Nutritional intake and gastro-intestinal symptoms in critically ill COVID-19 patients

Background & aims Critically ill COVID-19 patients seem hypermetabolic and difficult to feed enterally, due to gastro-intestinal (GI) symptoms such as high gastric residual volumes (GRV) and diarrhea. Our aim was to describe the association of nutritional intake and GI symptoms during first 14 days of ICU admission. Methods Observational study including critically ill adult COVID-19 patients. Data on nutritional intake [enteral nutrition (EN) or parenteral nutrition] and GI symptoms were collected during 14 days after ICU admission. Target energy and protein feeding goals were calculated conform ESPEN guidelines. GI symptoms included GRV (ml/d), vomiting, abdominal distension, and faeces (ml/d). High GRV's were classified as ≥2 times ≥150 ml/d and diarrhea as Bristol stool chart ≥6. GI symptoms were defined as mild if at least one symptom occurred and as moderate when ≥2 symptoms occurred. Acute gastrointestinal injury (AGI) grades of III were classified as GI dysfunction and grades of IV were considered as GI failure with severe impact on distant organs. Linear mixed model analysis was performed to explore the development of nutritional intake and GI symptoms over time at day (D) 0, 4, 10, and 14. Results One hundred and fifty patients were included [75% male; median age 64 years (IQR 54–70)]. BMI upon admission was 28 kg/m2 (IQR 25–33), of which 43% obese (BMI > 30 kg/m2). Most patients received EN during admission (98% D4; 96% D10-14). Mean energy goals increased from 87% at D4 to 93% D10-14 and protein goals (g/kg) were increasingly achieved during admission (84% D4; 93% D10-14). Presence of moderate GI symptoms decreased (10% D0; 6% D4-10; 5% D14), reversely mild GI symptoms increased. Occurrence of GI dysfunction fluctuated (1% D0; 18% D4; 12% D10; 8% D14) and none of patients developed grade IV GI failure. Development of high GRV fluctuated (5% D0; 23% D4; 14% D10; 8% D14) and occurrence of diarrhea slightly increased during admission (5% D0; 22% D4; 25% D10; 27% D14). Linear mixed models showed only an association between AGI grades III and lower protein intake at day 10 (p = 0.020). Conclusion Occurrence of GI symptoms was limited and seems no major barrier for EN in our group of critically COVID-19 patients. Nutritional intake was just below requirements during the first 14 days of ICU admission. The effect on nutritional status remains to be studied.


Introduction
The coronavirus disease 2019  is caused by the virus SARS-CoV-2 and can result in respiratory and gastro-intestinal (GI) complaints [1,2]. Approximately 20e30% of the hospitalized COVID-19 patients are being admitted to the intensive care unit (ICU) for respiratory and/or hemodynamic support [1,3]. In these patients, the median ICU stay ranges from 5 to 19 days [4] (see Fig. 1).
Initiating enteral nutrition (EN) within the first 24e48 h after ICU admission is recommended when patients are unable to eat [5,6]. Feeding critically ill patients is challenging due to several reasons such as GI dysfunction and mechanical problems (e.g., gastric tube occlusion or malposition) [7e9]. Optimal feeding goals are often not achieved [10,11], which can lead to a prolonged ICU stay and increased (infectious) complications [12]. Critically ill patients admitted with COVID-19 seem particularly hypermetabolic [13e15] and difficult to feed enterally, due to feeding intolerance resulting in high gastric residual volumes (GRV) and diarrhea [16,17]. Concerning feeding intolerance, different definitions for adult critically ill patients are used but there is no uniform definition [9]. Most commonly used definitions include the presence of GI symptoms such as diarrhea, bowel distension, vomiting, and absence of bowel sounds [18]. It is stated that in a conceptional framework for defining feeding intolerance in critically ill patients assessment of gastropareses and diarrhea have to be included [19]. GRV volumes over 300 ml (2x > 150 ml in preceding 24 h) are indicative of delayed gastric emptying [20], whereas faeces volumes over 350 g are likely to indicate malabsorption [21]. Due to a lack of clinical evidence a classification for feeding intolerance has been proposed based upon expert opinion. The proposed classification for acute gastrointestinal injury (AGI) ranges from grade I (Risk of developing GI dysfunction or failure) up to grade IV (GI failure with severe impact on distant organ function) [22].
Recently, AGI grades III and IV were observed in 50% of critically ill COVID-19 patients and were associated with a higher risk of prolonged mechanical ventilation and mortality [23,24]. Furthermore, in these patients, feeding intolerance based on clinical symptoms was frequently observed [16,23,25] and associated with poor clinical outcomes [16]. So far only one retrospective study in which 52 patients were included described the safety and GI tolerance of EN during the first week of ICU admission in critically ill COVID-19 patients [26]. Therefore, our aim was to further examine the association between EN and GI symptoms in a larger group of patients and beyond the first week up to first 14 days of ICU admission. We hypothesized that GI symptoms influenced achievement of adequate nutritional intake.

Study population and design
This observational study was conducted from April 2020 till November 2020 in the ICU of the Erasmus Medical Centre (MC), with the approval from the institutional review board of the Erasmus MC, Rotterdam, The Netherlands (MEC-2020-0336). The need for informed consent was waived by the Institutional review board. Conform our previous study [15], all adult patients (>18 years) admitted to the ICU with confirmed COVID-19 respiratory infection were included. Patients with a metabolic disease requiring a specific diet (e.g., Phenylketonuria) or (home)-parenteral nutrition (PN) starting >7 days before admission unrelated to COVID-19, were excluded.

Nutritional intake
Daily recorded nutritional intake included administered nutritional energy (kcal) and protein (g) and non-nutritional calories (e.g., propofol and glucose) intake per day during the first 14 days of admission. The initial goal was to increase nutritional intake gradually conform ESPEN guidelines [6,27,28]. Optimal feeding goals were determined individually per time point at D0, D4, D10, and D14 after admission by ICU dietitians [6]. The calculated nutritional goal (CNG) for energy corresponds with the estimated total energy requirement. Resting energy expenditure (REE) was measured with an indirect calorimeter (Q-NRGþ, Cosmed, Italy) conform the practical guidelines [29], unless contra-indications were present (e.g., hemodynamic instability, FiO2 > 70%) [30]. In case of contra-indications predictive formulas such as the WHO (BMI 30 kg/m 2 ) or Harris and Benedict (BMI > 30 kg/m 2 ) equation were used [31e33]. Correction of energy goal for influential factors were made individually and ranged from 0 to 30%. The CNG for protein generally ranges from 1.3 to 1.7 g/kg [5,6], taking into account the presence of wounds and continue renal replacement therapy [5,6]. Actual bodyweight (kg) upon admission was used to determine CNG, unless there was clinical manifestation of fluid overload, estimated usual bodyweight before admission was used [15]. The total delivered nutritional intake was calculated by adding all calories and grams of protein from EN, enteral liquid protein modules, PN, and non-nutritional sources.

Gastro-intestinal symptoms
Data of GI symptoms were collected at D0, D4, D10, and D14 after ICU admission. GI symptoms included GRV (ml/d), vomiting (y/n), abdominal distension (y/n), and faeces (ml/d). High GRV was defined as !2 times !150 ml/d conform our standard feeding protocol [5,6,20]. Diarrhea was defined as Bristol stool scale !6 [34]. GI symptoms were defined as mild if at least one symptom occurred and as moderate when !2 symptoms occurred. AGI grades of III were classified as GI dysfunction and grades of IV were considered as GI failure (with severe impact on distant organs) [19,22].

Route of feeding
EN was started 24e48 h after admission via nasogastric tubes, unless contraindications were present (e.g., haemodynamic instability). Generally polymeric high-protein formulas were administrated continuously [35], if necessary supplemented with enteral liquid protein modules. In case of high GRV, there was a lowthreshold switch to placement of nasoduodenal tubes. If EN was not (fully) tolerated, the use of (supplemental) PN was weighted case-by-case [6].

Statistical analysis
Data on baseline characteristics were analyzed by means of descriptive statistics: mean (SD), median [interquartile range (IQR)] and numbers [percentages (%)]. Development of nutritional intake over time was presented in figures. Linear mixed model analysis was performed to explore the development of nutritional intake and GI symptoms over time. With mixed models all available information of patients during the study period is used. Multivariable analyses were conducted with daily nutritional intake, as % of estimated REE, as a dependent variable (i.e., energy and protein intake). Four analyses were performed with four different independent variables: the single dichotomous GI symptoms (i.e., diarrhea and high GRV), the sum score of GI symptoms (i.e., moderate GI symptoms), and GI dysfunction (i.e., AGI grades > II). When a single variable occurred in less than 5% of the patients it was excluded for analysis because of limited information in this risk factor. Time was included as a continuous fixed effect. Associations were expected to differ per time point. Therefore, additional models were estimated with time as a factor and an interaction with the four different independent variables to obtain associations on the specific days. All models were adjusted for age, gender, BMI, comorbidities, SOFA score, and CRP, no variable selection procedures were used. To account for repeated measurements random intercepts were included in the models. Due to the linear mixed model analysis, no imputation was performed on missing data in the repeated measurements. The baseline characteristics were observed in all patients. Assumptions for linearity and normality of residuals were checked before interpretation of the results. Correction for multiple testing was performed. Data analysis was performed using IBM SPSS statistics for windows, version 25.0 (IBM Corp. Armonk, NY, USA). A two-sided p-value of less than a ¼ 0.05 was considered to be statistically significant.

Study population
A total of 150 patients critically ill COVID-19 were followed during the first 14 days of admission. The number of patients decreased over time, due to discharge from the ICU or death. As presented in Table 1, median age was 64 years (IQR 54e77) and 75% of the patients were male. Median BMI upon admission was 28 kg/ m 2 (IQR 25e33), of which 43% was obese (BMI>30 kg/m 2 ). Onethird of the patients had at least one comorbidity, in which diabetes mellitus type 2 (32%), cardiac (32%)-and respiratory disease (21%) were the most prevalent. Minority of patients (8%) had preexisting GI comorbidities (e.g., inflammatory bowel disease, gastrectomy, reflux oesophagitis). Most patients (71%) were transferred from other ICUs to the ICU in the Erasmus MC, with a median of 3 days (IQR 0e4) spend in another ICU.
A total of 38 patients (25%) died during ICU admission. Median ICU stay was 19 days (IQR 12e32) and surviving patients were ventilated for 18 days (IQR 10e30). Overall median length of hospital stay was 23 days (IQR 14e35).

GI symptoms
Presence of moderate GI symptoms decreased; reversely mild GI symptoms increased during admission ( Table 2). None of patients developed AGI grades IV. High GRV was highest at D4 and thereafter decreased. Diarrhea was seen in approximately a quarter of the patients at D4, 10, and 14. Overall vomiting and abdominal distention occurred in less than 5% of the patients. Use of laxatives remained respectively 75% during the first 10 days, reduced to 55% at D14. Prokinetics increased from 15% at D0 to 39% at D14.

Nutritional intake versus GI symptoms in the first 14 days
Linear mixed model analyses showed no significant association between diarrhea, high GRV, and AGI III and nutritional intake over time, as presented in Table 3. Moderate GI symptoms were associated with higher energy intake (p ¼ 0.008). Fourteen percent of the patients with high GRV received liquid enteral protein modules, compared to 86% in patients with low GRV. Even though in patients with high GRV who did not receive enteral liquid protein modules, the protein intake increased during admission (80% D4; 90% D10-14). Overall no difference for nutritional intake and GI symptoms was found between obese and nonobese patients, data not shown. Table 4 presents the associations between GI symptoms and nutritional intake per timepoint, after adjustments. At D0 moderate GI symptoms was significantly associated with higher energy intake (p ¼ 0.020) and diarrhea was significantly associated with higher energy-(p ¼ 0.039) and protein intake (p < 0.001). Thereafter these associations disappeared. AGI grades > II were significantly associated with lower protein intake on D10 (p ¼ 0.020). Overall, no significant associations were found at D14.

Discussion
This study emphasizes the possibility of adequate enteral feeding in critically ill COVID-19 patients with acceptable GI intolerance during the first 14 days of admission. Most patients could be fed enterally and nutritional intake was just below estimated goals during ICU admission. Moderate GI symptoms occurred in the minority of the patients on different time points (<10%) and none of the patients developed GI failure (AGI grade IV). Overall, patients presenting with moderate GI symptoms were associated with higher energy intake during admission. Furthermore, we found an association between AGI grades III and lower protein intake at D10.  Contrary to our hypothesis, GI symptoms had little influence on nutritional intake and vice versa. Both energy and protein intake were gradually increased during the observation period and mean delivery remained just below nutritional goals according to current guidelines (93% at D10-14). In our study around 80% of the patients received >80% of their nutritional goals during ICU stay. In those Table 3 Multivariable associations between GI symptoms c and nutritional intake in critically ill COVID-19 patients over time during first 14 days of ICU admission, presented separately for energy and protein intake (% calculated nutritional goal). AGI, acute gastrointestinal injury; BMI, body mass index; CI, confidence interval; CRP, C-reactive protein; GI, gastro-intestinal, GRV, gastric residual volume; SOFA, sequential organ failure assessment. Bold values in this table are considered significant (p < 0.05). a Exponent of the coefficients quantifies the increase or decrease in the average of the main outcome. b Adjusted for age, gender, BMI (kg/m 2 ), comorbidities(y/n), SOFA score, and CRP (mg/l). c Since vomiting and abdominal distention occurred in <5% of the patients, they were not taken into account. d P-values were corrected for multiple testing. Table 4 Multivariable associations between GI symptoms c and nutritional intake in critically ill COVID-19 patients at D0, 4, 10, 14 during ICU admission, presented separately for energy and protein intake (% calculated nutritional goal). patients who didn't reach these nutritional goals this might be explained by interruptions of feeding during medical procedures. This is in accordance with a recent study in which 90% COVID-19 patients received >80% of their nutritional needs with acceptable GI intolerance by D4 after ICU admission [25] and higher compared to a general ICU population in which it has been shown that 60e70% of the prescribed nutritional goals were delivered at D4 and D12 after admission [36]. This difference might be explained due to variations in disease course and its effect on nutritional support between a COVID-19 and a general non-COVID-19 ICU population.
Presence of moderate GI symptoms remained low in our study. Other studies in critically ill COVID-19 patients have used different definitions for feeding intolerance at different moments and therefore presence of reported feeding intolerance varies from 11 to 84% [16,23,26]. The liberal use of prokinetics and laxatives in our study could have decreased the presence of moderate GI symptoms. Comparison of results with non-COVID-ICU patients will be hampered by the heterogeneity of this population as several other factors might also influence GI symptoms (i.e., use of medication, opiates and pancreatic exocrine insufficiency) [19,21,37,38].
To harmonize the definitions for feeding intolerance an AGI classification has been proposed [22]. We found AGI grade III in 1e18% of the patients on the different time points and none of our patients developed grade IV (GI failure). This is line with a recent study in which none of the critically ill COVID-19 patients were diagnosed with AGI grade IV [39]. Other studies found higher percentages of GI dysfunction (AGI > II) in critically ill COVID-19 patients, 50e74% [23,24]. These higher percentages might be due to the subjective manner to classify AGI grades, which is a limitation of this classification system. In addition, medical policies and (GI) procedures per hospital may differ. Based on our results GI dysfunction seems no major issue in critically ill COVID-19 patients.
We observed an association between patients presenting with AGI grades III and lower protein intake at D10. Controversially, there was no association between these symptoms and energy intake. This might be explained by our caution in administrating enteral liquid protein modules in patients with AGI grades III. We recently observed a negative urinary protein balance in both the acute and the late phase of critically ill COVID-19 patients which might indicate a prolonged protein breakdown phase and the necessity to reconsider the need for a higher protein intake in this specific disease [15]. Based on these results it seems important to reassess the administration of enteral liquid protein modules to optimize protein intake in patients with AGI grades III.
Although we were able to feed our critically ill COVID-19 patients with acceptable GI symptoms conform our nutrition protocol, its effect on nutritional status and ICU-acquired weakness remains to be studied. It has been shown that muscle mass decreases rapidly during admission in critically ill patients due to the catabolic and immobilization state [40]. Future studies are recommended to assess nutritional intake in relation to body composition in COVID-19 patients during ICU admission and (long-term) followup after discharge from the hospital.
Some limitations of this study must be addressed. First, due to the observational study design all data concerning nutritional intake and GI symptoms/dysfunction were collected from medical charts and might be subject to underreporting. Second, the scope of this study was the first 14 days of ICU admission but provides no insights into the period afterwards. Taken into account the median length of ICU stay (9e15 days) of critically ill COVID-19 patients, a substantial part of ICU stay is included.

Conclusion
The occurrence of GI symptoms in our group of critically ill COVID-19 patients was low and seems no major barrier for providing EN. Nutritional intake was just below nutritional goals during ICU admission. It is recommended to monitor adequate protein intake in patients with more GI symptoms (AGI III). The effect on nutritional status and ICU-acquired weakness remains to be investigated.