Impact of Undernutrition Definitions on Its Prevalence in Hospitalized Patients

1. Rationale 1.1 Epidemiology and morbidity of undernutrition Undernutrition, whether related to disease, poverty, hunger, war or natural disasters, affects more than one billion people worldwide (1). In France, an empirical estimate suggests a prevalence of undernutrition among hospitalized patients next to 2 million or 30 to 50%. It should be underligned that undernutrition is very common in and out of hospital. It is widely accepted that undernutrition can be caused by a deficit in nutrient intake or malabsorption. However, it is now known that the presence of an inflammatory state during illness is associated with undernutrition (1). The inflammatory state leads to anorexia and decreased intake but also increases basal metabolic rate and muscle catabolism. Undernutrition is marked by a decrease in all markers of muscle mass (1). Thus, It is strongly associated with a decrease in the patient's functional capacity and an increase in morbidity and mortality and health care costs (1,4-6). 1.2 International consensus on the definition of undernutrition In 2019, the Global Leadership Initiative on Malnutrition (GLIM) published an international consensus report on the diagnostic criteria for undernutrition (1). The diagnostic criteria are numerous, and include aetiological criteria such as reduced food intake; inflammatory context; symptoms such as anorexia, weakness; phenotypic criteria such as weight loss, body mass index (BMI), fat to lean mass ratio, fluid retention and muscle function. Muscle mass is described as a major diagnostic criterion, since it is a direct indicator of protein catabolism linked to undernutrition, but also a reflection of patient functional impairment, as it is directly associated with functional capacity, autonomy and prognosis. Various measurement tools are suggested, such as bioelectrical impedancemetry, ultrasound or magnetic resonance imaging, giving the tool choice by the clinicians, depending on experience and local resources. Handgrip measurement is considered as a complementary measure. GLIM advises an assessment of the risk of undernutrition, using validated questionnaires such as the NRS (Nutritional Risk Screening) or the MNA-SF (Mini Nutritional Assessment-Short Form). 1.3 New definition of undernutrition in France The French National Authority for Health (HAS) revised the definition of undernutrition in adults (< 70 years) in 2019, which is very similar to, but differs in some important aspects, from the GLIM definition. It includes the use of aetiological and phenotypic criteria, and now includes the assessment of muscle strength, gait speed and bioelectrical impedancemetry. 1.4 Convergences and divergences between the international and French definitions In adults under 70 years of age, both the GLIM and the HAS recommend screening for undernutrition, but the HAS recommendations do not recommend the use of a precise tool for assessing the risk of undernutrition. The presence of at least one aetiological criterion and one phenotypic criterion is mandatory to make a diagnosis of undernutrition in both situations. The differences between the international consensus and the HAS recommendations concern the diagnostic criteria. 1.5 Assessment of muscle function Physiotherapists use different tools to assess different aspects of muscle function, such as strength or walking speed. Isometric grip strength is strongly correlated with lower limb muscle strength, and is a negative marker of mobility in case of weakness (14). A linear relationship has been observed between grip strength and patient disability as measured by the Katz independence scale. In addition, grip strength has been shown to be a good predictor of the patient's clinical condition (19). The measurement of grip strength, using a dedicated dynamometer, also has the advantage of being simple and reliable and has reference values adjusted for gender and BMI. Gait speed is another indicator used to assess muscle function. It is associated with lower limb strength and is a predictor of disability and mobility limitation (14). Usual walking speed is most often measured over a distance of 4 or 6 m and has reference values adjusted for gender and height (19). Dietitians are also involved in the assessment of muscle function by measuring body composition in terms of lean body mass and fat mass. They perform bioelectrical impedancemetry, which estimates the volumes of fat and lean body mass by measuring the resistance of biological tissues to a low-intensity, high-frequency sinusoidal electric current through electrodes (14). This measurement method is simple to use, inexpensive, reproducible and feasible in bedridden patients (14). It also correlates very well with MRI and reference values are obtained in a variety of adult populations (12,22). 1.6 Research hypothesis The investigators hypothesise that the prevalence of undernutrition in a population of patients hospitalised in diabetology-obesity, pneumology, oncology and gastro-nutrition, under 70 years of age is different according to the diagnostic criteria recommended by the French National Authority for Health or those recommended by the Global Leadership Initiative on Malnutrition. Moreover, these two definitions could have a different impact on patient morbidity and mortality. The choice of the method of assessment of muscle function could impact on this prevalence.
2. Objective 2.1 Main objective The main objective of this study is to compare the prevalence of global undernutrition based on the diagnostic criteria recommended by the French National Authority for Health with that based on the diagnostic criteria recommended by the Global Leadership Initiative on Malnutrition, in patients hospitalised in diabetology-obesity, pneumology, oncology and gastro-nutrition 2.2 Secondary objectives The secondary objectives will be, in patients hospitalised in diabetes-obesity, pneumology, oncology and gastro-nutrition units:

• To compare the prevalence of severe undernutrition according to the HAS and GLIM diagnostic criteria;
• To compare the morbi-mortality of moderate and severe undernutrition according to the HAS and GLIM diagnostic criteria:
• Length of hospital stay ;
• Mortality rate ;
• Autonomy at discharge.
• To compare the predictive capacity of muscle function assessment tools on undernutrition according to the HAS and GLIM diagnostic criteria.

1. Type of study This is a prospective, monocenter, obversational cohort study. The research will be conducted in accordance with the protocol. The duration of recruitment will be 20 months with follow-up of patients during their entire hospital stay in the care unit, i.e. approximately 30 days (average length of stay), i.e. a total duration of the study estimated at 21 months. The study will take place at the Forcilles Hospital, in diabetology-obesity, pneumology, oncology and gastro-nutrition units.
2. Usual assessment procedures 4.1 Nutritional assessment At the patient's admission, a nutrition assessment is carried out by the dietician following the recommendations issued by the French Association of Dieticians-Nutritionists (AFDN) and the HAS. The dietician collects all the data required to detect malnutrition, such as weight, height, BMI, albumin levels, as well as grip strength and walking speed (carried out by the care assistant and the physiotherapist). The dietician also looks for possible obstacles to feeding, such as dental condition, swallowing problems or digestive discomfort. The doctor looks for various pathologies affecting digestive absorption such as short stool syndromes, gastrectomies, pancreatic insufficiency etc. It also looks for factors of aggression such as cancer, chronic progressive diseases, infections etc. All of this data is used to estimate the patient's theoretical nutritional needs and to adapt the nutritional management. The dietician, in collaboration with the nurses' aides, monitors the patient's intake. 4.2 Anthropometric measurements When the patient is admitted, the nurses measure the patient's weight and height. If the patient can maintain the erect position, a conventional scale and a measuring rod are used for the measurements. If the patient cannot maintain this position, a weighing chair and a laser meter are used. In order to assess weight loss prior to hospitalisation, the previous weight is sought from a hospital report or other recent medical document. Contact with the attending physician is also made in order to obtain information about the patient's weight history. If this information is not available, the patient's previous weight is collected. 4.3 Biological measurements On admission, a biological sample is usually taken from patients at risk of undernutrition in order to measure pre-albumin, albumin and C-reactive protein (CRP). These elements make it possible to determine the severity of undernutrition when it is present, according to the diagnostic criteria of the HAS. 4.4 Muscle function measurements Measurement of grip strength ("Handgrip") 4.5 Gait speed measurement
3. Assessment procedure added by the research 5.1 Mini-Nutritional Assessment Short Version (MNA-SF) score 5.2 Bio-electrical impedance measurement 5.3 Diagnosis of undernutrition according to GLIM criteria The diagnosis of undernutrition will also be carried out in the framework of this research according to the GLIM criteria. It will be based on the search for at least 1 phenotypic criterion and 1 etiological criterion. Severe undernutrition will be based solely on phenotypic criteria. This diagnosis according to the GLIM definition will be made by one of the investigators not involved in the management of the patient. The patient's team will be blinded to these results so as not to influence the management of the patient.

5 Statistical aspects 5.1 Calculation of the number of subjects required According to our internal epidemiological data, the prevalence of undernutrition according to the HAS definition was 34%. Taking into account the wider period of weight loss and the higher BMI threshold used by GLIM in its definition of undernutrition, the investigators estimate a higher prevalence of undernutrition. In the absence of data in the literature, the investigators estimate a prevalence of undernutrition according to the GLIM definition at 49%, and a difference of at least 15%. The investigators wish to highlight this discordance. With a risk α = 0.05 and a power of 80%, it will be necessary to include 260 patients.

5.2 General aspects Descriptive statistics will be based on means (+/- standard deviation) or medians [interquartile range] depending on the distribution of quantitative variables. Categorical variables will be described in terms of numbers and percentages. Univariate comparisons will use the usual statistical tests after checking the distribution of the variables (Chi2 or Fisher's test, t-test, anova or their non-parametric equivalents Wilcoxon and Kruskal-Wallis tests). Paired-sample tests will be used if necessary.

Tests will be performed at the 5% significance level. The 95% confidence intervals will be provided for each estimate.

Calculations will be made using IBM SPSS v21 and R software (version 3.6.1, http://www.R-project.org ).

5.3 Main objective The difference in the prevalence of global undernutrition between the HAS and GLIM groups will be compared using the McNemar test for comparison of proportions in matched samples.

5.4 Secondary objectives 5.4.1 Comparison of the prevalence of severe undernutrition The difference in the prevalence of severe undernutrition between the HAS and GLIM groups will be compared using the test described in the primary objective.

5.4.2 Comparison of morbidity and mortality Morbidity and mortality will be compared between the HAS and GLIM groups, for moderate and severe undernutrition, by comparing length of hospital stay, mortality and independence at discharge.

Length of hospital stay There is a relationship between the incidence of mortality and all duration type variables. Length of stay will therefore be analysed using a competitive risk method, which allows for the fact that patients who die are no longer eligible for length of stay analysis (25).

The estimator also provides the probability of in-hospital death, so that the probability of hospital discharge versus death can be assessed simultaneously at any time after inclusion in the study.

Mortality rates Mortality rates will be estimated using the Kaplan and Meier method. Comparison of the 2 curves will be performed by a Logrank test. The mortality rates in the 2 groups will be estimated via the survival curves.

ADL The ADL score will be compared between the 2 groups using the discrete numerical variable comparison tests described in the general aspects.

5.4.3 Comparison of muscle function assessment tools The impact of the choice of muscle function assessment tool will be assessed by comparing the prevalences of undernutrition diagnosed by each tool and by comparing the presence of reduced muscle mass by each tool.

6 Expected results in terms of scientific and professional advances The results of this study will make it possible to verify whether the GLIM definition increases the prevalence of undernutrition compared to that of the HAS. The use of one or other of the definitions could have an impact on the medical management of undernutrition. On the other hand, muscle function benefits from various assessment tools. These tools assess different aspects of the muscle, such as strength, functional performance or quantity. Muscle function has an important place in the new definitions of undernutrition, involving physiotherapists in addition to dieticians, nurses, care assistants and physicians in the screening for undernutrition. The different tools could lead to a different estimation of the reduction of muscle mass and thus a different prevalence of undernutrition. The results of our study will help to evaluate this and guide professionals in the choice of tools for assessing muscle function.

7 Expected benefits and risks for patients 7.1 Minimal risks and constraints added by the research All medical examinations and management are usually carried out with the exception of the MNA-SF questionnaire and bioelectrical impedancemetry, which will be carried out systematically in the patients included. The MNA-SF questionnaire is very quick (less than 5 minutes) and therefore causes very little inconvenience to the patient apart from an additional visit by a carer. Bioelectrical impedancemetry is a tool for assessing body composition using electrodes applied to the skin surface. The examination takes an average of 15 minutes, and is non-invasive and painless. The minimal stress is related to the placement of the electrode and the duration of the examination.

7.2 Expected benefits for the patient This study may highlight a difference in the prevalence of undernutrition according to the definitions and tools used. The results of this study will not provide direct benefits for the patient, but could allow the clinician to be more vigilant if undernutrition is underestimated, according to the criteria used in France. The benefit would then be a broader screening of undernutrition and would allow the adaptation of the patient's nutritional and rehabilitation management.