastric residual volume in the supine position (reduction of the gastric residual volume by 23.3% in the supine position versus 43.9% in the prone position; p < 0.01). Two studies evaluated the frequency of vomiting; one found that it was higher in the prone position (30 versus 26 episodes; p < 0.001), the other study no significant difference (p > 0.05). [35] France Before-after study, single-center To evaluate an intervention for improving the delivery of early enteral nutrition (EEN) in patients receiving mechanical ventilation with prone positioning (PP).
Eligible patients receiving EEN and mechanical ventilation in PP were included within 48 h after intubation in a before-after study. Patients were semi-recumbent when supine.
An intervention including PP with 25 • elevation, an increased acceleration to target rate of EN, and erythromycin improved EN delivery. Compared to the before group, larger feeding volumes were delivered in the intervention group (median volume per day with PP, 774 ml (IQR 513-925) vs. 1170 mL (IQR 736-1417); p < 0.001) without increases in residual gastric volume, vomiting, or ventilator-associated pneumonia. Canada Systematic review and meta-analysis To evaluate the effectiveness and safety of post-pyloric feeding versus gastric feeding for critically ill adults who require enteral tube feeding.
14 eligible studies including 1109 ICU patients There was no difference in mortality or duration of mechanical ventilation between the groups. Post-pyloric feeding is associated with lower rates of pneumonia compared with gastric tube feeding.
(moderate quality of evidence; RR 0.65, 95% confidence interval (CI) 0.51 to 0.84. To study the effect of implementation of fasting guideline 74 ICU patients 77% of staff were familiar with the guidelines, whilst 42% requested further education. The main barriers to guideline compliance were delays and unpredictable timing of procedures, and differing guidance from senior staff and non-ICU teams. Significant improvement in enteral nutrition (EN) delivery and reduced duration of feed breaks when using a protocol. [48] Germany Before and after design, single-center study To examine whether early enteral nutrition (EN) of critically ill patients could be improved by a nurse-driven implementation of an existing feeding protocol.
A total of 101 and 97 patients were included, respectively, before and after the intervention.

ICU nurses
Most patients did not achieve the prescribed daily calorie goal, associated with the use of midazolam and assistance by a reduced nursing staff. The most important barrier was "Not enough nursing staff to deliver adequate nutrition" (M = 4.80, SD = 1.81, 60%), followed by "Fear of adverse events due to aggressively feeding patients" (M = 4.59, SD = 1.50, 56%). [52] USA Cross-sectional survey, multicenter study To describe the barriers to enterally feeding critically ill patients from a nursing perspective and to examine whether these barriers differ across centers.
A total of 138 of 340 critical care nurses completed the questionnaire.
No or not enough dietitian coverage during weekends and holidays. The 5 most important barriers to nurses were (1) other aspects of patient care taking priority over nutrition, (2) not enough feeding pumps available, (3) enteral formula not available on the unit, (4) difficulties in obtaining small bowel access in patients not tolerating enteral nutrition, and (5) no or not enough dietitian coverage during weekends and holidays. [53] China Cross-sectional descriptive multicenter study To investigate the barriers in administering enteral feeding to critically ill patients from the nursing perspective. 808 nurses from 10 comprehensive hospitals Frequency of enteral nutrition (EN)-related training, full-time ICU nutritionist, hospital level, specific protocols for enteral feeding and position were significantly influencing the enteral feeding of ICU patients. 8 surgery residents completed the nutrition education program. Pre-and post-testing were performed to assess short-term comprehension.
The nutrition education program improved both short-term and long-term ICU nutrition knowledge of surgery residents (p < 0.01). [55] Australia Online questionnaire To explore Australian nurses' enteral nutrition (EN) knowledge and sources of information.

responses of registered nurses
Most respondents reported their EN knowledge was good (n = 205, 60.1%) or excellent (n = 35, 10.3%), but many lacked knowledge regarding the effect of malnutrition on patient outcomes. Dietitians and hospital protocols were the most valuable sources of enteral nutrition information, but were not consistently utilized. [56] South Korea Quasi-experimental, one-group study with a preand post-test design, multicenter study To evaluate the effects of an education program to improve critical care nurses' perceptions, knowledge, and practices towards providing enteral nutritional support for ICU patients.
Nurses' knowledge about enteral nutritional support showed a significant improvement after the education program (mean change = 11.2%, p < 0.001). Nurses' total practice score significantly improved after the program (mean change = 2.54, p < 0.001). [57]

Brazil
Prospective, non-blinded single-center study To evaluate the impact of a multifaceted nutritional educational intervention on the quality of nutritional therapy and clinical outcomes in critically ill patients.
16-bed ICU Phase 1: the quality of NT was evaluated in 50 newly admitted ICU patients in a pre-educational program (Pre-EP). Phase 2: nutritional protocols were created and an education program was implemented. Phase 3: another 50 patients were enrolled and observed in a post-educational program (Post-EP) using phase.
188 patient electronic medical records (EMR) The intervention of an education program reduced the documented discrepancy between the pump readings and charted volumes from 44 to 33%. A correlation analysis also showed a tighter relationship post-intervention (rpost = 0.84 vs. rpre = 0.76, both had a p < 0.01). [59]

USA
Prospective clinical trial The experimental group (EG) received targeted education consisting of strategies to increase delivery of early enteral nutrition.
Strategies included early enteral access, avoidance of nil per os (NPO) and clear liquid diets (CLD), volume-based feeding, early resumption of feeds post procedure, and charting caloric deficits. The control group (CG) did not receive targeted education but was allowed to practice in a standard ad hoc fashion.
Patients (n = 121) assigned to 1 of 2 trauma groups EG received a higher percentage of measured goal calories (30.1 ± 18.5%, 22.1 ± 23.7%, p = 0.024) compared with the CG. Mean caloric deficit was not significantly different between groups (−6796 ± 4164 kcal vs. −8817 ± 7087 kcal, p = 0.305). CLD days per patient (0.1 ± 0.5 vs. 0.6 ± 0.9), length of stay in the intensive care unit (3.5 ± 5.5 vs. 5.2 ± 6.8 days), and duration of mechanical ventilation (1.6 ± 3.7 vs. 2.8 ± 5.0 days) were all reduced in the EG compared with the CG (p < 0.05). EG patients had fewer nosocomial infections (10.6% vs. 23.6%) and less organ failure (10.6% vs. 18.2%) than did the CG, but these differences did not reach statistical significance. [60] Canada Questionnaire Multicenter Evaluation, enquiry study This study describes the results of an evaluation of educational strategies used to implement a novel enteral feeding protocol.
The response rate to the questionnaire was 166 of 434 or 38.2%.
More than 70% of respondents rated 5 of the educational strategies as very useful or somewhat useful. The percentage of nurses who found the bedside protocol tools of the enteral feeding order set, gastric feeding flowchart, and volume-based feeding schedule either "very easy" or "somewhat easy" to use were 64.0%, 60.5%, and 59.1%, respectively. The identified studies focused primarily on assessments or interventions and less on risks or safety issues. These risks and safety issues were derived from the studies directly and indirectly. A detailed description of the search process can be found in Figure 1 (by Moher et al. (2009) [14] adapted by Hoffmann).
The identified studies focused primarily on assessments or interventions and less o risks or safety issues. These risks and safety issues were derived from the studies direct and indirectly. A detailed description of the search process can be found in Figure 1 (b Moher et al. (2009) [14] adapted by Hoffmann).    2019) primarily recommended a general clinical assessment, including a history, report of weight loss or decrease in physical performance before ICU admission, physical examination, etc., in addition to screening and assessment instruments. Presence of frailty is considered to be significant in ICU patients, and should therefore be considered in nutritional management. Recording of muscle mass is also used as a parameter for assessing the nutritional status [3]. In a systematic review, Lew [1]. Further studies focused on instruments such as the NUTRIC score and the mNUTRIC score [16][17][18][19][20][21]. Given that the studies investigated different illnesses, different instruments, various concepts such as nutrition screening or assessment, and that there is no clear definition of critical illness-associated malnutrition [3], various risks can arise in practice. However, studies and guidelines recommended a clinical assessment supplemented by an easy-to-use validated screening instrument such as the NRS-2002(NRS-2002 and the NUTRIC score-determine both nutrition status and disease severity, while the use of a frailty scale can be helpful for elderly patients.

Inadequate Tube Management and Position
Singer et al. (2019) recommended the use of gastric access as the standard approach to initiate EN. Post-pyloric, mainly jejunal, feeding is possible for patients deemed to be at high risk of aspiration [3]. Ultrasonography, camera-assisted technology with real-time video guidance and X-rays are used to check the position of the tube after insertion and thus obtain a positive outcome [22][23][24]. In clinical practice further tube placement testing methods are used such as aspirate appearance, aspirate pH or auscultation, although X-rays and real-time video guidance are considered to be the most adequate methods. Different types of tubes designed for use with imaging procedures should also be considered. However, certain questions remain unanswered with regard to risk management, for example how long the tubes remain in the correct position, or after what period of time or type of intervention a further positional check is needed, including for jejunal tubes. It is important for individual ICUs to develop a protocol to guide their tube management policy.

Missing Energy Target
For mechanically ventilated patients, guidelines and studies recommended that EN should be determined by indirect calorimetry [25][26][27]. In the absence of indirect calorimetry, VO2, or VCO2 measurements and simple weight-based equations (such as 20-25 kcal/kg/d) should be used. In order to prevent risks, it is important that the prescribed quantity should match the calorie requirement, and that this should be re-evaluated in regular intervals [28], as the measured energy expenditure increased in the course of time with great individual variation [29]. De Waele et al. (2012) recommended a dedicated nutrition support team for a more systematic use of indirect calorimetry in long-term mechanically ventilated patients [30]. However, in order to determine patients' energy target, their nutritional status before admission to ICU should not be used [3].

Verification
Missing a Nutritionist at the ICU Two studies, examining the clinical impact of a two-step interdisciplinary nutrition program and enteral feeding protocols in the ICU, have found that interventions by a dietician significantly improved patient energy balances by day 7 [31,32]. Additionally, the presence of a dietician in the ICU has been associated with better nutrition performance, with a multi-professional approach reducing risks through shared responsibility, interdisciplinary quality programs, and re-evaluations of EN [33]. A systematic review by Mistiaen et al. (2020), stated, that there is weak evidence that Nutrition Support Teams increase appropriate EN use in ICU patients. A decrease of the duration of PN could not be shown [34].

Preparation Insufficient Hygiene and Handling
In their study, Perry et al. (2015) compared open systems, "ready-to-hang"-systems (RTH), and modular hospital-built tube feeding systems (MTF), in a normothermic (23 • C) and hypothermal ICU environment. The contamination in both environments/systems does not differentiate between open and closed feeding systems for up to 8 h. However, adding modules to open systems can lead to an unacceptable risk of contamination in hyperthermic (i.e., particularly warm) environments [35]. Training in preparing the setups, maintaining constant temperatures before and after preparation, as well as storage were important factors.

Wrong Time Management, Speed and Route
Medical nutrition therapy should be considered for all ICU patients, mainly for those staying for more than 48 h [3]. An oral diet is preferable to EN or PN for critically ill patients who are able to eat. If oral intake is not possible, early EN (within 48 h) in critically ill adult patients should be performed/initiated without delay [3,[36][37][38]. To avoid overfeeding, early full EN and PN should not be used in critically ill patients but should be prescribed within three to seven days. Singer et al. (2019) suggest to be caution in critically ill patients with uncontrolled shock, uncontrolled hypoxemia and acidosis, uncontrolled upper gastrointestinal bleeding, gastric aspirate >500 mL/6 h, bowel ischemia, bowel obstruction, abdominal compartment syndrome, and high-output fistula without distal feeding access. A systematic review and a meta-analysis showed that, when given within 48 h after admission, EN itself is efficient and safe for those patients with predicted severe acute pancreatitis [39]. An episode of vomiting was observed in patients with sepsis [40]. Blaser et al. (2017) and Zheng et al. (2019) confirmed that early EN reduced infectious complications in unselected critically ill patients, and in traumatic brain injury, severe acute pancreatitis, gastrointestinal (GI) surgery and abdominal trauma [41,42]. However, their recommendations are weak because of the poor quality of the evidence, with several information based only on expert opinion. In order to actively counter the risks, early EN should be monitored like a vital sign. Though the implementation of the ENFit™-standard it is still possible that enteral medication in Luer injections or tube feeding can be administered intravenously. A systematic review and meta-analysis by Alkhawaja et al. (2015) evaluated the effectiveness and safety of post-pyloric feeding versus gastric feeding. There was no difference in mortality or duration of mechanical ventilation but post-pyloric feeding is associated with lower rates of pneumonia compared with gastric tube feeding [43].

Nutritional Interruptions
Lee et al. (2018) reported 332 episodes of feeding interruptions, this means 12.8% (4190 h) of the total 1367 nutrition days. Each ICU patient experienced feeding interruptions for a median of three days. Total duration of feeding interruptions for the entire ICU stay: 24.5 h, which resulted in an energy and protein deficit. They therefore recommended an evidence-based feeding protocol and a nutrition support team [44]. Williams et al. (2013) investigated the number of nutritional interruptions. They cited education, audit, leadership support, interprofessional collaboration and the use of guidelines as starting points for reducing these interruptions [45]. Based on a chart review, Uozumi et al. (2017) also proposed the development of a protocol for nutritional interruptions, this could possibly reveal deficits in the administration of the EN at an early stage [46]. Prolonged fasting before and after surgery, airway procedures, dressing changes, feed intolerance, and tube malfunction were identified as the most important causes of delays by Segaran et al. (2016) [47,48].

Wrong Body Position
To reduce gastric residual volume in ICU patients, Farsi et al. (2020) recommended positioning patients in the right lateral and supine, semi recumbent positions rather than in the supine position [49]; however, this remains contradictory. There is insufficient literature on this subject. The use of a protocol based on the elevation of the patient's head, the use of fixed prokinetics and reduced speed of the diet allowed the application of early EN and faster attainment of the planned energy target in prone position, this was found by Regnier et al. (2009) [50]. However, the literature regarding the effect of EN while in the prone position is also sparse and of limited quality [51]. No studies on agitated patients, whose position can change continuously, have been found. Therefore, to prevent risks, ongoing clinical observations are needed.

Understaffing
Risk factors for inadequate nutrition support as described by Honda

Lack of Education
Studies show that multifaceted nutritional education programs and protocols are able to improve the knowledge of the healthcare professionals. These tools and training programs should be versatile, easy to use, and can be web-based [68][69][70][71][72][73][74][75].

Discussion
Our narrative synthesis highlights the risks of EN in ICU. Each process step, from admission to discharge, demonstrated certain risks, like no use of clinical assessment or screening nutrition tools, inadequate tube management and position, missing energy target, missing a nutritionist at the ICU, bad hygiene and handling, wrong time management and speed, nutritional interruptions, wrong body position, gastrointestinal complication and infections, missing or not using guidelines, standards or protocols, understaffing, and lack of education which are often intertwined and mutually dependent. Due to the heterogeneity of the ICU patient population, no consensual evidence-based protocols about EN were found. However, studies indicated, that for all the aforementioned risks, safety measures exist.
In health care systems in recent years, patient safety has become a priority issue [6]. In addition to individual measures, national and international strategies and protocols have attempted to overcome the most prominent hazards. In clinical risk management it is important to identify, analyze, and manage potential risks. The implementation of measures into routine procedures within complex hospital organizations like ICUs is challenging and have to be monitored regularly as adherence or compliance can be lacking [8].
Based on our findings we developed a short checklist (see Table S1) which can be used by key groups like ICU staff to detect risks in ICUs. Diverse reviewers, with different years of work experience, should use the checklist independently (e.g., nurses, physicians, and dieticians), because a person's knowledge is not the department's knowledge. Application of the checklist should be followed by a discussion of the results, implementation of relevant measures and ways of how a distinctive measure can be checked in the routine.
In our results, we did not describe all identified risk which were reported by staff and CIRS, such as interaction between medication or other additives and EN, involvement of relatives and missing discharge plans. This does not imply that these risks and safety issues do not exist, but reflects the lack of adequate literature. Due to a very high turnover in ICU staff [76] and a very complex setting, the implementation of multifaceted nutritional education programs and protocols based on the newest guidelines [3] are necessary to improve and secure the knowledge of healthcare professionals. These tools and training programs should be versatile, easy to use, can be web based [68][69][70][71][72][73][74][75] and trained at regular intervals, followed by internal and external audits [8].

Strengths and Limitations
Our narrative synthesis highlighted risks of EN in ICU. These risks can be observed, and institutional approaches exist for minimizing these risks, for example, by raising awareness, evidence-based protocols, guidelines, consulting professionals, and education. The narrative synthesis should prompt the readers to reflect on their own way of working and on the actual and potential risks. However, our study has several limitations. First, due to the lack of systematic reviews, we also included other studies with different methods and missing high-quality evidence. Depending on the study type, bias is possible. Many of these previous nutrition trials were open to bias because they were unblinded, very small or had other confounders. Second, our literature languages were restricted to German and English and the search terms were also limited due to a large number of studies. Published guideline recommendations for the management of nutrition in ICU patients remain largely supported by expert opinion and only a minority of the studies and reports includes high-quality evidence [77]. Finally, few of the identified studies addressed the risks and safety issues of EN directly, therefore further research is needed.

Conclusions
The aim was to identify risks in the management of EN in critically ill patients in ICU. Based on our results, numerous risks related to the management of EN in the ICU were discovered. Clinical experts can use the risk checklist and the recommendations drawn up to carry out their own risk analysis in clinical practice. Once risks have been identified, appropriate measures can be taken. From the authors' point of view, risk management with tools such as checklists or other risk analysis tools are important for improving patient safety in the ICU and secure knowledge. Further research is needed on how risk management can be implemented in the daily routine, so that the staff reflects on and reviews their own EN management.