Nutritional status of patients with COVID-19 1-y post-ICU stay: A prospective observational study

Objectives Patients discharged from the intensive care unit (ICU) often experience physical complaints and poor nutritional intake, which negatively affect their nutritional status (NS). The aim of this study was to describe the NS of patients with COVID-19 1-y post-ICU stay. Methods This was an observational study of adult patients with COVID-19 1-y post-ICU stay. NS assessment (nutrient balance, body composition, and physical status) was performed. We examined nutritional intake and nutrition-related complaints. Nutritional requirements were determined with indirect calorimetry and body composition with bioelectrical impedance. Fat-free mass index (FFMI) and fat mass index (FMI) were calculated. Physical status was determined using handgrip strength, the 6-min walk test, and the 1-min sit-to-stand test. Descriptive statistics and paired sample t tests were used for analysis. Results We included 48 patients (73% men; median age 60 y [IQR 52;65]). Median weight loss during the ICU stay was 13%. One-y post-ICU stay, 12% of weight was regained. Median body mass index was 26 kg/m2 and 23% of the patients were obese (body mass index >30 kg/m2 and high FMI). Of the patients, 50% had high FMI and 19% had low FFMI. Median reported nutritional intake was 90% of measured resting energy expenditure. Nutrition-related complaints were seen in 16%. Percentages of normal values reached in physical tests were 92% of handgrip strength, 95% of 6-min walking distance, and 79% of 1-min sit-to-stand test. Conclusions Despite almost fully regained weight and good physical recovery in adult patients 1-y post-ICU stay, NS remained impaired because of elevated FMI, even though reported nutritional intake was below the estimated requirements.


Introduction
Malnutrition and weight loss are frequently observed in critically ill patients with COVID-19 [1,2] most likely due to the concomitant hypermetabolic and catabolic state [3]. It is well known that muscle mass rapidly decreases during intensive care unit (ICU) stay and is associated with impaired muscle function, a longer hospital length of stay (LOS), and higher mortality [4]. One of the factors contributing to this is inadequate protein and energy intake. Inadequate protein and energy supply is associated with risk for mortality [5]. Approximately 60% of critically ill patients with COVID-19 survive [6] and return home [7].
After ICU discharge, these patients often continue to experience physical impairment (e.g., decreased muscle and lung function), fatigue, nutrition-related complaints (e.g., dysphagia), and reduced cognitive function and quality of life as main problems [8,9]. These aspects are also known as post-intensive care syndrome (PICS) [10]. It has been shown that 60% to 80% of ICU survivors remain functionally impaired 12 to 18 mo after their stay in the ICU [9,11] This work was partially funded by Nutricia Advanced Medical Nutrition, ZonMw grant (10430022010026), Rijndam Rehabilitation and Laurens. The funders had no advisory role in the study design; data collection, management, analysis, and interpretation of data; writing of the report; or the decision to submit the report for publication. PLML was responsible for the investigation, project administration, data curation, software, writing and preparation of the original draft. JFO and KFMJ were responsible for the supervision. All authors were responsible for the conceptualization, methodology, writing, reviewing, and editing of the manuscript. PLML received grant funding from Nutricia Advanced Medical Nutrition. RvdB

Contents lists available at ScienceDirect
Nutrition journa l homepage: www.nutritionjrnl.com and that the majority of (hospitalized) patients with COVID-19 is diagnosed with PICS [12À14]. Many patients do not return to work in the first year after hospital admission, thus placing a great burden on the patients, their family, the health care system, and society. Nutritional support strategies may decrease the burden of critical illness [15,16]. Therefore, it is recommended that the nutritional status (NS) of patients with COVID-19 post-ICU stay be carefully monitored to reduce complications [17]. Nutritional assessment (NA) is the quantitative evaluation of NS. This assessment involves interpretation of the following domains: Nutrient balance (i.e., nutritional intake, requirements, and nutrient losses); Body composition; Physical status (including muscle strength and functionality); and Biochemical status [18].
Based on limited short-term follow-up data, overall NS is not restored in postÀCOVID-19 patients. It has been shown that patients with COVID-19 were able to achieve an adequate reported nutritional intake 3 to 5 mo after ICU admission [19] despite frequently observed nutrition-related problems such as decreased appetite and feelings of being full [2,20]. Most of these patients (almost) fully regained their unintended weight loss 3 to 5 mo post-ICU stay [2,19]. Concerning body composition, normal fatfree mass index (FFMI) [21] and high fat mass [22] were observed in most patients 3 mo after recovery from acute COVID-19. Physical status was impaired 3 to 6 mo after ICU stay in most patients with COVID-19 [12,23].
So far no further (long-term) follow-up data of complete NA is available on patients with COVID-19 post-ICU stay [24]. Therefore, our aim was to investigate NS, specifically measured resting energy expenditure (REE), body composition, and physical status in patients with COVID-19 1 y after ICU discharge. We hypothesized that NS is still impaired 1-y post-ICU stay.

Study population and design
This prospective observational study was conducted from March 2021 through July 2021 in the Erasmus Medical Centre (MC) or at home if this was more convenient for the patient, with approval from the institutional review board of the [Blinded for peer review]. Patients with COVID-19 (>18 y of age) post-ICU stay were included. Some of these patients were involved in the ongoing 3-y prospective multicenter cohort study, CO-FLOW (COVID-19 Follow-up care paths and Long-term Outcomes Within the Dutch health care system) [25]. Patients with a metabolic disease requiring a specific diet (e.g., phenylketonuria) or (home) parenteral nutrition starting >7 d before admission unrelated to COVID-19, were excluded [3]. All patients were admitted to the ICU for invasive ventilation and in accordance with international guidelines, patients were generally discharged from the ICU when vital signs were stable without support and continuous monitoring and or when treatment was no longer required. Data on patient characteristics and NA were collected 1 y after ICU discharge. Written informed consent was obtained before data collection. Data management was performed using Castor Electronic Data Capture, version 2021.1 (Castor EDC, the Netherlands).

Patient characteristics
Patient characteristics included age, weight, height, sex, body mass index (BMI), ICU LOS, duration of mechanical ventilation and continuous renal replacement therapy (CRRT) during ICU stay, mortality risk (APACHE IV) upon ICU admission, rehabilitation trajectory after discharge (y/n), and comorbidities. Current weight (kg) 1 y after ICU stay was measured on a calibrated scale to the nearest 0.1 kg. For comparison reasons, body weight history before, during (ICU) admission, and after hospitalization was assessed from medical charts and checked with the patient.

Nutritional status assessment
A trained dietitian performed NA including nutrient balance, body composition, and physical status.

Nutrient balance
During ICU stay, all patients were fed according to our nutrition protocol, which is based on the guidelines of the European Society for Clinical Nutrition and Metabolism (ESPEN) [4]. Nutritional intake 1-y post-ICU was obtained by using a 24-h diet recall and a global diet history. Nutrition-related complaints (i.e., taste, smell, and/or swallowing problems, and losses) were examined by interviews. Appetite degree was evaluated on a numeric rating scale (0À10). Data on gastrointestinal (GI) symptoms and feeding intolerance included vomiting (y/n), abdominal distension (y/n), and diarrhea (Bristol stool scale 6) [26]. Optimal nutritional requirements were determined individually. The calculated energy requirement corresponded with the measured REE (mREE) using indirect calorimetry (Q-NRG+, Cosmed, Italy) during resting state according to the guidelines [27,28]. The total energy requirement was determined individually based on influential factors such as physical activity levels [29] and ranged from 10% to 50% (minimum of 10% in case of sedentary lifestyle to maximum of 50% in case of vigorously active lifestyle). Comparison of mREE and predicted REE (pREE) with formulas (mREE / pREE £ 100%) was performed to explore hypometabolism (<90%) and hypermetabolism (>110%) [30]. Furthermore, if available mREE during ICU admission was compared with mREE 1-y post-ICU stay. One-y post-ICU stay protein requirement ranged from 0.8 to 1.2 g/kg, with an average norm of 1 g/kg [4,31]. Actual body weight (kg) was used to determine requirements, unless there was fluid overload, estimated dry body weight was used [3].

Body composition
Body composition was measured with a multifrequency bioelectrical impedance analysis (InbodyS10, Inbody Co. Ltd., Seoul, Korea) with patients in a supine position. Based on segmental impedance and reactance whole body and segmental fat-free mass (FFM), fat mass (FM), skeletal muscle mass, and extracellular water (ECW) and intracellular body water were determined. Phase angle (PhA) was deduced and whole body (50 kHz) values were compared with reference values presented by Mattiello [32], in which age-dependent PhA values <5.6 to 7.2°for men and <5.1 to 6.2°for women were considered too low. FFMI and fat mass index (FMI) were calculated (i.e., FFM/m 2 and FM/m 2 ) and if indicated, FFM was corrected for fluid overload (i.e., ECW/total body water ratio >0.390). In case of fluid overload, estimated values derived from ECW were recalculated to dry weight using a standardized ratio of 0.380 for healthy persons [33]. Values <10th percentile were considered too low and >90th too high [34]. For interpretation of skeletal muscle mass index reference values from the European Working Group on Sarcopenia in Older People (EWGSOP2) were used [35], in which values <7 kg/ m 2 for men and <5.5 kg/m 2 for women were defined as low.

Physical status
Physical status was determined using handgrip strength (HGS) in all patients. HGS was measured with a handhold dynamometer (Jamar; Patterson Medical, Bolingbrook, IL, USA) and the test was performed with patients in a sitting position. The highest value of six measurements was compared with the reference values of Dodds [36], in which values <10th percentile were considered too low [37]. Additional functional tests (i.e., 6-min walk test [6MWT] and 1-min sit-to-stand test [1MSTST]) were obtained in patients who also participated in the CO-FLOW study in the [Blinded for peer review] [25]. For logistical reasons it was not possible to perform these functional tests in all patients. During the 6MWT, patients were instructed to walk as far and as fast as possible back and forth on 30-m track to obtain the 6-min walk distance (6MWD) [38,39]. The 6MWD was compared to reference values of Enright and Sherrill [40]. During the 1MSTST, patients performed as many repetitions of sit-to-stand (1MSTST repetitions) in 1 min as possible without using arm support [41]. 1MSTST repetitions were compared with reference values of Strassmann et al. [42].

Nutritional status assessment
Assessment of NS upon ICU admission was based on the ESPEN guidelines (i.e., low BMI and/or unintended weight loss) before ICU stay [43]. NS 1-y post-ICU stay was based on, among others, the Global Leadership Initiative on Malnutrition (GLIM) [44] and EWGSOP2 criteria [35] as presented in Table 1.

Statistical analysis
Data of characteristics and NA were analyzed by means of descriptive statistics: mean (SD) in case of normally distributed values or otherwise as median (interquartile range [IQR]), and numbers (%), where appropriate. Difference of BMI and mREE upon admission and 1-y post-ICU stay were analyzed with a paired sample t tests. A two-sided P < 0.05 was considered statistically significant. Data analysis was performed using SPSS version 25.0 (IBM, Armonk, NY, USA).

Study population
Of the 85 eligible patients, 35 (41%) withdrew because they were too tired for (further) participation. As two patients had incomplete NA (no indirect calorimetry), 48 patients were included ( Supplementary Fig. 1), of whom 35 patients (73%) came from the CO-FLOW study. As presented in Table 2, the study population consisted mostly of middle-aged men (73%). Upon ICU admission, median BMI was 29 kg/m 2 [IQR 25;30], 7 of the patients (15%) were classified as moderately malnourished, 12 as severely malnourished (25%), 24 as obese (50%; BMI >30 kg/m 2 ), and 5 (10%) as normal (not malnourished; Supplementary Fig. 2 Table 2). The prevalence of obesity based on BMI had decreased to 23%. The obese patients had high FMIs and were diagnosed with obesity according to our criteria. None of the patients were diagnosed with (obese) sarcopenia and none displayed either a phenotypic and a etiologic criteria for diagnosing malnutrition (Table 3 and Supplementary Fig. 3). Thirty-six of the patients (75%) followed a rehabilitation process after discharge, but indicated that they were unable to continue exercising independently afterward. The most common comorbidity was type 2 diabetes mellitus, as reported in 11 patients (24%).

Nutrition-related complaints and losses
Eight patients (16%) had at least one nutrition-related complaint. Two (4%) patients had combination of three complaints.  Table 1). Table 4 presents whole body composition values for all patients and Figure 1 [45,46] shows an overview of the measured FM, FFM, and TBW compared with the dry weight reference values for men and women separately. In 6 patients (13%), FFM was corrected for fluid overload. Nine patients (19%) had low FFMI and 24 (50%) had high FMI. A combination of high FFMI and FMI was present in 7 patients (15). Of these, 6 (13%) were living with obesity (BMI >30 kg/m 2 ). Only seven patients had low PhA (15%) and one had low SMI (2%). Percentage of normative values reached in HGS was 92%, and 4% of the patients had low HGS. In a subgroup of 35 patients (73%), 6MWT and 1MSTST were performed. In these patients, the mean 6MWD was 516 m ( §105). Overall, patients performed 95% ( §19) of predicted level of normal and six patients (17%) had a 6MWD below the lower limit of normal, of which two (33%) patients had low FFMI. The mean 1MSTS repetitions was 30 ( §12) and patient performance was 79% ( §30) of predicted.

Discussion
The main finding of this study with adult post-ICU COVID-19 patients is that unintended weight loss was almost fully regained 1y after ICU discharge, often in combination with a high FMI (50%). A low nutritional intake was reported in 80% to 90% of the patients compared with the total calculated or measured REE. Nutritionrelated complaints and GI symptoms were seen in 16% of the patients, of which swallowing problems were most often reported. A low physical status was only seen in the minority of the patients.
Most of our cohort almost fully regained their weight 1y after their stay in the ICU. This observation is in line with a study where 97% of weight loss in patients with COVID-19 was regained by 3 mo post-ICU stay [19]. In our study, obesity based on BMI significantly decreased to 23% compared with ICU admission, which is in line with other reports in which 7% to 40% of patients with COVID-19 were living with obesity up to 3 mo after discharge [49,50]. Because we found a high FMI combined with a decrease in obesity prevalence, most observed weight gain was likely composed of fat mass. Unfortunately, body composition was not measured during ICU stay in most of these patients, limiting our ability to make intraindividual comparisons.
None of our patients could be classified as being malnourished, however, the elevated FMI reflects an unhealthy NS. The presence of malnutrition in other studies varied from 15% to 100% in patients with COVID-19, but follow-up was only 3 mo after ICU discharge and in these studies also different definitions for malnutrition other than the GLIM criteria were used [19,49,50]. In contrast, 29% of our patients had one phenotypic criteria without a combination with one etiologic criteria for malnutrition. Although these patients were not diagnosed as malnourished, 19% had (prolonged) low muscle mass. These results underline the importance of measuring body composition during and after ICU stay to evaluate NS.
During critical illness, most patients with COVID-19 were found to be hypermetabolic [51] and in a catabolic state for a prolonged period [3]. However, how metabolism evolves after hospital discharge is unknown. In a subgroup of 22 patients in our study, we measured the REE during their ICU stay. One y after discharge, we found a broad range of mREE fluctuations ranging from a 41% decrease to a 59% increase of which 36% remained hypermetabolic. No association was found between intraindividual REE measurements. In this subgroup, this might be due to both the resolution of the hypermetabolic state and/or the altered body composition with a high fat mass and lower FFMI of the patients.
Concerning nutrient balance, the reported nutritional intake 1y post ICU-stay was »90% of mREE and 80% compared with total calculated energy and protein requirements. Whereas this might be influenced by underreporting, nutrition-related complaints and GI symptoms seemed no barrier to achieve a higher nutritional intake. Furthermore, physical activity level values may be lower than estimated for energy requirements since most patients with COVID-19 might experience exercise intolerance post-ICU stay [52]. Concerning physical status, only 4% of our patients had a low HGS and 17% of the patients had a 6MWD under the lower limit of normal, even though low FFMI was not frequently observed in these patients. Overall, most patients reached 95% of predicted 6MWD and 79% of predicted 1MSTST. This is in line with other studies reporting 90% to 110% of norm 6MWD in patients with COVID-19 1-y post-ICU stay [53,54]. In fact, significant improvement in these physical status parameters has been reported after 1 to 3 mo rehabilitation [22,49,55] and this trend continues during long-term follow-up. Importantly the rate and degree of recovery seems much better than in non-COVID ICU survivors [9,56,57]. This might be explained by a different effects of the primary disease on the patient's body (viral pneumonia versus for instance severe septic shock due to a bacterial infection) or by the development of dedicated rehabilitation programs for post COVID-19 patients [56]. Despite good physical recovery, chronic fatigue might prevent frequent physical exercise in our study population. It should be noted that 41% of the patients did not participate in this follow-up study because they were too tired. Most of our patients did follow a rehabilitation program after discharge but were possibly unable to continue exercising independently afterward despite good physical recovery because of persistent tiredness. This might not be detected by our assessment of physical status, but might the combination of (prolonged) fatigue and limited activity as seen in patients with COVID-19 [52,58À60] could explain the relatively high FMI in weight regain in our patients. This high FMI needs attention in relation to the lower reported nutritional intake than calculated and a rehabilitation program which must be extended for >2 to 3 mo. A high FMI is not favorable for long-term health outcomes [61À63]. Combined lifestyle interventions, especially nutrition in combination with exercise, are needed to improve body composition and thus NS and health outcomes.
Although to our knowledge, this is the first long-term follow-up study to assess complete NA in patients with COVID-19, some limitations should be addressed. First, the number of patients was limited and there was no control group; therefore, results should be interpreted with caution. However even without a control group, these data are of importance as it provides initial insight into the recovery of the COVID-group. Later studies should include comparisons with a control group. Second, selection bias may have occurred since participation in follow-up was voluntary. A considerable number of the eligible patients withdrew because they were too tired for (further) participation. Third, we used norm values largely based on a healthy population despite our population having a high prevalence of comorbidities. We opted for this approach because of a lack of pre-morbid data.

Conclusion
Because of the persistent elevated FMI, the NS remained impaired in our group of adult post-ICU COVID-19 patients, even though most patients fully regained body weight and had good physical recovery 1-y after their ICU stay. This might be explained by limited physical exercise due to persistent fatigue. Our findings emphasize the importance of NA measurements during and after ICU discharge to provide tailored individual nutritional support. Future studies must focus on the effect of nutritional and exercise interventions on maintaining muscle mass and preventing increase of fat mass after discharge from the ICU.   [44]. High FMI defined as P>90th percentile and both low FFMI [34] and low HGS defined as P<10th percentile [36]. Reduced nutritional intake defined as <50% of energy requirement for >1 wk [44].