Title: Nutritional risks predicts recovering full oral intake among elder convalescent stroke patients undergoing enteral nutrition: data from multicenter cohort study (the APPLE study).
Abstract
Background & Aims: Whether nutritional risks correlated with recovery of swallowing function of convalescent stroke patients was uncertain. This study was conducted to clarify whether nutritional risks predicts achievement of full oral intake in convalescent stroke patients undergoing enteral nutrition.
Methods: This retrospective cohort study were included 466 convalescent stroke patients undergoing enteral nutrition aged 65 years or over extracted from the database of “Algorithm for Post-stroke Patients to improve oral intake Level; APPLE” study investigated in the Kaifukuki (convalescent) rehabilitation wards. Individuals with missing data, suspected dehydration, and presence of cirrhosis or renal diseases were excluded. Nutritional risks were determined by the Geriatric Nutritional Risk Index (GNRI) as follows; severe (< 82), moderate (82 to <92), mild (92 to <98) and no nutritional risks (≥ 98). Swallowing function were assessed by the Fujishima’s swallowing grade(FSG) on admission and discharge. The primary outcome was achievement of full oral intake defined by FSG ≥ 7. Binary logistic regression analysis were examined to clarify predictive factors including nutritional risks for achieving full oral intake. Estimated hazard risk was computed by Cox’s hazard model.
Results: Of the 466 individuals, 257 were included in this study. Participants with severe nutritional risk showed significantly lower proportion of achievement of full oral intake than other groups (p = 0.001). After adjusted potential confounders, binary logistic regression analysis revealed that being severe nutritional risk was significantly reduced achieving full oral intake [Adjusted odds ratio: 0.419, 95% confidence interval (95%CI): 0.220—0.798] as well as pre-stroke functional dependency or incidence of pneumonia. Similar results were confirmed by Cox’s proportional hazard model for full oral intake (Adjusted hazard ratio: 0.587, 95%CI: 0.396—0.869). However, no significant correlation were observed between moderate or mild nutritional risks and achieving full oral intake.
Conclusion: Severe nutritional risk, but not moderate and mild nutritional risks, was independently predicts the possibility of achieving full oral intake in the convalescent stroke patients undergoing enteral nutrition.
Introduction
Stroke is second cause of death [1] and fourth leading cause of disease burden [2] in the world, thus it has been global health problem. Moreover, up to 62% of stroke patients experienced malnutrition[3] due to dysphagia [4], abnormal eating speed[5], stress response[6] and others. Previous study reported that malnutrition was correlated with poorer long-term mortality and functional dependency[7] among stroke patients. Similarly, we had been reported nutritional improvement was significantly correlated with functional recovery[8] in the convalescent stroke patients. Thus, nutrition assessment and management should be provided to all stroke patients. Above all causes of malnutrition, the predominant one in stroke patients may be dysphagia which had been reported strong correlation with malnutrition[4], aspiration and pneumonia[7]. Recovering swallowing function is therefore important issue for preventing life-threating complication such as malnutrition and pneumonia, promoting functional recovery, as well as regaining joy of eating in all stroke patients.
However, predictive factor of recovering swallowing function of convalescent stroke patients had not been well-established. Previous study suggested that activities of daily living on admission[9,10], age[9,10], and days after stroke onset[9] were correlated with swallowing recovery of convalescent stroke patients. On the other hands, whether nutrition status affects swallowing recovery had not been investigated to the best of our knowledge. Dysphagia may develop malnutrition due to inability to intake food or water appropriately [4], whereas few articles were mentioned that loss of muscle mass caused by nutritional deficit may worsened swallowing function [11].
Does nutritional risks predicts recovering swallowing function in the convalescent stroke patients? How much convalescent stroke patients needed to enteral nutrition were experienced malnutrition? To clarify these questions, we conducted retrospective study using data from multicenter cohort study for stroke patients with dysphagia in the Kaifukuki (convalescent) rehabilitation wards in Japan.
Materials and Methods
The APPLE study
The current study used the data which were picked out from the cohort study conducted by multicenter investigation “Algorithm for Post-stroke Patients to improve oral intake Level: APPLE”[12]. The APPLE study aimed to collect basic data to establish best nutritional care and swallowing therapy for stroke patients with dysphagia which investigated swallowing function, type of swallowing therapy, nutritional status and type of nutritional management. The APPLE Study was done at convalescent rehabilitation wards which covered by public health care insurance[13] in the 25 hospitals in Japan, February 2013. This study included consecutively stroke patients with dysphagia who admitted to the Kaifukuki rehabilitation wards between March 2011 and March 2013.
In Japan, public health insurance approved The Kaifukuki rehabilitation wards (KRWs) in 2000. The KRWs provided comprehensive, intensive, interdisciplinary in-patients rehabilitation for individuals who needs to rehabilitation after experienced diseases or injuries such as stroke, traumatic head injuries, and proximal femoral fractures. The interdisciplinary KRWs team composed on rehabilitation physician, rehabilitation nurses, physical therapist, occupational therapist, speech-language-hearing therapist, social worker, registered dietitian, care worker, dental hygienist, and other professionals. Intensive in-patients rehabilitation by therapist were provided up to 3h/day, for 180days in the KRW.
Study participants
The current study used 466 individuals from the APPLE study. Eligibility criteria were aged 65 years or older, receiving enteral nutrition and presence of recorded swallowing function on admission to the KRWs. Exclusion criteria were missing data (swallowing function at discharge, nutritional status, pre-stroke physical function status, history of stroke, and activities of daily living on admission), suspected dehydration (serum albumin ≥ 5.0g/dL), and presence of liver cirrhosis or renal diseases. Because hospitalization costs of stroke patients who elapsed over 60 days from stroke onset, or stayed in the KRWs for over 180 days were not covered by national insurance system in Japan, they were also excluded. Demographic and clinical characteristics data such as age, gender, diagnosis, stroke legion, history of stroke, days from stroke onset to admission to the KRWs, comorbidity, and pre-stroke mobility expressed by modified Rankin scale (MRS)[14] were collected. All patients received intensive rehabilitation including swallowing rehabilitation up to 3h/day by physiotherapist, occupational therapist, or speech language hearing therapist in accordance with facilities’ protocol. Nutrition support were also provided in each facilities’ approach. All participant were followed up swallowing function and other basically data at the time of discharging from the KRWs.
The current study had been approved by the local ethical committees of Hatsudai Rehabilitation Hospital. Since the study was retrospectively collected anonymous data, a waiver of consent was obtained from the ethics committee.
Assessing nutritional risks
Nutritional risks was assessed by the Geriatric Nutritional Risk Index (GNRI) [15] which comprised following parameters; serum albumin (Alb), actual body weight (ABW), and ideal body weight (IBW). Previous studies showed that GNRI predicts muscle dysfunction [16] and long-term mortality [17] among institutionalized elderly, as well as 6 month mortality in hospitalized elderly [18]. In the current study, GNRI was calculated as following equation.
GNRI = [14.89 × Alb(g/dL)] + 41.7 × [ABW(kg)/IBW(kg)]
IBW = [height(m)]2 × 22
Although ordinary method of GNRI used IBW calculated by Lorenz formula [15], this equation did not validated in Asian people. Therefore the present study used IBW calculated by Body Mass Index (BMI) 22kg/m2 which familiarly formula with Asian people and was demonstrated equivalency to Lorenz formula for calculated GNRI [19]. If ABW exceeded IBW, We set ABW/IBW = 1 based on the equation of GNRI[15]. In accordance with GNRI categories from original article[15], study participants were classified into 4 groups based on GNRI on admission; severe nutritional risk (GNRI < 82), moderate nutritional risk (82 to < 92), mild nutritional risk (92 to < 98), and no nutritional risk ( ≥ 98).
Assessing swallowing function
Swallowing function was assessed by Fujishima’s Swallowing Grade (FSG) which is 10-grade scale of swallowing ability[20]. FSG was ordinary tool of Food Intake LEVEL Scale[21] which was one of the most common swallowing assessment tool in Japan. The FSG were expressed as grading from 1 (completely disabled with swallowing function and not indicated swallowing rehabilitation) to 10 (normal swallowing function)(Table 1) in accordance with patients’ swallowing ability. Incidence of pneumonia was examined as complication of swallowing rehabilitation.
Assessing activities of daily living
Degree of swallowing dysfunction depended on stroke severity itself. The current study used the Functional Independence Measure (FIM) on admission which was validated measurement of ADLs [22] instead of stroke severity. Because previous study showed the FIM was significantly correlated with stroke severity [23], the FIM was deemed as potential confounder of the achievement of full oral intake.
Main outcome measurements
The primary outcome of the current study was whether participants achieved oral intake without enteral nutrition or not, because many stroke patients with dysphagia and their families need to ingest food without any artificial nutrition. According to FSG grading method, individuals with grade 7 or over could oral intake without tube feeding. Therefore, the study participants with grade 7 to 10 were deemed as possible to achieve oral intake without enteral nutrition.
Sample size calculation
Sample size was calculated using Power and Sample Size Calculation software (version 3.0, 2009, William D. Dupont, PhD, and Walton D. Plummer, Department of Biostatistics, Vanderbilt University). Because few reports were available regarding achievement of oral intake in the convalescent stroke patients, sample size was determined by the data from original APPLE study[12]. According to this article, over-all rate of full oral intake was 53.8%.We inferred that differences of 25% between participants with lowest nutritional risk and those with highest risk was clinically significant. If ratio of number of participants among groups is 1:1, 61 participants would be needed in each groups with 80% power and type 1 error probability of 0.05.
Statistical analysis
All statistical analyses were performed by the IBM SPSS Statistics (version21, 2012, IBM Corporation). Parametrical data were presented as mean (standard deviation), while non-parametrical data were expressed as median and interquartile range. Univariate analyses were performed using one-way ANOVA, Kruskal-Wallis test, chi-squared test and Fisher’s exact test to compare the outcomes between different GNRI categories, subsequently multiple comparison test were also performed if applicable. Binary logistic analysis was performed for achievement of full oral intake to adjust potential confounders such as age, gender, pre-stroke MRS, days from stroke onset to admit to rehabilitation wards, history of stroke, stroke lesions, total FIM on admission, FSG on admission, and GNRI categories. Kaplan-Meier curves and a log-rank test were used to assess crude correlation between cumulative rate of achieving full oral intake without enteral nutrition and different GNRI categories. Cox’s hazard model was used to estimate the adjusted hazard ratio. Significance was established as P-value < 0.05.
Results
Of total 466 individuals, 122 were excluded for missing data, 78 were excluded for not applied to public healthcare insurance policy of the KRWs, 9 were excluded for presence of dehydration, cirrhosis or renal diseases. Finally, 257 participants were included in the analyses. There were no differences in age, gender, GNRI, and FSG on admission between included and excluded individuals (Table2).
Table 3 showed the characteristics of the study participants. Based on GNRI on admission, 72 (28.0%) were classified into severe nutritional risk, 120 (46.7%) were moderate nutritional risk, 47 (18.3%) were mild nutritional risk, and 18 (7.0%) were no nutritional risk. Overall, 93.0% of participants had nutritional risks with any degrees. Participants with severe and moderate nutritional risks were older than other groups (p < 0.001), other characteristics did not difference among the groups except for GNRI components such as serum albumin and BMI (p < 0.001, p < 0.001, respectively).
Table 4 shows FSG and swallowing function among 4 groups. There were no differences in FSG on admission between groups. At discharge, participants with severe nutritional risk were demonstrated lowest FSG (p < 0.001) and lowest proportion of oral intake without enteral nutrition (p = 0.001). Additionally, participants with severe nutritional risk had higher proportion of pneumonia incidence than other groups (p =0.042).
Table 5 indicates that binary logistic regression analysis for achievement of oral intake without enteral nutrition (R2=0.438, p < 0.001). There was no multicollinearity between the variables. Being severe nutritional risk (GNRI < 82) was independently correlated with worse swallowing functional outcome [adjusted odds ratio (AOR): 0.419, 95% confidence interval (95%CI): 0.220—0.798, p = 0.008], whereas being moderate and mild nutritional risks did not show significant correlation with complete oral intake. Similarly, functional dependency before stroke onset [pre-stroke MRS, AOR: 0.634, 95%CI: 0.498—0.808], days from onset to admission [AOR: 0.950, 95%CI: 0.927—0.973] and pneumonia incidence [AOR: 0.248, 95%CI: 0.111—0.551] were negatively correlated with achievement of full oral intake without enteral nutrition. Conversely, FSG on admission and FIM on admission were positively affected to achievement of full oral intake.
According to the Kaplan-Meier method for cumulative rate of achievement of oral intake without enteral nutrition, participants with severe nutritional risk demonstrated lowest rate of achievement full oral intake (log lank: p < 0.001) (Figure 1). After adjusted covariate using Cox’s proportional hazard model, participants with severe nutritional risk showed lower rate of full oral intake [HR: 0.587, 95%CI: 0.396—0.869, p =0.008], whereas those with other nutritional risks had been shown no correlation for swallowing functional outcome(Table 6).
Discussion
In the present study, two clinical findings were observed. First, severe nutritional risk, but not moderate and mild nutritional risks, was independently predicts the possibility of achieving oral intake without enteral nutrition. Second, 93.0% of the convalescent stroke patients aged 65 years or over and underwent enteral nutrition had various degree of nutritional risks.
Severe nutritional risk, but not moderate and mild nutritional risks, was independently predicts the possibility of achieving complete oral intake without enteral nutrition. This findings had been observed both binary logistic regression analysis and Cox’s proportional hazard model after adjusted confounders. To the best of our knowledge, this was the first study which indicates nutritional risks affects regaining swallowing function among convalescent stroke patients underwent enteral nutrition. Previous studies reported that presence of dysphagia strongly correlated with being malnourished in the stroke patients[4,24], particularly in rehabilitation setting[4]. The results from the current study were in line with these articles. These results led us to consideration that reduced muscle mass caused by malnutrition may interrupt recovering swallowing function after stroke. To date, few articles available which had been mentioned altering muscle and nerve function occurred by protein-calorie malnutrition may promote dysphagia[11]. More recently, interesting with “sarcopenic dysphagia”, namely that decreased muscle mass and muscle function could deteriorate swallowing function, has been increasing[25]. We hypothesized that if stroke patients were severely malnourished, muscle mass regarding deglutition may also decrease separate from neurological deficit, and subsequently interrupt recovery of swallowing function. Alternatively, participants with severe nutritional risk had a higher incidence of pneumonia, and the incidence was negatively correlated with full oral intake in the present study. There were possibilities that stroke patient with severe nutritional risk easily experienced pulmonary infection, gastrointestinal bleeding, or other infection [7,24], thus may interrupted continuous swallowing rehabilitation.
93.0% of the convalescent stroke patients aged 65 years or over and underwent enteral nutrition had various degree of nutritional risks. Previous study indicated that prevalence of malnutrition in stroke patients was vary[3]. Regarding rehabilitation setting, prevalence of malnutrition ranges 6.1—49% in this review, whereas all of them did not use validated measurement of malnutrition[3]. In case of stroke patients with dysphagia, decreased visceral proteins had been observed in 50—72%[26]. Additionally, several study suggested that prevalence of malnutrition were increased during acute care hospital among stroke patients [6, 27]. Although GNRI did not expressed nutritional status but nutritional risks[15], there were possibility that almost all of the convalescent stroke patients undergoing enteral nutrition were at risk of malnutrition. The high prevalence of nutritional risks in the current study’s participants could be explained by dysphagia [4], abnormal eating speed[5], stress response[6] and others.
Undergoing nutrition support along with swallowing rehabilitation were necessary for stroke patients. In the current study, no differences were observed in energy and protein intakes regardless of degree of nutritional risks. On the other hands, approximately 60kcal/kg/day was required for gaining 50g/day of body cell mass among malnourished patients aged 80’s whereas only 25kcal/day was required among those aged 20’s[28]. More recently, Wakabayashi et al. proposed the concept of “Rehabilitation Nutrition” and demonstrated that individuals with disability who undergo rehabilitation would require more nutrients than those without rehabilitation[29]. It remains possibility that 1200kcal/day feeding among the current study’s participants with severe nutritional risk was relatively low to regain their body cell mass. Since we did not examine efficacy of nutritional support in this study, further intervention study will be needed that aggressive nutritional support with swallowing rehabilitation for stroke patients aiming for improving recovery of swallowing function.
The current study had several limitations. First, this study did not include all patients with dysphagia. Since median days from stroke onset to admission to rehabilitation wards were 37 days, participants who already recover swallowing function did not included in this study, while those with severer stroke who were not admitted to the rehabilitation wards were also excluded. Therefore, the results from present study cannot imply stroke patients with milder or severer dysphagia. Second, it is said that GNRI is not valid for convalescent stroke patients as indicator of nutritional status. Since GNRI had been showed poor agreement with the Mini Nutritional Assessment in the institutionalized elderly[30], whether nutritional status affects achievement of full oral intake remained unclear. Third, FSG did not confirmed validation as swallowing function indicator although the Food Intake LEVEL Scale which was validated measurement of severity of dysphagia had been developed based on FSG[21]. Fourth, the data about what types of nutrition support were implemented were not collected. As mentioned above, there were no differences in energy and protein intakes on admission among 4 groups. There was the possibility that amount of provided energy and nutrient were not enough to regain muscle mass in participants with severe nutritional risks. Whether appropriate nutritional support were provided or not might be confounder of this study.
In conclusion, this study clearly revealed following two findings. Severe nutritional risk, but not moderate and mild nutritional risks, was independently predicts the possibility of achieving complete oral intake without enteral nutrition. 93.0% of the convalescent stroke patients aged 65 years or over experienced moderate or severe nutritional risks. These findings add the value for convalescent stroke rehabilitation for individuals with dysphagia and nutritional risks. Further study will be needed aiming for resolving whether aggressive nutritional support in collaboration with comprehensive rehabilitation is effective for recovery of swallowing function in the convalescent stroke patients with severe malnutrition and dysphagia.
Acknowledgements
We are grateful to directors of the KRW association for advise and expertise to the study, researchers of APPLE (Algorithm for Post-stroke Patients to improve oral intake Level) for coordination of the original study, staffs of Hirosaki Stroke and Rehabilitation Center, Tokyo Bay Rehabilitation Hospital, Iwate Rehabilitation Center, Kinkai Rehabilitation Hospital, Iyo Hospital, Nagasaki Rehabilitation Hospital, Chikamori Rehabilitation Hospital, Tsukuba Memorial Hospital, Seiai Rehabilitation Hospital, Yufuin Hospital, Ohta General Hospital, Okinawa Rehabilitation Center Hospital, NTT East Corporation Izu Hospital, Sapporo Nishi-Maruyama Hospital, Yawata Medical Center, Ukai Rehabilitation Hospital, Kyoto Ohara Memorial Hospital, Showa University Fujigaoka Rehabilitation Hospital, Kawasaki Medical School Hospital, Nishinomiya Kyouritsu Rehabilitation Hospital, Morinomiya Hospital, Kumamoto Kinoh Hospital, Fujita Health University Nanakuri Sanatorium, Nishi-Hiroshima Rehabilitation Hospital, and Hatsudai Rehabilitation Hospital for data collection.
Statement of Authorship
SN: conception and design of the study, acquisition of data, analysis and interpretation of data, drafting the article, final approval of the version to be submitted.
TO: analysis and interpretation of data, revising the paper critically for important intellectual content, final approval of the version to be submitted.
MT, MU, MW, YK, KS: the conception and design of the study, acquisition of data, revising the paper critically for important intellectual content, final approval of the version to be submitted.
HN, NK: the conception and design of the study, revising the paper critically for important intellectual content, final approval of the version to be submitted.
Conflict of Interest Statement
The authors have no conflicts of interest.
Funding Sources
Data from the current study were picked out from the Multicenter Investigation “Algorithm for Post-stroke Patients to improve oral intake Level: APPLE” which funds were offered from EN Otsuka Pharmaceuticals Co. Ltd.
Figure legend
Figure1. Cumulative rate of achievement of full oral intake among 257 convalescent stroke patients with different nutritional risks. The participants were divided into 4 groups based on the Geriatric Nutritional Risk Index(GNRI) as follows: Severe (GNRI < 82), moderate (82 to < 92), mild (92 to < 98) and no risks (≥ 98). Significance was established by log lank test (p < 0.001).