Nutrition Support (Tube Feeding) as a Rehabilitation Intervention

Article Outline

• Abstract
• Methods
  • Patient Sample
  • Patient Variables
  • Process Variables and Interventions
  • Outcome Variables
  • Statistical Methods
• Results
  • Patient Variables
    • Demographic and health plan characteristics
    • Health and functional status characteristics
    • FIM and CSI
  • Process Variables
    • Length of stay
    • Therapy
  • Outcome Variables
    • Nutritional status
    • Depression
    • Discharge FIM and CSI scores
  • Site Variation
  • Regression Analyses
• Discussion
• Conclusions
• Acknowledgments
• References
• Copyright

Abstract 

James R, Gines D, Menlove A, Horn SD, Gassaway J, Smout RJ. Nutrition support (tube feeding) as a rehabilitation intervention.

Objective

To describe site variation in use of enteral feeding and its association with stroke rehabilitation outcomes, controlling for a variety of confounding variables.

Design

Prospective observational cohort study.

Setting

Six inpatient rehabilitation facilities in the United States.

Participants

Patients (N=919) from the Post-Stroke Rehabilitation Outcomes Project database with moderate or severe stroke who were discharged to home, community, or skilled nursing facility.

Interventions

Not applicable.

Main Outcome Measures

Change in total, motor, and cognitive FIM instrument scores and change in severity of illness.

Results

Monitoring of nutritional status and the frequency of tube-feeding interventions for patients with moderate and severe stroke varied significantly among sites. Patients with tube feeding had higher severity of illness and lower functioning on admission compared with patients who did not receive tube feeding. However, when we controlled for severity of illness, admission FIM score, and other important covariates, we found that patients with severe strokes who were tube fed for more than 25% of their stay had greater increases in total, motor, and cognitive FIM scores and greater improvement in severity of illness by discharge.

Conclusions

Nutrition support (tube feeding) is an effective therapy in rehabilitation service for patients with severe strokes and is associated with greater motor and cognitive improvements, even in patients with the most severe strokes.

Key Words:  Cerebrovascular accident , Nutrition , Outcome assessment (health care) , Rehabilitation , Severity of illness index , Tube feeding

MANY PATIENTS IN STROKE rehabilitation have some degree of malnutrition, either from prior poor food intake or from demands imposed by the stroke and hospitalization.¹ It is a known problem in stroke patients, with a prevalence of 16% on admission that increases to 22% to 35% at about 2 weeks and up to 50% at 2 to 3 weeks.², ³, ⁴ Malnourished people may lack the energy, stamina, strength, and mental focus to participate fully in therapies.⁵, ⁶ The effect of poor nutritional status on patients with stroke has been associated with poorer outcomes, such as reduced functional improvement, longer lengths of stay (LOSs), increased rates of complications, and mortality.⁷, ⁸ However, nutrition is often overlooked or not included as an important poststroke rehabilitation intervention, although it has been shown to be a foundation for effective therapy.⁹

Most inpatient rehabilitation facility patients have varying degrees of limitations related to eating, such as dysphagia, cognitive impairment, limited mobility, and movement.¹⁰ Often enteral (using the gastrointestinal system) feeding by mouth, although the most natural and desirable feeding method, is complicated by a lack of ability to self-feed, chewing or swallowing difficulties, poor appetite, and prior food preferences and patterns that are closely linked to malnutrition, weight loss, and decreased strength.¹¹ Altered-consistency diets (eg, puree, blended, or ground foods) and thickened liquids frequently are given to patients with chewing and swallowing difficulties but may be unappealing and contribute to poor intake. Nutritional support (enteral feeding by tube) provides adequate nutrition and is not affected by reduced appetite, swallowing problems, limited self-feeding, or disease conditions in which the gastrointestinal system is compromised. Tube feeding enhances continued physical recovery, supports earlier initiation of rehabilitation efforts, and may reduce aspiration related to dysphagia.¹²

Optimal timing to initiate tube feeding and for which patients have not been defined clearly.¹³, ¹⁴ The decision begins, in most cases, during the acute hospital stay before rehabilitation. A physician assesses each patient and if the patient is deemed to be at risk for aspiration based on diagnosis or poor tolerance of oral intake, a referral is made for a bedside (or clinical) swallow study. Approximately 60% of aspiration occurs without sensation or outward signs and symptoms and often would not be diagnosed from clinical examination.¹⁵ Therefore, if a patient appears at risk for silent aspiration or if further assessment is warranted, an instrumental swallow study is performed. If a safe diet can be established during a clinical or instrumental examination, with or without the use of compensatory strategies and diet modification, all attempts are made to maintain an oral route. If a patient is unable to meet nutritional needs and/or adequately protect his/her airway with oral intake, then temporary alternative measures for nutrition and hydration, such as tube feeding, are recommended.¹⁶, ¹⁷, ¹⁸ However, the time between trying to feed by oral intake and identifying that oral intake is unable to meet a patient’s nutritional needs delays the initiation of tube feeding, which may result in nutritional decline.¹⁹

Clinicians must weigh the perceived risk of silent aspiration caused by dysphagia against the known association of nasogastric tubes that are used commonly for short-term tube feedings with increased rates of aspiration pneumonia.²⁰, ²¹, ²², ²³, ²⁴ To truly weigh the risks, it is important to understand the full benefit of tube feeding on functional status and cognition.

The aim of this article was to describe the variation of nutritional support interventions, specifically tube feeding, used during poststroke inpatient rehabilitation and to determine the association of tube-feeding interventions with functional and clinical outcomes and patients’ abilities to participate in physical therapy (PT), occupational therapy (OT), and speech and language pathology (SLP) sessions. Our hypothesis was that patients who present with similar functional and cognitive levels (measured by FIM instrument scores) and severity of illness (measured by the Comprehensive Severity Index [CSI]) and who receive tube feeding during acute rehabilitation to provide additional nutritional support have better participation in therapies and, hence, better outcomes. This may be the first study to explore site variation in tube-feeding practice in inpatient rehabilitation and associate tube-feeding interventions with rehabilitation outcomes.²⁴

Back to Article Outline

Methods 

Patient Sample 

We used the Post-Stroke Rehabilitation Outcomes Project (PSROP) database²⁵ to explore tube feeding as a stroke intervention and its association with amount of therapy time and rehabilitation outcomes for patients with moderate or severe strokes. The PSROP database contains patient characteristics and process and outcome data for 1161 consecutively enrolled adult (>18y) stroke patients in 6 inpatient rehabilitation facilities across the United States. Stroke was defined as having an International Classification of Diseases, 9th Revision (ICD-9),²⁶ code of 430 to 438.99, 997.02, or 852 to 853 for the rehabilitation admission. Mild, moderate, and severe strokes were classified by case-mix groups (CMGs) of 100 to 103, 104 to 107, and 108 to 114, respectively. Patients with mild stroke were not included in the sample because of the low frequency of tube feeding (2.9%). In addition, patients with missing CMGs were not included. The PSROP sample was reduced from a total of 1161 to 919 patients, with 474 moderate strokes and 445 severe stroke patients who were discharged to home, the community, or a skilled nursing facility.²⁵

Functional status was determined using total and component (motor, cognitive) FIM scores on admission and discharge. Severity of illness was measured for 3 time periods: admission (first 24h of rehabilitation), discharge (last 24h of rehabilitation), and maximum (entire rehabilitation stay including admission and discharge), as measured by the CSI based on signs and symptoms extracted from the medical record.

Tube-feeding data were abstracted from inpatient rehabilitation charts after patient discharge. Data abstractors entered the date of, type of, and reason for tube placement. Start and stop times of an enteral formula, amount, and delivery frequency were collected from physician orders. Based on those data, the 919-patient sample of moderate and severe stroke patients was divided into 4 tube-feeding groups: (1) no tube feeding (n=758); (2) tube feeding for less than 25% of the rehabilitation stay (n=17) with tube feeding discontinued before discharge; (3) tube feeding for 25% or more of the rehabilitation stay with tube feeding discontinued before discharge (n=109); and (4) tube feeding for 100% of stay and discharged on tube-feeding support (n=35). The last group was defined by the following criteria: (1) a patient’s last ordered diet type was nothing by mouth or a speech and language pathologist was supervising all oral intake, (2) the patient was unable to swallow liquids or solids 24 hours before discharge, (3) a percutaneous endoscopic gastrostomy or other gastrostomy tube had been placed, and (4) the patient was discharged to a skilled nursing facility or home health agency. Group 4 patients never advanced to sufficient oral intake.

Patient Variables 

Patient variables include sex; age; race; payer source; previous stroke; hypertension diagnosis; diabetes diagnosis; depression diagnosis; any other mental disorder diagnosis; body mass index (BMI) on admission (categorized as underweight [<18.5kg/m²], normal [18.5–24.9kg/m²], overweight [25.0–29.0kg/m²], obese [≥30.0kg/m²]); type, location, and side of stroke; number of days from onset of stroke symptoms to admission to rehabilitation; admission motor and cognitive FIM scores; admission CSI continuous scores; and nutritional risk. Higher FIM scores indicate higher functioning levels, and higher CSI scores indicate increased severity of illness, or sicker patients. Nutritional risk was defined by using the first serum albumin or prealbumin level measurement. A patient was considered to be at high nutritional risk if the first serum albumin level was less than 2.5g/dL or if the first prealbumin level was less than 15mg/dL, moderate risk if the first albumin level was 2.5 to 3.0g/dL or the first prealbumin level was 15 to 20g/dL, low risk if the first albumin level was 3.0 to 3.5g/dL or the first prealbumin level was 20 to 25mg/dL, or no risk if the first albumin level was greater than 3.5g/dL or the prealbumin level was greater than 25mg/dL.

Process Variables and Interventions 

Process variables include inpatient rehabilitation LOS, tube feeding, and average number of therapy minutes per patient per day for PT, OT, and SLP sessions calculated by dividing the number of total minutes in each type of therapy by the LOS.

Outcome Variables 

Primary outcome measures include improvements in functioning using the difference between admission and discharge total, motor, and cognitive FIM scores. Secondary outcome measures include improvements in severity of illness (CSI) from admission to discharge (net medical improvement), increase in severity of illness (CSI) from admission to maximum, change in weight from the first to last weight measurments, mental disorder diagnosis (ICD-9 codes starting with 781, 294, 305, 309, 310, 311), depression diagnosis (ICD-9 codes 311−311.99), pneumonia diagnosis (ICD-9 codes starting with 480−486, 507), and improvement in nutritional status, defined as a decrease in nutritional risk from the first to last measured albumin and/or prealbumin level. Larger differences in FIM score and net medical improvement (CSI) scores signify more recuperation. Larger increases in CSI from admission to maximum signify a worsening condition at some time during the rehabilitation stay. Discharge destination was not used as an outcome variable in regression analysis because it was used to define the group of patients discharged with tube feeding.

Statistical Methods 

Descriptive statistics were used to compare patient characteristics, therapy interventions, and outcomes for the 4 tube-feeding groups, along with site variation in tube feeding and common nutrition assessment measures. Chi-square tests were used for categoric data, and analysis of variance was used for continuous data.

Ordinary least squares (OLS) regression was used to identify associations between patient characteristics and the tube-feeding groups with continuous outcome variables. To avoid multicollinearity, only variables with a correlation coefficient less than .50 were allowed to enter regression models for patients with moderate and severe stroke. Stepwise selection was used at a significance level of .10 to allow independent variables to enter and leave each model. The final models were the most parsimonious, with the maximum number of variables with significance levels less than .05. All analyses were performed with SAS statistical software, release 8.2.^a

Back to Article Outline

Results 

To understand the associations of tube feeding with outcomes, we begin by assessing similarities and differences in patients with moderate (CMG 104–107) and severe (CMG 108–114) strokes who received tube feeding compared with those who did not.

Patient Variables 

Patient characteristics by tube-feeding group are presented in table 1. Almost 6% of patients with moderate strokes and 30% of patients with severe strokes received tube feeding during rehabilitation.

Table 1. Patient Variables for Moderate and Severe Strokes by Tube-Feeding Group

Variables	Moderate Stroke (CMGs 104–107)					Severe Stroke (CMGs 108–114)
	No Tube Feeding (n=447)	Tube Feeding 1%–24% of Stay (n=4)	Tube Feeding ≥25% of Stay (n=18)	Tube Feeding 100% of Stay; Discharged with Tube Feeding (n=5)	Variation Significance (P)	No Tube Feeding (n=311)	Tube Feeding 1%–24% of Stay (n=13)	Tube Feeding ≥25% of Stay (n=91)	Tube Feeding 100% of Stay; Discharged with Tube Feeding (n=30)	Variation Significance (P)
Demographics
Male (%)	49.7	50.0	55.6	60.0	.931⁎	51.1	61.5	55.0	53.3	.829⁎
Age (y)	65.3±14.9	60.7±5.3	66.4±14.7	68.8±14.5	.861†	68.1±13.9	64.3±15.1	66.4±14.5	70.2±14.5	.428†
Race (%)					.342⁎					.023⁎
White	55.9	75.0	61.1	80.0		51.5	61.5	67.0	73.3
Black	29.1	25.0	11.1	0.0		28.6	30.8	14.3	10.0
Other	15.0	0.0	27.8	20.0		19.9	7.7	18.7	16.7
Payer (%)					.015⁎					.008⁎
Medicare	54.6	25.0	66.7	60.0		64.6	23.1	56.0	73.3
Medicaid	11.6	0.0	5.6	0.0		10.6	0.0	12.1	16.7
Commercial	32.0	75.0	22.2	20.0		21.2	61.5	26.4	6.7
Self-pay	1.3	0.0	0.0	20.0		2.6	15.4	4.4	3.3
Unknown/missing	0.5	0.0	5.6	0.0		1.0	0.0	1.1	0.0
Health and functional status characteristics
Nutritional risk by first albumin/prealbumin level (%)					<.001⁎					<.001⁎
No nutritional risk	11.0	0.0	5.6	20.0		10.9	7.7	4.4	3.3
Low nutritional risk	13.4	0.0	22.2	0.0		18.0	23.1	38.5	13.3
Moderate nutritional risk	7.8	25.0	11.1	0.0		9.6	30.8	18.7	23.3
High nutritional risk	3.6	0.0	22.2	40.0		10.9	7.7	17.6	20.0
No albumin/prealbumin measurement	64.2	75.0	38.9	40.0		50.5	30.8	20.9	40.0
BMI (%)					<.001⁎					.035⁎
Underweight (<18.5kg/m2)	2.9	0.0	11.1	40.0		1.6	7.7	8.8	6.7
Normal (18.5–24.9kg/m2)	43.6	25.0	16.7	0.0		38.6	38.5	34.1	46.7
Overweight (25–29kg/m2)	32.9	25.0	61.1	60.0		36.3	38.5	40.7	40.0
Obese (≥30kg/m2)	20.8	50.0	11.1	0.0		23.5	15.4	16.5	6.7
Stroke risk factors (%)
Previous stroke (exclude TIA)	27.5	0.0	44.4	40.0	.224⁎	33.4	30.8	20.9	26.7	.141⁎
Hypertension diagnosis	77.4	75.0	72.2	80.0	.961⁎	83.6	76.9	81.3	76.7	.721⁎
Diabetes diagnosis	27.1	0.0	27.8	40.0	.590⁎	33.8	15.4	35.2	33.3	.564⁎
Type of stroke (%)					.314⁎					.042⁎
Hemorrhagic	21.5	50.0	16.7	0.0		24.4	46.2	35.2	16.7
Ischemic	78.5	50.0	83.3	100.0		75.6	53.8	64.8	83.3
Side of stroke (%)					.193⁎					.557⁎
Right	46.1	25.0	33.3	40.0		44.7	38.5	42.9	33.3
Left	42.5	75.0	44.4	40.0		43.7	53.9	38.5	50.0
Bilateral	8.7	0.0	22.2	0.0		9.0	7.7	16.5	16.7
Unknown	2.7	0.0	0.0	20.0		2.6	0.0	2.2	0.0
Location of stroke (%)					.006⁎					.447⁎
Brainstem/cerebellum	22.6	25.0	22.2	0.0		15.4	0.0	11.0	13.3
Subcortical	34.0	0.0	27.8	20.0		43.7	53.9	36.3	33.3
Brainstem + subcortical	4.5	50.0	16.7	0.0		3.2	7.7	5.5	0.0
Lobar	33.8	25.0	33.3	60.0		32.5	38.5	41.8	50.0
Unknown	5.2	0.0	0.0	20.0		5.1	0.0	5.5	3.3
Days onset to rehabilitation admission	10.6±11.5	35.3±30.9	13.2±17.0	5.6±1.5	.002†	15.8±28.7	12.8±8.0	21.9±29.5	23.6±35.9	.186†
FIM
Admission total FIM score	72.5±9.8	61.5±9.0	67.4±10.4	64.2±16.2	.006†	47.7±11.7	32.9±9.1	35.8±10.7	33.2±11.3	<.001†
Admission motor FIM score	48.0±5.6	42.5±1.0	47.9±5.8	46.2±7.2	.238†	29.0±6.4	21.3±5.3	22.6±6.6	22.3±7.7	<.001†
Admission cognitive FIM score	24.5±7.1	19.0±8.5	19.4±7.1	18.0±9.4	.002†	18.7±7.5	11.6±4.5	13.2±6.7	10.8±5.8	<.001†
CSI
Maximum CSI indicator dysphagia (%)					<.001⁎					<.001⁎
Unable to swallow solids	6.0	25.0	22.2	20.0		7.7	15.4	17.6	6.7
Unable to swallow liquids	0.9	25.0	38.9	80.0		4.5	38.5	52.8	90.0
Dysphagia NOS	34.2	50.0	38.9	0.0		56.9	46.2	27.5	3.3
Normal swallowing	30.2	0.0	0.0	0.0		11.3	0.0	1.1	0.0
Unknown/missing	28.6	0.0	0.0	0.0		19.6	0.0	1.1	0.0
Admission CSI continuous score	15.4±9.7	16.8±9.1	24.4±13.3	34.6±11.5	<.001†	22.4±12.0	28.5±13.4	35.6±14.7	43.0±17.3	<.001†
Maximum CSI continuous score	22.2±12.9	26.0±16.6	41.1±17.3	56.2±30.5	<.001†	34.3±19.4	47.4±23.0	52.5±20.7	59.9±21.8	<.001†

NOTE. Values are mean ± standard deviation (SD) unless otherwise indicated.

Abbreviations: NOS, not otherwise specified; TIA, transient ischemic attack.

There were no significant differences in demographic characteristics for sex or age by tube-feeding groups for moderate or severe stroke patients. Racial differences were significant in the severe stroke group. Most patients discharged on tube-feeding support had Medicare insurance (60.0% and 73.3% for moderate and severe strokes, respectively).

Nutritional risk, as defined by albumin/prealbumin level, varied significantly by tube-feeding group for patients with moderate and severe stroke (P<.001). Thirty-three percent of patients with moderate stroke and 38% of patients with severe stroke who received any tube feeding were at moderate or high nutritional risk. For patients at high nutritional risk, 73% with moderate strokes and 60% with severe strokes received no tube feeding during rehabilitation. Many patients who received tube feeding (44% with moderate stroke, 26% with severe stroke) had no albumin or prealbumin measurements taken during rehabilitation.

In the moderate stroke group, more patients who received tube feeding were overweight or obese (70.4%) compared with those without tube feeding (53.5%, P<.001). For the severe stroke group there were proportionately more overweight and obese patients in the non–tube-feeding group (59.8%) than in the tube-feeding group (54.5%) (P=.035).

Stroke risk factors and stroke type and side were not significantly different among tube-feeding groups. Most tube-fed patients with moderate strokes had brainstem/subcortical and lobar strokes (P=.006).

Patients who received tube feeding differed significantly in function and severity of illness from patients who did not receive tube feeding. Tube-fed patients with moderate and severe stroke had significantly lower admission total and cognitive FIM scores and higher CSI admission scores. In addition, patients with severe stroke also had lower admission motor FIM scores.

When examining criteria used to determine CSI severity of illness for the tube-feeding groups, we found that the most severe symptoms of dysphagia (unable to swallow solids or liquids) were associated with tube-feeding groups. That is, almost all patients who were tube fed were unable to swallow solids or liquids or had dysphagia “not otherwise specified.” However, there were 31 patients with moderate stroke and 38 patients with severe stroke who were unable to swallow solids or liquids and were not tube fed (see table 1, Maximum CSI Indicator Dysphagia).

Process Variables 

Process variables by tube-feeding group are presented in table 2.

Table 2. Process Variables for Moderate and Severe Strokes by Tube-Feeding Group

Process Variables	Moderate Stroke (CMGs 104–107)					Severe Stroke (CMGs 108–114)
	No Tube Feeding (n=447)	Tube Feeding 1%–24% of Stay (n=4)	Tube Feeding ≥25% of Stay (n=18)	Tube Feeding 100% of Stay; Discharged With Tube Feeding (n=5)	Variation Significance (P)	No Tube Feeding (n=311)	Tube Feeding 1%–24% of Stay (n=13)	Tube Feeding ≥25% of Stay (n=91)	Tube Feeding 100% of Stay; Discharged With Tube Feeding (n=30)	Variation Significance (P)
Rehabilitation LOS	15.0±7.1	20.0±5.4	18.7±8.5	16.8±5.5	.079⁎	23.6±9.3	34.2±13.5	28.7±12.1	23.1±11.4	<.001⁎
PT
No. of days	10.6±6.0	16.3±3.3	13.1±7.3	12.6±5.4	.085⁎	16.8±8.1	20.3±12.3	19.7±10.7	15.6±8.2	.018⁎
No. of minutes	665±384	950±256	704±398	622±404	.491⁎	1065±576	1204±949	1140±741	862±451	.016⁎
No. of min/patient/day	44.3±16.7	49.4±17.3	36.6±15.4	36.4±22.4	.162⁎	44.1±13.7	35.3±18.2	39.1±14.6	39.0±14.4	.003⁎
OT
No. of days	9.2±5.8	13.3±4.7	12.3±6.8	10.4±9.3	.077⁎	14.5±8.0	16.7±12.0	16.9±10.5	13.9±7.7	.098⁎
No. of minutes	603±389	860±254	670±354	603±606	.530⁎	958±585	1088±869	1018±735	826±526	.441⁎
No. of min/patient/day	39.5±17.7	44.3±15.9	35.0±15.3	30.7±22.9	.454⁎	39.8±16.1	32.0±21.1	35.0±15.9	35.6±13.8	.025⁎
SLP
No. of days	7.8±6.2	13.5±4.5	13.7±6.1	11.0±3.2	<.001⁎	13.7±8.4	18.8±10.0	18.2±10.8	14.4±9.8	<.001⁎
No. of minutes	396±355	656±291	755±374	564±152	<.001⁎	749±530	1107±710	1024±687	755±550	<.001⁎
No. of min/patient/day	24.7±16.2	32.4±11.2	38.4±12.3	34.9±10.6	.004⁎	30.8±15.8	31.6±13.4	34.1±14.2	32.5±15.2	.384⁎

NOTE. Values are mean ± SD.

Rehabilitation LOSs in tube-feeding groups differed for patients with moderate and severe stroke. For patients with moderate stroke, the mean LOS for patients who were discharged with tube-feeding support (16.8d) was closer to the mean of patients not tube fed (15.0d), both of which were less than that for patients who were tube fed for 1% to 24% and 25% or more of their stay (20d and 8.7d, respectively; P=.079). For patients with severe stroke, again the mean LOS for patients who were discharged with tube-feeding support (23.1) was closer to the mean LOS for patients not tube fed (23.6d), both of which were shorter than for patients who were tube fed for 1% to 24% and 25% or more of their stay (34.2, 28.7d, respectively; P<.001).

For the moderate stroke group, tube feeding was not associated with the average number of days, total minutes, or minutes per patient per day spent in PT or OT sessions. However, for the severe stroke group, patients with tube feeding spent significantly fewer minutes per day in PT and OT compared with patients with no tube feeding (P<.025). For both stroke severity levels, SLP minutes per day were higher for patients receiving tube feeding, although the difference was significant for patients with moderate strokes only.

Outcome Variables 

Outcome variables by tube-feeding group are presented in table 3.

Table 3. Outcome Variables for Moderate and Severe Strokes by Tube-Feeding Group

Outcome Variables	Moderate Stroke (CMGs 104–107)					Severe Stroke (CMGs 108–114)
	No Tube Feeding (n=447)	Tube Feeding 1%–24% of Stay (n=4)	Tube Feeding ≥25% of Stay (n=18)	Tube Feeding 100% of Stay; Discharged With Tube Feeding (n=5)	Variation Significance (P)	No Tube Feeding (n=311)	Tube Feeding 1%–24% of Stay (n=13)	Tube Feeding ≥25% of Stay (n=91)	Tube Feeding 100% of Stay; Discharged With Tube Feeding (n=30)	Variation Significance (P)
Improvement in nutritional status based on first and last albumin/prealbumin levels (must have >1 serum albumin/prealbumin measurement) (%)	4.0	0.0	11.1	20.0	.163⁎	4.5	15.4	8.8	16.7	.022⁎
Any mental disorder diagnosis (%)	51.7	50.0	61.1	60.0	.862⁎	53.4	61.5	51.7	60.0	.809⁎
Depression diagnosis (%)	9.0	25.0	27.8	20.0	.037⁎	12.2	30.8	22.0	16.7	.047⁎
Weight change (last − first) (kg)	−0.3±4.4	2.6±7.1	−0.4±4.9	2.2±6.2	.389†	−1.8±6.1	−0.4±3.1	−0.8±4.4	−0.9±4.6	.444†
Discharge total FIM score	98.6±11.6	96.3±9.9	98.4±15.2	68.4±28.8	<.001†	77.2±19.3	78.0±11.2	69.9±23.2	47.6±21.4	<.001†
Increase in total FIM score (discharge − admission)	26.1±9.7	34.8±13.0	30.5±15.2	4.2±23.7	<.001†	29.4±14.4	45.1±12.6	33.9±20.3	14.7±15.7	<.001†
Discharge motor FIM score	70.4±8.8	70.3±10.0	73.1±12.6	49.6±22.2	<.001†	53.5±14.6	57.9±10.8	49.6±18.5	33.6±16.6	<.001†
Increase in motor FIM score (discharge − admission)	22.4±8.2	27.8±10.6	25.0±12.1	3.4±20.1	<.001†	24.5±12.3	36.6±13.0	26.9±16.8	11.7±13.1	<.001†
Discharge cognitive FIM score	28.2±5.9	26.0±7.4	25.1±6.1	18.8±10.0	<.001†	23.7±7.0	20.1±5.5	20.2±7.6	13.9±7.4	<.001†
Increase in cognitive FIM score (discharge − admission)	3.6±3.6	7.0±2.5	5.7±4.5	0.8±4.6	.009†	4.9±4.3	8.5±2.7	7.0±5.1	3.1±4.4	<.001†
Discharge CSI continuous score	5.5±6.3	6.25±2.6	10.8±6.8	26.4±13.8	<.001†	11.8±11.6	13.6±10.3	16.7±11.5	30.6±17.0	<.001†
Increase in severity (maximum − admission CSI score)	6.8±7.3	9.3±8.1	16.6±10.9	21.6±19.7	<.001†	11.9±11.4	18.8±18.3	16.8±12.4	16.9±13.7	.001†
Net medical improvement (admission − discharge CSI score)	9.9±8.0	10.5±10.5	13.6±11.3	8.2±5.2	.282†	10.6±9.3	14.9±7.3	18.9±11.3	12.4±10.3	<.001†
Pneumonia diagnosis (%)	1.8	0.0	5.6	20.0	.028⁎	5.5	15.4	22.0	20.0	<.001⁎
Discharge destination (%)					.307⁎					<.001⁎
Community discharge including home	95.1	100.0	100.0	80.0		77.8	69.2	68.1	43.3
Inpatient institutional discharge	4.9	0.0	0.0	20.0		22.2	30.8	31.9	56.7

NOTE. Values are mean ± SD unless otherwise indicated.

Significant differences in improvements in nutritional status (change in albumin or prealbumin levels from admission to discharge) for patients who did or did not receive tube feeding were found in the severe stroke group (P=.022). Patients with tube feeding showed more improvement.

We found that significantly more patients with moderate and severe stroke who were tube fed had a depression diagnosis, compared with patients without tube feeding (P=.037 vs P=.047, respectively).

At discharge, patients receiving tube feeding for 1% to 24% and 25% or more of their stay had similar discharge motor, cognitive, and total FIM scores as the non–tube-feeding group. However, these scores differed significantly from those of patients who were discharged on tube feeding (P<.001). Larger increases in FIM (total, motor, cognitive) scores were seen in the 2 tube-fed patient groups (tube fed for 1%–24% and ≥25% of stay); increases were significantly less for patients discharged on tube feeding. At discharge, patients receiving tube feeding for 1% to 24% and 25% or more of their stay had similar discharge CSI scores as the non–tube-feeding group. However, these scores were significantly lower than scores for patients who were discharged on tube feeding (P<.001).

Site Variation 

We examined variation in tube-feeding interventions and nutritional assessment measures across the 6 facilities (table 4).

Table 4. Site Variation in Nutrition Process Variables

Facility Processes	Moderate Stroke (CMGs 104–107)							Severe Stroke (CMGs 108–114)
	Site 1	Site 2	Site 3	Site 4	Site 5	Site 6	P	Site 1	Site 2	Site 3	Site 4	Site 5	Site 6	P
Patients (n)	73	91	47	112	59	92		82	66	119	46	91	41
Enteral tube feeding (%)							<.001⁎							<.001⁎
No tube feeding	97.3	100.0	95.7	95.5	72.9	97.8		75.6	84.9	76.5	87.0	35.2	73.2
Tube feeding 1%–24% of stay	1.4	0.0	0.0	1.8	1.7	0.0		4.9	3.0	2.5	2.2	2.2	2.4
Tube feeding ≥25% of stay	1.4	0.0	4.3	0.9	20.3	2.2		18.3	9.1	16.8	2.2	45.1	19.5
Tube feeding 100% of stay and discharged with tube feeding	0.0	0.0	0.0	1.8	5.1	0.0		1.2	3.0	4.2	8.7	17.6	4.9
Serum albumin measurements (%)							<.001⁎							<.001⁎
0	20.6	73.6	72.3	81.3	44.1	78.3		7.3	77.3	54.6	54.4	31.9	56.1
1	45.2	18.7	21.3	13.4	39.0	16.3		39.0	18.2	30.3	34.8	36.3	22.0
>1	34.3	7.7	6.4	5.4	17.0	5.4		53.7	4.6	15.1	10.9	31.9	22.0
Serum prealbumin measurements (%)							<.001⁎							<.001⁎
0	84.9	100.0	100.0	97.3	100.0	82.6		68.3	100.0	90.8	91.3	100	61.0
1	13.7	0.0	0.0	2.7	0.0	15.2		29.3	0.0	7.6	8.7	0.0	22.0
>1	1.4	0.0	0.0	0.0	0.0	2.2		2.4	0.0	1.7	0.0	0.0	17.1
Weight measurements (%)							<.001⁎							<.001⁎
0	5.5	5.5	10.6	20.5	0.0	8.7		4.9	0.0	6.7	21.7	0.0	4.9
1	39.7	14.3	29.8	13.4	8.5	23.9		17.1	4.6	24.4	4.4	4.4	19.5
>1	54.8	80.2	59.6	66.1	91.5	67.4		78.1	95.5	68.9	73.9	95.6	75.6
Mean no. of weight measurements ± SD	2.1±1.1	2.7±1.4	2.1±1.1	2.9±1.8	5.2±4.9	3.1±3.1	<.001†	2.7±1.2	4.6±2.1	2.7±1.5	3.5±1.8	8.7±7.6	4.1±4.0	<.001†

For patients with moderate stroke, the percentage of patients with tube feeding for 1% to 24% of stay varied little by site (0%–1.8%), whereas the number of those tube feeding for 25% or more of stay ranged from 0% to 20.3% of a site’s patients (P<.001). The number of patients discharged with tube feeding was very small in all sites, with only 5 patients in total.

The percentage of patients with severe stroke, who received tube feeding was statistically different by site (P<.001). Approximately 65% of patients in site 5 received tube feeding compared with 13% to 27% for the other sites. Over 17% of site 5 patients were discharged with tube feeding, which was double the percentage of the site with the second highest percentage. It is important to note that patients in site 5 had the lowest average admission total FIM score (lowest functioning) and the highest maximum CSI score (highest severity of illness), and the site had the second largest percentage of patients with severe stroke.²⁵

Nutritional assessment methods of monitoring serum albumin and serum prealbumin levels and weight varied significantly by site (P<.001 for each). For patients with moderate strokes, the percentage of patients with more than 1 albumin level measurement varied by site from 5.4% to 34.3%; the percentage with no serum albumin measurements ranged from 20.6% to 81.3%. For patients with severe strokes, the variation by site for no serum albumin measurements ranged from 7.3% to 77.3%. No serum prealbumin measurements were recorded at sites 2 and 5. The number of patients with moderate strokes who were weighed more than once varied by site from 54.8% to 91.5%; it varied from 73.9% to 95.6% for patients with severe strokes. Two sites weighed all patients with severe strokes at least once (range of no weight measurements, 0%–21.7%).

Regression Analyses 

Patients who received tube feeding for their entire rehabilitation stay and were discharged with tube feeding (group 4) were not included in OLS regression analyses, because they did not advance to oral intake or progressed little in other rehabilitation outcomes, as shown in table 3.

Table 5, Table 6 present significant variables in OLS regression analyses to predict the outcomes of increase in total FIM, motor FIM, and cognitive FIM scores and net medical improvement for patients with moderate and severe stroke. Patient characteristics and rehabilitation LOSs were controlled for in all analyses. Tube feeding for 1% to 24% of stay had only a borderline positive association (P=.08) with an increase in total FIM score in the moderate-stroke group and was not significant for the other outcomes (see table 5). For patients with severe stroke, tube feeding for 25% or more of stay had a significant positive association with each outcome; tube feeding for 1% to 24% of stay had a significant positive association with increase in total FIM and motor FIM scores only (see table 6).

Table 5. Improving Outcomes: Regression Analysis Results for Moderate Stroke (CMGs 104–107)^⁎

Independent Variables	Dependent Variables
	Increase Total FIM Score			Increase Motor FIM Score			Increase Cognitive FIM Score			Net Medical Improvement (Admission − Discharge CSI)
	Coeff	F	P	Coeff	F	P	Coeff	F	P	Coeff	F	P
Tube feeding 1%–24% of stay	3.991	3.1	.080
Partial R2		.006
Tube feeding ≥25% of stay
Partial R2
Age	−0.174	35.5	<.001	−0.147	35.0	<.001	−0.020	4.5	.035
White	1.774	4.3	.039	1.549	4.4	.036
Other race										1.360	4.2	.042
Incomplete low paraplegia or worse/complete hemiplegia	−3.605	4.6	.032				−1.386	6.5	.011
Incomplete hemiplegia										3.073	22.9	<.001
Stroke location: brainstem/cerebellum										1.442	6.6	.010
Comorbidity diabetes							0.629	4.1	.043
Admission motor FIM score	−0.493	37.0	<.001	−0.474	47.5	<.001
Admission cognitive FIM score	−0.247	16.1	<.001				−0.307	224.0	<.001
Admission CSI score							−0.039	7.1	.008	0.657	778.4	<.001
Rehabilitation LOS	−0.223	12.3	<.001	−0.225	17.8	<.001
R2		.186			.148			.357			.627

Abbreviation: Coeff, coefficient.

Table 6. Improving Outcomes: Regression Analysis Results for Severe Stroke (CMGs 108–114)^⁎

Independent Variables	Dependent Variables
	Increase Total FIM Score			Increase Motor FIM Score			Increase Cognitive FIM Score			Net Medical Improvement (Admission − Discharge CSI)
	Coeff	F	P	Coeff	F	P	Coeff	F	P	Coeff	F	P
Tube feeding 1%–24% of stay	13.290	9.8	.002	12.225	12.1	.006
Partial R2		.013			.018
Tube feeding ≥25% of stay	5.424	8.2	.005	4.497	8.4	.004	1.335	6.8	.009	3.567	11.0	.001
Partial R2		.011			.013			.008			.012
Age	−0.337	39.5	<.001	−0.332	55.3	<.001	−0.044	9.1	.003
Incomplete low paraplegia or worse/complete hemiplegia	−9.127	20.8	<.001	−7.419	19.5	<.001				−2.997	7.8	.005
Incomplete hemiplegia
Monoplegia UE/complete monoplegia LE/normal				4.902	4.4	.037
Stroke type: hemorrhagic							−1.126	6.5	.011	2.262	6.2	.013
Stroke side: right							0.861	4.8	.029
Stroke side: left				3.370	7.2	.008
Comorbidity diabetes	−4.790	9.8	.002	−3.695	8.3	.004	−1.256	9.5	.002	−1.781	4.4	.038
BMI underweight (<18.5kg/m2)										4.556	4.2	.042
Admission motor FIM score	0.249	4.4	.038				0.097	7.6	.006	0.187	6.6	.011
Admission cognitive FIM score				0.329	14.6	<.001	−0.288	87.1	<.001	0.136	4.7	.031
Admission CSI score							−0.044	6.7	.010	0.427	143.5	<.001
Rehabilitation LOS	0.210	7.6	.006				−1.344	6.5	.011	0.157	14.8	.001
R2		.201			.206			.301			.405

Abbreviations: LE, lower extremity; UE, upper extremity.

Back to Article Outline

Discussion 

One of the most telling findings in our study is the large site variation in nutritional assessment measures and use of tube feeding, which indicate that clinicians differ in the importance they place on nutrition as a stoke rehabilitation intervention. Because low functioning and higher severity of illness are associated with increased rates of tube feeding in our data, it is not surprising that site 5 had more patients who received tube feeding. However, sites 1, 2, and 3 had a similar mix of severe strokes (40%–68%) as site 5 (51%) but did not use tube-feeding interventions as frequently.

There does not appear to be an association between sites that have more patients with tube feeding and the measuring of serum albumin and prealbumin levels, because many patients with low albumin or prealbumin levels were not tube fed. Albumin and prealbumin values were obtained by retrospective chart review. A portion of site variability could be due to results not being reported clearly or documented in the chart at the time of review. In addition, other measures may have been used to assess nutritional risk, and clinicians may question whether high nutritional risk, as defined by low albumin or prealbumin level, is sufficient to initiate tube feeding. Studies have shown that malnutrition is associated with poorer outcomes, so it would seem that nutrition and nutritional assessment measures would be a standard of care implemented with similar frequencies across not only hospital organizations but also in all other health care–providing institutions. However, in our sample of rehabilitation centers, this is not the case.

A possible reason for the disparity in tube feeding is the lack of data showing benefits of nutrition or tube feeding in stroke rehabilitation. Dávalos et al³ found that nutritional status declined despite adequate enteral feeding after acute stroke. Dávalos concluded that caloric intake was not the only factor affecting malnutrition; therefore, the impact of nutritional intervention on stroke outcomes remained unclear.

Other articles suggest that the unclear relation may be due to the fact that serum albumin level, which usually is used in the definition of nutritional status, also is associated with underlying disease processes.²⁴, ²⁷ To minimize reliance on interpretation of albumin level, we related amount of tube feeding to rehabilitation outcomes directly. Tube feeding is associated with improved rehabilitation outcomes for patients with severe strokes, although not for patients with moderate strokes. One reason an effect may not have been detected in the moderate-stroke group is the small number of patients who received tube feeding—22 of 469 (4.7%) compared with 104 to 415 (25.1%) patients with severe stroke (omitting patients discharged with tube feeding). Intuitively, if tube feeding is beneficial for patients with severe strokes, it should be beneficial for patients with moderate strokes when deemed necessary by clinical staff.

One problem with tube feeding is the perceived risk of aspiration that may lead to mild symptoms, such as small increases in white blood cell count, higher temperatures, and lethargy or more severe symptoms associated with pneumonia.²⁸ These symptoms may decrease a patient’s ability to participate fully in therapy, thus affecting outcomes. The challenge with assessing aspiration risk is knowing what caused the aspiration. Feeding tubes often are placed when patients are at risk for aspiration, but as their swallowing improves, they begin therapeutic oral feedings while continuing tube feeding. Our data were not collected in a manner that would allow us to differentiate reasons for aspiration in these types of patients, so the cause-and-effect association of tube feeding and aspiration would be difficult to interpret. This limits our ability to understand if the improvement in outcomes associated with tube-fed patients is related to nutrition or to the reduction of adverse aspiration signs and symptoms that negatively affect therapy. However, tube-fed patients with moderate and severe strokes did have a significantly larger increase in severity of illness (maximum and admission CSI) and did have a pneumonia diagnosis code significantly more often during rehabilitation than patients who were not tube fed. We did not determine if pneumonia was diagnosed before or after tube placement. The general rule is that sicker patients have poorer outcomes, but this did not hold in our sample. It appears that tube feeding provided enough of a benefit to counteract the negative effects of aspiration that may cause greater increase in severity of illness and more pneumonia, leading to the assumption that nutrition is an intervention that improves outcomes.

The PSROP provided a first look at how nutrition may be an important component of stroke rehabilitation. The primary objective of this article was to describe and test our hypothesis that nutrition risk assessment and use of tube feeding is associated with better outcomes for patients with moderate and severe strokes. Our data indicate that use of tube feeding is associated with greater increases in function and decreases in severity of illness for patients with severe strokes, but clearly further examination is warranted.

Back to Article Outline

Conclusions 

This study identifies nutritional support (tube feeding) as an effective therapy in stroke rehabilitation. Use of nutritional intervention was associated with greater motor and cognitive improvements in the most severely impaired patients. The high level of site variation in the use of nutritional assessments and interventions for patients with stroke speaks loudly to the need for additional study of the role of nutrition as an important component of stroke rehabilitation. Further study, including a consistent measure of nutritional status during the rehabilitation stay, seems warranted.

Supplier

Back to Article Outline

Acknowledgments 

We acknowledge the role and contributions of the collaborators at each of the clinical sites represented in the Post-stroke Rehabilitation Outcomes Project: Brendan Conroy, MD (Stroke Recovery Program, National Rehabilitation Hospital, Washington, DC); Richard Zorowitz, MD (Department of Rehabilitation Medicine, University of Pennsylvania Medical Center, Philadelphia, PA); David Ryser, MD (Neuro Specialty Rehabilitation Unit, LDS Hospital, Salt Lake City, UT); Jeffrey Teraoka, MD (Division of Physical Medicine and Rehabilitation, Stanford University, Palo Alto, CA); Frank Wong, MD, and LeeAnn Sims, RN (Rehabilitation Institute of Oregon, Legacy Health Systems, Portland, OR); Murray Brandstater, MD (Loma Linda University Medical Center, Loma Linda, CA); and Harry McNaughton, MD (Wellington and Kenepuru Hospitals, Wellington, NZ).

Back to Article Outline

References 

1. Finestone HM , Greene-Finestone LS , Wilson ES , Teasell RW . Prolonged length of stay and reduced functional improvement rate in malnourished stroke and rehabilitation patients . Arch Phys Med Rehabil . 1996;77:340–345
   • View In Article
   • Abstract
   • Full-Text PDF (704 KB)
   • CrossRef
2. Axelsson K , Asplund K , Norberg A , Alafuoff I . Nutritional status in patients with acute stroke . Acta Med Scand . 1988;224:217–224
   • View In Article
   • MEDLINE
   • CrossRef
3. Dávalos A , Ricart W , Gonzalez-Huix F , et al.   Effect of malnutrition after acute stroke on clinical outcome . Stroke . 1996;27:1028–1032
   • View In Article
   • MEDLINE
   • CrossRef
4. Finestone HM , Greene-Finestone LS , Wilson ES , Teasell RW . Malnutrition in stroke patients on the rehabilitation service and at follow-up (prevalence and predictors) . Arch Phys Med Rehabil . 1995;76:310–316
   • View In Article
   • Abstract
   • Full-Text PDF (847 KB)
   • CrossRef
5. Albiin N , Asplund K , Bjermer L . Nutritional status of medical patients on emergency admission to hospital . Acta Med Scand . 1982;212:151–156
   • View In Article
   • MEDLINE
   • CrossRef
6. Reilly JJ , Hull SF , Albert N , Waller A , Bringardener S . Economic impact of malnutrition (a model system for hospitalized patients) . JPEN J Parenter Enteral Nutr . 1988;12:371–376
   • View In Article
   • MEDLINE
   • CrossRef
7. Dominioni L , Dionigi R . Immunological function and nutritional assessment . JPEN J Parenter Enteral Nutr . 1987;11:70S–72S
   • View In Article
   • MEDLINE
   • CrossRef
8. Stone SP , Whincup P . Standards for the hospital management of stroke patients . J R Coll Physicians Lond . 1994;28:52–58
   • View In Article
   • MEDLINE
9. Gines D . Nutrition management in rehabilitation . Rockville: Aspen; 1990;
   • View In Article
10. Newmann S , Miller M , Daniels L , Crotty M . Malnutrition in rehabilitation (prevalence and outcomes) . [abstract] Arch Phys Med Rehabil . 2004;85:21
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (131 KB)
    • CrossRef
11. Ickenstein GW , Kelly PJ , Furie KL , et al.   Predictors of feeding gastrostomy tube removal in stroke patients with dysphagia . J Stroke Cerebrovasc Dis . 2003;12:169–174
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (74 KB)
    • CrossRef
12. In:  Gottschlich MM ,  Fuhrman MP ,  Hammond KA ,  Holcombe BJ ,  Seidner DL editor. The science and practice of nutrition support (a case-based core curriculum) . Dubuque: Kendall Hunt Publisher; 2000;
    • View In Article
13. Barer DH . Lower cranial nerve motor function in unilateral vascular lesions of the cerebral hemisphere . [letter] BMJ . 1984;289:1621–1622
    • View In Article
    • CrossRef
14. Veis SL , Logemann JA . Swallowing disorders in persons with cerebrovascular accident . Arch Phys Med Rehabil . 1985;66:372–375
    • View In Article
    • MEDLINE
15. Smith C , Logemann J , Colangelo L , Rademaker A , Pauloski BR . Incidence and patient characteristics associated with silent aspiration in the acute care setting . Dysphagia . 1999;14:1–7
    • View In Article
    • MEDLINE
    • CrossRef
16. American Speech-Language Hearing Association . Role of the speech language pathologist in the performance and interpretation of endoscopic evaluation of swallowing (guidelines) . Rockville: ASHA; 2004;
    • View In Article
17. Hiss S , Dikeman K , Green J . In: Guidelines for speech-language pathologists performing videofluoroscopic swallowing studies . Vol 24: Rockville: American Speech-Language-Hearing Association; 2004;p. 77–92
    • View In Article
18. American Speech-Language-Hearing Association . In: Clinical indicators for instrumental assessment of dysphagia . Vol 3: Rockville: ASHA; 2003;p. 225–233 (guidelines) ASHA Desk Reference.
    • View In Article
19. Norton B , Homer-Ward M , Donnelly MT , Long RG , Holmes GK . A randomized prospective comparison of percutaneous endoscopic gastrostomy and nasogastric tube feeding after acute dysphagic stroke . BMJ . 1996;312:13–16
    • View In Article
    • MEDLINE
    • CrossRef
20. Wolfsen H , Kozarek R . Percutaneous endoscopic gastrostomy. Ethical considerations . Gastroenterol Clin North Am . 1993;2:259–271
    • View In Article
21. Ciocon JO , Silverstone FA , Graver M , Foley CJ . Tube feeding in elderly patients . Arch Intern Med . 1988;148:429–433
    • View In Article
    • MEDLINE
22. Wicks C , Gimson A , Vlavianos P , et al.   Assessment of the percutaneous endoscopic gastrostomy feeding tube as part of an integrated approach to enteral feeding . Gut . 1992;33:613–616
    • View In Article
    • MEDLINE
    • CrossRef
23. Rees RG , Payne-James JJ , King C , Silk DB . Spontaneous transpyloric passage and performance of ‘fine bore’ polyurethane feeding tubes (a controlled clinical trial) . JPEN J Parenter Enteral Nutr . 1988;12:469–472
    • View In Article
    • MEDLINE
24. Gariballa SE , Parker SG , Taub N , Castleden CM . Influence of nutritional status on clinical outcome after stroke . Am J Clin Nutr . 1998;68:275–281
    • View In Article
    • MEDLINE
25. Gassaway J , Horn SD , DeJong G , Smout RJ , Clark C , James R . Applying the continuous practice improvement approach to stroke rehabilitation (methods used and baseline results) . Arch Phys Med Rehabil . 2005;86(12 Suppl 2):S16–S33
    • View In Article
    • MEDLINE
26. In:  Hart AC ,  Schmidt KM ,  Aaron WS editor. ICD-9-CM code book . Reston: St. Anthony’s Publishing; 1999;
    • View In Article
27. Baker JP , Detsky AS , Wesson DE , et al.   Nutrition assessment . N Engl J Med . 1982;306:969–972
    • View In Article
    • MEDLINE
    • CrossRef
28. Abernathy G , Heizer W , Holcombe B , et al.   Efficacy of tube feeding in supplying energy requirements of hospitalized patients . JPEN J Parenter Enteral Nutr . 1989;13:387–391
    • View In Article
    • MEDLINE

• a SAS Institute Inc, 100 SAS Campus Dr, Cary, NC 27513.