Artículo Original / Original Article Associations between nutritional risk at hospital admission and incidence of complications, hospitalization time and mortality

The early identification of nutritional risk is essential for the earliest possible implementation of nutritional therapy. The aim of the present study was to evaluate the association between nutritional risk and the incidence of clinical complications. An observational study was conducted at the internal medicine infirmary of a Brazilian public hospital involving patients admitted between January 2016 and December 2017. The NRS-2002, anthropometrics and laboratory exams (initial and final) were used for the evaluation of nutritional risk. The prevalence of nutritional risk was 63.6%. Nutritional risk upon admission to hospital was associated with the occurrence of complications, a longer hospitalization time and death. In correlation analysis, considering anthropometrics and laboratory exams compared to hospitalization time and time elapsed prior to the onset of the diet, serum albumin was inversely correlated with both hospitalization time and nutritional risk. Based on the present findings, knowledge on the nutritional status of patients and adequate nutritional therapy can lead to fewer complications during hospitalization.


INTRODUCTION
Malnutrition in the hospital setting is related to a greater incidence of complications, delayed recovery time, a longer hospital stay, an increase in hospital costs, readmission of the patient and mortality 1,2 . When a patient's nutritional status is compromised, organic, cardiac, respiratory, intestinal, renal and immunological functions are affected, which could lead to the occurrence of infectious complications that have repercussions regarding clinical outcomes 3 . Undernourished hospitalized patients undergo a decline in functioning and are bedridden most of the time, which favors the occurrence of pressure ulcers; moreover, malnutrition hinders the healing process 4 .
According to Aquino and Phillippi 5 , the earliest possible identification of nutritional risk (NR) is an important care objective in the global treatment of hospitalized patients.
Adequate diagnosis is essential to the timely implementation of nutritional therapy.
The assessment of NR is of immense importance when performed upon admission to hospital. The present paper reinforces the importance of this assessment for better screening and a reduction in clinical complications. Therefore, considering the need for the assessment of NR for the earliest possible intervention, the aim of the present study was to evaluate the frequency of NR upon hospital admission and its association with the incidence of complications, hospitalization time and mortality at a public hospital located in northeast Brazil.

MATERIALS AND METHODS
An observational study was conducted, in which patients were evaluated upon hospital admission through the last day of hospitalization. Data collection involved the analysis of patient charts and nutritional follow-up charts of adults admitted to the internal medicine infirmary between January 2016 and December 2017 at the Miguel Arraes Hospital in the city of Paulista, which is located in the metropolitan region of Recife, Brazil.
The inclusion criteria were male and female adults with nutritional follow-up charts with complete information on diagnosis, admission date, date of discharge or death, nutritional screening, anthropometric nutritional assessment, biochemical analysis upon admission, subsequent analyses and dietary evolution. Individuals with amputations, those with edema (who have liver or kidney disease), those submitted to bariatric surgery and those undergoing corticoid therapy were excluded from the study.
A nutritional risk screening (NRS-2002) was performed within the first 48 hours of admission to the hospital and patients were divided into two groups: with and without NR. The anthropometric evaluation was performed upon admission, with reevaluations on a weekly basis. Admission and followup routines included the determination of weight using a mechanical scale (FILIZOLA, model 31) with a capacity for 150 kg and height using a stadiometer. For bedridden individuals, weight was estimated using the formula proposed by Chumlea et al 6 . without the subscapular skin fold. Arm circumference (AC) was determined with a non-elastic tape measure with the elbow flexed at 90º and measured the midpoint between the acromion and olecranon. Weight and height were used to calculate the body mass index (BMI) considering the reference cutoff points established by the World Health Organization 7 . AC was evaluated based on Blackburn et al 8 .
Patients who received exclusive oral feeding and the use of supplementation, patients in enteral nutritional therapy and those in mixed nutritional therapy participated in the present study.
For the biochemical evaluation (albumin, C-reactive protein and leukocytes), we considered the values established by the outsourced laboratory that offers this service to the Miguel Arraes Hospital (CERPE Diagnostics), which used the following cutoff points: albumin: 3.5-5.0 mg/dl; CRP: ≤ 5 mg/dl; and leukocytes: 3500-11000/UL. The first and last exams of each patient during hospital stay were considered.
We evaluated the delay until the onset of the diet due to hemodynamic instability or the need to undergo exams. For the evaluation of clinical complications, we investigated infectious conditions (analysis of leukocytes), inflammation (analysis of CRP), lowered level of consciousness and admission to the intensive care unit. The hospital outcome (discharge or death) was identified through information obtained from the nursing staff. The databank was created and analyzed using the Statistical Package for the Social Sciences (SPSS, version 13.0, SPSS Inc., Chicago, IL, USA). The Kolmogorov-Smirnov test was used to test the normality of the data distribution. Variables with normal distribution were expressed as mean and standard deviation values. Variables with non-Gaussian distributions were expressed as median and interquartile range. In the description of proportions, binomial distribution was approximated to normal distribution by the 95% confidence interval.
The paired t-test was used for the comparison of means of the anthropometric and biochemical variables at both evaluation times (admission and outcome). The paired Wilcoxon test was used for the comparisons of non-parametric data. Proportions were compared using Pearson chi-square test, the linear trend chi-square test or Fisher's exact test.
Either Pearson's or Spearman's correlation coefficients were calculated to determine the strength of correlations between the anthropometric/laboratory data and both hospitalization time and number of days elapsed until the onset of the diet adjusted for sex and the occurrence of NR upon admission to hospital.
This study received authorization from the administration of the Miguel Arraes Hospital. The study also received approval from the Human Research Ethics Committee of the Federal University of Pernambuco (certificate number: 81871717.1.0000.5208) in accordance with Resolution nº 466/2012 of the Brazilian National Board of Health and was conducted following the ethical precepts established in the Declaration of Helsinki. As the data were collected from patient follow-up charts, the ethics committee waived the need for a signed statement of informed consent.

RESULTS
Approximately 750 patients are admitted to the hospital per year. The present study was conducted between January 2016 and December 2017. Therefore, we started with 1500 patients, among whom 806 were eligible for the study after the application of the exclusion criteria. Mean age was 46.4 ± 12.5 years. Male sex accounted for 58.3% of the sample. The main causes of hospitalization were infectious-parasitic disease (28.6%), followed by chronic non-communicable disease (12.1%) ( Table 1).
Nutritional risk upon admission was 63.6%. Median hospitalization time was 26.0 days (range: 17 to 41) and 54.6% of patients had no complications. Diet was administered mainly through the oral route and 80.0% of the patients were discharged from hospital (Table 1). Tables 2 and 3 display the results of the analysis of the associations between NR and aspects of hospitalization. Among men (Table 2), NR upon admission was associated with a longer time prior to the onset of the diet, greater number of days of zero diet for exams, the occurrence of complications, a longer hospitalization time, negative outcome (death) as well as leukopenia and high C-reactive protein (CRP) at the end of follow-up. Among women (Table 3), NR upon admission was associated with a larger number of days of zero diet for exams, the occurrence of complications, a longer hospitalization time, the occurrence of weight loss, the negative outcome (death) and hypoalbuminemia at the end of follow-up.
In correlation analysis, considering anthropometrics and laboratory exams compared to hospitalization time and time elapsed prior to the onset of the diet adjusted by sex and the occurrence of NR upon admission, serum albumin was inversely correlated with both hospitalization time and time until the onset of the diet in men and women with NR (Table 4).

DISCUSSION
In the present study, 63.6% of patients were at NR upon admission to hospital. This high prevalence may be explained by the fact that the patients were at a public emergency hospital in the city of Paulista in the metropolitan region of Recife (northeast Brazil), the majority of whom had little access to basic healthcare services. Moreover, patients affected by conditions that require emergency care generally have a high hypercatabolic state, which contributes to the aggravation of poor nutritional status.
Previous studies that also employed the NRS-2002, report much lower frequencies of NR compared to that found in the present study. Among patients at a public hospital in the city of Recife, 16.7% were at NR 9 . In a study conducted by Lisboa et al 10 in a internal medicine infirmary at a university hospital in the state of Pernambuco (PE), 39% of patients were diagnosed with NR in the first 48 hours after admission. At a general surgery infirmary of another university hospital, the frequency of NR among patients was 31.1% 11 . Raslan et al 12 evaluated 705 patients and found that 27.9% were at NR. The high frequency of NR in the present study may have been due to the fact that infectious-parasitic diseases were the main cause of hospitalization, along with the high frequency of respiratory, kidney and liver diseases. Thus, admission to hospital was due to the exacerbation of these diseases, increasing the hypercatabolic state, with a consequent negative impact on nutritional status.
The higher number of male patients compared to females has also been reported in previous studies 13,14 . This reflects the chronicity and disease complications due to the failure on the part of men to seek basic healthcare services, only doing so when conditions become so Table 4. Correlation between anthropometric/laboratory variables and hospital stay/number of days until onset of diet adjusted for sex and nutritional risk in patients admitted to a hospital in Paulista, Brazil, between January 2016 and December 2017.

Variables
Hospitalization is also in agreement with data found in the previous studies, which describe a mean of 45 to 59 years 9,12,15,16 . The mean hospitalization time was 26 days, which may have been due to the most frequent type of adverse health condition and the occurrence of complications, along with the fact that the internal medicine infirmary is an investigational setting in which patients spend more time for the confirmation of the diagnosis. Previous studies report a shorter hospitalization time, with a mean of 15 days 12,16,17,18 . Moreover, hospitalization time was longer among those with NR upon admission. According to Kyle and collaborators (2001) 19 , both the depletion of lean mass and an excess of fat mass exert a negative impact on hospitalization time.
In the analysis of the associations with the variables analyzed during hospitalization, NR upon admission was associated with a later onset of the diet among the men. It is possible that patients at NR have more serious conditions, with hemodynamic instability, digestive hemorrhages and the preparation for exams or surgical procedures. The literature reports that patients at NR and with a late onset of the diet have greater catabolism and a higher risk of weight loss, especially lean mass, leading to malnutrition and its probable complications 4 .
Among women, no association was found between NR and later onset of the diet. However, the greater need for nutritional support, complications and death were more probable among both men and women at NR. The most frequent complications were respiratory and urinary infections, which can be caused by an unfavorable nutritional status 20,21 .
In a review of the literature published in 2005, authors concluded that patients at risk for the development of pressure ulcers can be identified early through the nutritional assessment, including a biochemical analysis, anthropometric measurements, clinical signs, dietary history and energy expenditure 22 . Pressure ulcers are more frequent in seniors, hospitalized patients and individuals with chronic diseases, such as stroke, cancer and spinal cord injuries.
Clinical outcome is an important aspect to consider. Patients at NR and those with malnutrition have worse outcomes, including an increased risk of mortality. In the present study, deaths were more frequent among the men and women with NR upon admission to hospital 1 .
The prevalence of hypoalbuminemia was high at both evaluation times. Amaro and collaborators (2016) 1 also found albumin levels below the reference value in patients at two evaluation times. The association between NR and hypoalbuminemia was found in the women. Previous studies have also found hypoalbuminemia more frequently in malnourished patients and those at NR 23,24,25 . In contrast, CRP was higher than the laboratory reference value. Previous studies report similar results, with mean CRP values of 3 mg/dL 26,27,28 . These findings are due to inflammation caused by either the baseline disease or an associated condition, reflecting the clinical status and severity of the disease.
Albumin is a negative acute-phase protein, meaning that it is reduced in the inflammatory phase, whereas CRP is a positive acute-phase protein, meaning that it is increased during the inflammatory process of a disease 29 .
The late onset of the diet and/or longer fasting time for the execution of exams was negatively correlated with albumin (patients with a longer fasting time had albumin below the reference range). Albumin was correlated with hospitalization time among the patients at NR upon admission.
In the present study, 13.2% of male and 13.9% of female patients at NR died. In the literature, this frequency is around 18.0% among patients at NR 29 . In undernourished patients and those at NR, there is a greater possibility of the occurrence of clinical complications due to the greater susceptibility to infection and surgical complications and a greater risk of death.
Nutritional screening performed upon admission to the hospital is an important factor to ensure better outcomes, as preexisting NR together with clinical complications increase morbidity and mortality rates among patients.

CONCLUSION
Determining the nutritional status of patients upon admission and performing adequate nutritional follow-up improves patient status during hospitalization, thereby minimizing the occurrence of complications, lowering the risk of morbidity/mortality and shortening the hospital stay. However, being in the ideal range for BMI is not necessarily a protective factor against NR. The present findings underscore the need for the earliest possible screening for NR to ensure the wellbeing of hospitalized patients.