Timing of Percutaneous Endoscopic Gastrostomy for Acute ...stroke.ahajournals.org/content/strokeaha/early/2016/12/13/STROKE... · gastrostomy (PEG) feeding compared ... Key Words:

1

Dysphagia, or difficulty swallowing, is a common occur-rence after stroke, occurring in ≥65% of stroke patients.1

About half of the patients with dysphagia improve within 2 weeks and 15% continue to have persistent dysphagia after 1 month. Prior studies evaluating early percutaneous endoscopic gastrostomy (PEG) feeding compared with nasogastric feed-ing and delayed PEG< if needed, have shown that there was no benefit and, possibly, a borderline increase in the absolute risk of death or poor outcome with a policy of early PEG place-ment (48% within 3 days) compared with early nasogastric feeding followed by PEG if needed.2 Because of this and the limited data available to predict outcomes among dysphagic stroke patients, providers often use time-limited trials of arti-ficial nutrition. The American Stroke Association Guidelines for the Early Management of Patients With Acute Ischemic Stroke recommend nasogastric feeding over PEG tube feeding for a period of ≥2 to 3 weeks after stroke onset.3 Prior studies have found significant variation in the surgical placement of

feeding tubes after stroke, with many hospital factors influenc-ing the decision to place a PEG tube.4 However, little is known about the current practices in time to PEG tube placement for stroke admissions and which factors may impact the timing of PEG placement. In the setting of health system pressures to reduce length of stay and maintain efficiency, we hypoth-esized that many patients receive their PEG tubes early on in their hospital stay. We sought to describe the timing of surgi-cal feeding tube placement for patients hospitalized for acute ischemic stroke, to identify characteristics associated with early PEG tube placement, and to assess variability between hospital practices in timing of PEG placement.

Methods

DatabaseData were obtained from the Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project, Nationwide Inpatient

Background and Purpose—Stroke guidelines recommend time-limited trials of nasogastric feeding prior to percutaneous endoscopic gastrostomy (PEG) tube placement. We sought to describe timing of PEG placement and identify factors associated with early PEG for acute ischemic stroke.

Methods—We designed a retrospective observational study to examine time to PEG for ischemic stroke admissions in the Nationwide Inpatient Sample, 2001 to 2011. We defined early PEG placement as 1 to 7 days from admission. Using multivariable regression analysis, we identified the effects of patient and hospital characteristics on PEG timing.

Results—We identified 34 623 admissions receiving a PEG from 2001 to 2011, 53% of which received the PEG 1 to 7 days from admission. Among hospitals placing ≥10 PEG tubes, median time to PEG for individual hospitals ranged from 3 days to over 3 weeks (interquartile range 6–8.5 days). Older adult age groups were associated with early PEG (≥85 years versus 18–54 years: adjusted odds ratio 1.68, 95% confidence interval 1.50–1.87). Those receiving a PEG and tracheostomy were more likely to receive the PEG beyond 7 days, and these patients were more often younger compared with PEG only recipients. Those admitted to high-volume hospitals were more likely to receive their PEG early (≥350 versus <150 hospitalizations; adjusted odds ratio 1.26, 95% confidence interval 1.17–1.35).

Conclusions—More than half of the PEG recipients received their surgical feeding tube within 7 days of admission. The oldest old, who may benefit most from time-limited trials of nasogastric feeding for ≥2 to 3 weeks, were most likely to receive a PEG within 7 days. (Stroke. 2017;48:00-00. DOI: 10.1161/STROKEAHA.116.015119.)

Key Words: feeding tube ◼ gastrostomy tube ◼ ischemic stroke ◼ PEG ◼ variation

Timing of Percutaneous Endoscopic Gastrostomy for Acute Ischemic Stroke

An Observational Study From the US Nationwide Inpatient Sample

Benjamin P. George, MD, MPH; Adam G. Kelly, MD; George P. Albert; David Y. Hwang, MD; Robert G. Holloway, MD, MPH

Received August 15, 2016; final revision received November 7, 2016; accepted November 18, 2016.From the Department of Neurology, University of Rochester Medical Center, NY (B.P.G., A.G.K., R.G.H.); College of Arts and Science, University of

Rochester, NY (G.P.A.); and Department of Neurology, Division of Neurocritical Care and Emergency Neurology, Yale School of Medicine, New Haven, CT (D.Y.H.).

The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA. 116.015119/-/DC1.

Correspondence to Robert Holloway, MD, MPH, Department of Neurology, University of Rochester Medical Center, 601 Elmwood Ave, Box 673, Rochester, NY 14642. E-mail [email protected]

© 2016 American Heart Association, Inc.

Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.116.015119

Original Contribution

by guest on June 14, 2018http://stroke.ahajournals.org/

Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from


Dow

nloaded from

mailto:[email protected]

http://stroke.ahajournals.org/










2 Stroke February 2017

Sample, 2001–2011.5 Detailed description of the Nationwide Inpatient Sample database can be found in the online-only Data Supplement.

Hospitalization SelectionWe selected adult (age ≥18 years) acute ischemic stroke admissions from January 1, 2001, to December 31, 2011, using International Classification of Diseases-Ninth Revision-Clinical Modification (ICD-9-CM) primary diagnosis codes 433.x1, 434.x1, 436 for isch-emic stroke (Figure 1). Observations with missing data for vari-ables such as age, sex, race, length of stay, and time to procedure were excluded from the analysis. We further identified patients who underwent surgical placement of a PEG tube (Agency for Healthcare Research and Quality Clinical Classification Software code 71). Patients receiving their PEG tube >30 days from the time of admis-sion (n=449) were considered outliers who may be more prone to coding error and were therefore excluded. Description of patient and hospital characteristics used can be found in methods in the online-only Data Supplement.

Outcome MeasuresWe defined early PEG placement as those receiving a PEG tube within 1 to 7 days from admission, which marks the median time to PEG placement for US ischemic stroke patients in the sample and the time within which the majority of stroke patients received a PEG in the PEG versus nasogastric trial.2 We determined covariate asso-ciation with early PEG tube placement using descriptive, univariate, and multivariable analyses. Only variables with <10% missing data that were significant in the univariate analysis were included in the multivariable model. Variability between hospitals in average time to PEG and rates of early PEG were studied among all stroke admis-sions within hospitals placing ≥10 PEG tubes for stroke patients. We performed a small subgroup analysis examining the effect of age and receiving a tracheostomy on early versus later (>7 days from

admission) PEG placement. Finally, we investigated the association between early PEG tube placement and length of stay among survi-vors, inpatient mortality, and discharge disposition.

Sensitivity AnalysisTwo sensitivity analyses were performed. Given the high level of missing data on race, we examined the results of the multivariable model predicting early PEG tube placement on the subset of admis-sions with race available. In addition, 4.5% of patients received a PEG on either day 1 or 2 of admission, and given that some of these patients may have been readmitted for the sole purpose of PEG place-ment, we examined the multivariable model predicting early PEG placement, excluding this minority of patients.

Statistical AnalysisWe compared categorical variables using chi-squared tests and con-tinuous variables with Student t test or Wilcoxon rank-sum tests as indicated. Trends were assessed using a Cochran–Armitage test of trends. Multivariable logistic regression analysis was used to deter-mine factors associated with early PEG tube placement. Hospital practice variability was measured using the coefficient of variation, range and interquartile range. Statistical significance was set a pri-ori at P<0.05. Analyses were performed using Stata v13.0. (Stata, College Station, TX). All Healthcare Cost and Utilization Project data have been deidentified, and therefore, the analysis is exempt from federal regulations for the protection of human research participants, and institutional review board approval was not necessary. The analy-sis is in compliance with the Healthcare Cost and Utilization Project data use agreement.

ResultsA total of 34 623 PEG tube placements for acute ischemic stroke adult inpatients were identified from 2001 to 2011 (Figure 1). The average time to PEG tube placement for stroke hospitalizations was 8.4 days from the time of admission (median 7 days, interquartile range =5–11). Over half (53.5%) of stroke hospitalizations receiving a PEG tube had the pro-cedure early or within 7 days of admission. The rate of early PEG did not change substantially over time (P trend =0.38), ranging between 52% and 55% from 2001 to 2011 (Figure I in the online-only Data Supplement). However, among those receiving a PEG tube during the time period examined, there was a trend toward young and fewer Medicare patients, fewer females, greater comorbidity, and greater representation of teaching, urban, and higher volume hospitals (Table I in the online-only Data Supplement).

In the unadjusted analysis (Table 1), patient characteristics, including older age, female sex, white race, Medicare insur-ance, and low- to middle-income origins, were associated with early PEG placement. Hospital characteristics associated with early PEG in the unadjusted analysis included nonteaching, rural, and low stroke volume hospitals, as well as those hos-pitals that intubate a smaller proportion of stroke patients and have lower hospital charges on average for stroke patients.

There were 2936 stroke hospitalizations (8.5%) that received a tracheostomy in addition to a PEG tube, which rep-resents just 0.3% of the entire ischemic stroke sample from 2001 to 2011. Over 90% of those receiving both a PEG and a tracheostomy were <85 years old, compared with 73% of those with a PEG only (P<0.001). Nearly one quarter of the youngest stroke patients with a PEG tube (age 18–54 years) had a PEG and tracheostomy compared with just 2% of those

Figure 1. Patient selection method. PEG indicates percutaneous endoscopic gastrostomy. *Key variables include age, sex, death, hospital location, and teaching status. †There were 22 840 acute ischemic strokes with PEG tubes from hospitals with ≥10 PEG tubes for stroke. ‡There were 29 455 acute ischemic strokes with PEG tubes with complete data on race.


Dow

nloaded from


George et al Timing of PEG Tube Placement for Ischemic Stroke 3

Table 1. Patient and Hospital Characteristics for Stroke Hospitalizations Receiving an Early Versus Later PEG Tube

Patient Characteristics

No. of Hospitalizations, % Univariate Analysis

Early PEG (1–7 days) Later PEG (>7 days)Unadjusted Odds

Ratio 95% CI

Total 18 526 (100) 16 097 (100)

Age, y

18–54 1148 (6.2) 1567 (9.7) Reference

55–64 1900 (10.3) 2130 (13.2) 1.21 1.10–1.34

65–74 3614 (19.5) 3560 (22.1) 1.39 1.27–1.51

75–84 6486 (35.0) 5430 (33.7) 1.63 1.50–1.77

≥85 5378 (29.0) 3410 (21.2) 2.15 1.97–2.35

Sex

Male 8084 (43.6) 7424 (46.1) Reference

Female 10 442 (56.4) 8673 (53.9) 1.11 1.06–1.15

Race*

White 10 442 (56.4) 8352 (51.9) Reference

Black 3253 (17.6) 3417 (21.2) 0.76 0.72–0.81

Hispanic 1559 (8.4) 1442 (9.0) 0.86 0.80–0.93

Other 973 (5.3) 1086 (6.7) 0.72 0.65–0.79

Missing 2299 (12.4) 1800 (11.2) NA

Insurance

Medicare 14 736 (79.5) 11 840 (73.6) Reference

Private 2094 (11.3) 2019 (12.5) 0.83 0.78–0.89

Medicaid 1085 (5.9) 1501 (9.3) 0.58 0.54–0.63

Self pay, no charge, other 611 (3.3) 737 (4.6) 0.67 0.60–0.74

Patient income

Low to middle income 14 001 (75.6) 11 768 (73.1) Reference

High income 4130 (22.3) 3951 (24.5) 0.88 0.84–0.92

Stroke Charlson Comorbidity Index

Score: 0 4314 (23.3) 3622 (22.5) Reference

Score: 1–2 10 631 (57.4) 8998 (55.9) 0.99 0.94–1.05

Score: ≥3 3581 (19.3) 3477 (21.6) 0.86 0.81–0.92

Tracheostomy 636 (3.4) 2300 (14.3) 0.21 0.19–0.23

Atrial fibrillation 6197 (33.5) 5846 (36.3) 0.88 0.84–0.92

Hospital characteristics

Teaching status

Nonteaching 11 039 (59.6) 7695 (47.8) Reference

Teaching 7487 (40.4) 8402 (52.2) 0.62 0.60–0.64

Location

Rural 1942 (10.5) 1201 (7.5) Reference

Urban 16 584 (89.5) 14 896 (92.5) 0.69 0.64–0.74

Stroke volume

30–149 8526 (46.0) 7083 (44.0) Reference

150–249 4889 (26.4) 4195 (26.1) 0.97 0.92–1.02

(Continued )


Dow

nloaded from



≥85 years of age (P<0.001; Figure II in the online-only Data Supplement). The average time to PEG placement for those with a tracheostomy was 12.8 days compared with 8.0 days without a tracheostomy (P<0.001). In addition, those receiv-ing a tracheostomy had the procedure an average of 12.6 days from the time of admission. For those receiving a PEG only, the oldest old (≥85 years) were most likely to receive a PEG early (P<0.001), whereas among those receiving both a PEG and tracheostomy, the oldest old were least likely to receive a PEG early (P=0.02; Figure 2).

Among 1401 hospitals placing ≥10 PEG tubes (n=22 840 admissions), the median time to PEG placement for hospitals

ranged from 3 days to 23 days from admission, and the median rate of early PEG placement was 56.4% (interquartile range 41.1%–66.7%; Figure 3).

Within the multivariable regression model (Table 2), the strongest predictor of early PEG was older age, with those ≥85 years old (versus 18–54 years; adjusted odds ratio 1.70, 95% confidence interval 1.46–1.98) having the greatest odds of early PEG. Receiving a tracheostomy on the same admis-sion was the strongest predictor for receiving a later PEG tube (versus no tracheostomy; adjusted odds ratio 0.27, 95% con-fidence interval 0.24–0.29). The model also demonstrated an association with later PEG placement and receiving care within

250–349 2700 (14.6) 2385 (14.8) 0.94 0.88–1.00

≥350 2411 (13.0) 2434 (15.1) 0.82 0.77–0.88

Hospital intubation use

1st quartile 2169 (11.7) 1223 (7.6) Reference

2nd quartile 4161 (22.5) 2740 (17.0) 0.87 0.79–0.93

3rd quartile 5738 (31.0) 4629 (28.8) 0.70 0.65–0.76

4th quartile 6458 (34.9) 7505 (46.6) 0.49 0.45–0.52

Mean hospital charges†

1st quartile 2483 (13.4) 1664 (10.3) Reference

2nd quartile 3865 (20.9) 2786 (17.3) 0.93 0.86–1.01

3rd quartile 5263 (28.4) 4303 (26.7) 0.82 0.76–0.88

4th quartile 6768 (36.5) 7194 (44.7) 0.63 0.59–0.68

Missing 147 (0.8) 150 (0.9) NA

CI indicates confidence interval; NA, not available; and PEG, percutaneous endoscopic gastrostomy.*Race was missing in 12% of cases and 42% of hospitals (n=1956) had incomplete data on race.†Mean hospital charges represent the average of all hospital charges for stroke admissions within an individual hospital in a

given year. Charges were adjusted using the Medical Consumer Price Index for all urban consumers prior to averaging to account for inflation. The highest and lowest 1% were excluded to avoid influence of outliers.

Figure 2. Early PEG tube placement for acute ischemic stroke hospitalizations (A) without and (B) with tracheostomy by age group. PEG indicates percutaneous endoscopic gastrostomy. *Later PEG placement was defined as >7 days from the time of admission, whereas early PEG placement is defined as 1 to 7 days from admission.

Table 1. Continued


No. of Hospitalizations, % Univariate Analysis

Early PEG (1–7 days) Later PEG (>7 days)Unadjusted Odds

Ratio 95% CI


Dow

nloaded from



teaching hospitals, hospitals with high utilization of intuba-tion among stroke patients, and hospitals with high average hospital charges for stroke patients. However, in the adjusted model, hospitals with larger stroke volume were associated with early PEG (≥350 versus <150 hospitalizations; adjusted odds ratio 1.26, 95% confidence interval 1.17–1.35). The results of sensitivity analyses demonstrate that inclusion of race in the model and exclusion of individuals receiving PEG tubes within the first 2 days of admission did not substantially change the significance or values of other covariates (Tables II and III in the online-only Data Supplement).

When examining the effect of early PEG tube placement as a predictor of various outcome measures, we found those receiving an early PEG tube were more likely to have shorter length of stay and had greater odds of discharge to home or acute rehabilitation (Table 3). The odds of discharge to skilled nursing or long-term care were less for early PEG during a stroke hospitalization. However, when these same outcomes were examined in the subset of patients ≥85 years (n=8788), there was no association with discharge to home, acute reha-bilitation, and skilled nursing or long-term care (Table IV in the online-only Data Supplement). There was no association with early versus later PEG tube placement and inpatient mor-tality for all patients.

DiscussionAmong acute ischemic stroke patients hospitalized in the United States who receive a PEG tube, we found that more than half (53%) receive their surgical feeding tube within 7 days of admission. Similar to findings from other smaller studies,6 this suggests a mismatch between practice reality and stroke guidelines, which recommend time-limited trials of nasogastric feeding for a period of ≤2 to 3 weeks prior to placing a surgical feeding tube.7 We found that age was the greatest determinant of early PEG tube placement. In fact, the oldest PEG tube recipients (≥85 years) had 1.7× the adjusted

odds of receiving an early PEG compared with the youngest (18–54 years), and there were no differences in discharge dis-position for those receiving their PEG early versus later in this subset of older patients.

Our study demonstrates that ≈2 out of every 3 patients ≥85 years old with a stroke and PEG tube receive their PEG within 7 days of their hospitalization. Prior studies have found a greater likelihood of PEG tube complications and hospital readmis-sions with advanced age.8 Also, from the FOOD trials (Feed or Ordinary Diet), a comparison of early feeding via PEG versus nasogastric tube and delayed PEG, if needed, demonstrated that early PEG feeding was associated with a borderline increased risk of death or poor outcome.2 Furthermore, the study showed that among patients randomized to nasogastric feeding, only 28% later received a PEG tube.2 This suggests that time-lim-ited trials, which may represent a period of ≤2 to 3 weeks, may affect whether a patient will receive a feeding tube—an impor-tant decision because many patients have strong preferences regarding their use.9,10 In fact, 1 study recently found that >50% of seriously ill hospitalized patients viewed relying on a feed-ing tube to live as equal to or worse than death, although this study did not include stroke patients and preference may change as death draws near.11 Also, surgical feeding tube placement is a marker for an overall poor prognosis, with only about half of patients receiving a PEG tube in prior studies surviving to 6 months, and 85% of those surviving had moderate to severe disability.2 Therefore, these results should provide pause when considering a surgical feeding tube for older adults early on in their hospitalization because this group of patients is potentially the most likely to benefit from time-limited trials of nasogastric feeding for ≤2 to 3 weeks, allowing for greater time spent on patient/family discussions.

For younger PEG recipients, receiving an early PEG tube was associated with discharge to home or acute rehabilita-tion. Eligibility for these dispositions may ultimately drive the decision to place a surgical feeding tube early in the

Figure 3. Variation in (A) median time to PEG and (B) rate of early (1–7 day) PEG placement for stroke hospitalizations within individual hospitals. IQR indicates interquartile range; PEG, percutaneous endoscopic gastrostomy; and SD, standard deviation. *Median time to PEG and early PEG tube rates were calculated for stroke admissions within 1401 individuals in hospital placing at least 10 PEG tubes in the data set. The volume of PEG placement among these hospitals ranged from 10 to 53 admissions (median=15; IQR=12–20). The x axis intersects the y axis at the median value.


Dow

nloaded from



hospitalization for the appropriate patient. However, the decision should be weighted with patient/family preferences after an informed discussion. If artificial nutrition is needed beyond 2 to 3 weeks, PEG feeding has traditionally been the preferred approach because of its association with improved patient safety and outcomes; however, these studies are out-dated and small.12,13 Given the shortcomings of this evidence, the time may still be right for more prolonged alternatives to PEG tubes for select patients, and the use of nasogastric feeding during rehabilitation should be explored further given advancements in outpatient management.14–16 For all patients with dysphagia early on in their hospitalization, eliciting pref-erences and arranging alternative means of feeding (eg, care-ful hand-feeding) if desired is an essential aspect of poststroke care.10,17

In our study, the youngest patients receiving a PEG tube after an acute stroke more often additionally required a tra-cheostomy during their admission compared older patients (24% in those aged 18–54 years versus 2% in those aged ≥85 years). This could be a result of more severe strokes, leading to respiratory failure with intensive care unit stays in younger patients receiving a PEG tube, whereas older patients requiring a PEG tube may more often experience dysphagia in isolation because of age-related effects on swallow functioning18; however, our data set was limited in the inability to assess stroke severity based on National Institutes of Health stroke scale or imaging, so this is diffi-cult to confirm. For these patients receiving a tracheostomy, only ≈20% received a PEG tube early in their hospitaliza-tion. The timing of PEG and tracheostomy seemed to coin-cide at ≈12 days from the time of admission and seem to be similar in terms of timing for critical care patients, in gen-eral, including those having trauma or sepsis.19 However, the optimal timing of tracheostomy for patients mechani-cally ventilated with stroke remains unclear and is an active area of research.20

Hospitals with larger stroke volume were associated with early PEG placement in the multivariable model, yet, later PEG placement in the unadjusted analysis. This reversal is likely because of confounding factors, such as disease sever-ity (eg, greater intubations and tracheostomies) within high-volume centers. The tendency to place PEG tubes early in the adjusted model could reflect the efficiency within large stroke centers, which translates to reduced length of stay and more

Table 2. Multivariable Logistic Regression Model for Predictors of Early (1–7 Days) PEG Tube Placement in Acute Ischemic Stroke Hospitalizations

Patient Characteristics Adjusted Odds Ratio 95% CI

Age, y

18–54 Reference

55–64 1.09 0.98–1.21

65–74 1.12 1.01–1.25

75–84 1.28 1.16–1.43

≥85 1.68 1.50–1.87

Sex

Female (vs male) 0.98 0.93–1.02

Insurance

Medicare Reference

Private 1.09 1.01–1.18

Medicaid 0.79 0.72–0.87

Self pay, no charge, other

0.86 0.76–0.98

Patient income

Low to middle income

Reference

High income 0.85 0.80–0.89

Stroke Charlson Comorbidity Index

Score: 0 Reference

Score: 1–2 0.97 0.92–1.03

Score: ≥3 0.86 0.81–0.93

Tracheostomy (vs none) 0.27 0.24–0.29

Atrial fibrillation (vs none)

0.79 0.75–0.82

Hospital characteristics

Teaching (vs nonteaching)

0.70 0.66–0.73

Urban location (vs rural)

0.98 0.90–1.07

Stroke volume

30–149 Reference

150–249 1.08 1.02–1.15

250–349 1.22 1.14–1.30

≥350 1.26 1.17–1.35

Hospital intubation use

1st quartile Reference

2nd quartile 0.90 0.83–0.99

3rd quartile 0.80 0.73–0.87

4th quartile 0.67 0.61–0.73

Mean hospital charges*

1st quartile Reference

2nd quartile 1.00 0.92–1.08

(Continued )

3rd quartile 0.96 0.89–1.04

4th quartile 0.86 0.80–0.93

Analysis was conducted on 33 554 admissions with complete availability for the above variables. The model demonstrates acceptable fit with Hosmer–Lemeshow goodness of fit, P=0.50. C-statistic=0.71. CI indicates confidence interval; and PEG, percutaneous endoscopic gastrostomy.

*Mean hospital charges represent the average of all hospital charges for stroke admissions within an individual hospital in a given year. Charges were adjusted using the Medical Consumer Price Index for all urban consumers prior to averaging to account for inflation. The highest and lowest 1% were excluded to avoid influence of outliers.

Table 2. Continued

Patient Characteristics Adjusted Odds Ratio 95% CI


Dow

nloaded from



process-driven care.21 The unanswered question is, does qual-ity of patient/family discussions regarding feeding decisions suffer at the expense of efficiency? Strategies that incorporate family discussions into processes of care22 and standardization of the approach with tools such as decision aids23 may help to prevent overlooking key dialogues that patient’s and their families may find important.24

In addition to overall high utilization of early PEG tube placement, we found wide variation in practices between hospitals. For hospitals placing at least 10 PEG tubes in the Nationwide Inpatient Sample, the median time to PEG place-ment ranged from ≈3 days to 3 weeks from the time of admis-sion. While 75% of hospitals placed PEG tubes 1 to 7 days from admission between 42% and 67% of the time for stroke patients, most of the variation came from hospitals in the upper and lower quartiles (Figure 3). Indeed, some provid-ers or institutions may lack awareness of the potential ben-efits of delayed PEG placement, leading to greater variation. Additional clinical factors that may alter the timing of PEG tube placement include anticoagulation and also intolerance or frequent interruptions of nasogastric feeding. Further research is needed to better understand reasons for this significant prac-tice variation, beyond the patient and hospital predictors used in our analysis.

Our study has several limitations. We used an administra-tive data set, which relies on International Classification of Diseases-Ninth Revision coding to identify patients, which lack granularity and may contain some inaccuracies in identi-fying diagnoses and procedures. In addition, the data set does not have robust clinical data to describe stroke severity, func-tional disability, and size or location of the infarct, which may importantly impact the timing of PEG tube placement. In addi-tion, we used time from admission to PEG placement, which may be different from stroke onset—although the majority of stroke patients present within hours of stroke onset.25 We used risk adjustment using a Charlson Comorbidity Index tailored for stroke outcomes studies, which may be a mean-ingful covariate to capture general illness severity in stroke patients. We accounted for the presence of atrial fibrillation, a likely marker for embolic disease. In addition, we included

tracheostomy placement indicating critical respiratory failure, which in this population is most commonly attributed to severe neurological disease as a result of a stroke. Complications, such as aspiration pneumonia, hemorrhagic transformation, and other comorbid diagnoses, were not used in this analysis of predictors because it would not be possible to understand the true sequence of events using these data and, therefore, dif-ficult to interpret the effect on PEG tube placement. Perhaps, most importantly, our study does not investigate how often a discussion of PEG tube placement occurs for stroke admis-sions, nor does it provide insight into the quality of patient/family discussions, which lead to the shared decision to place a surgical feeding tube.

Despite these limitations, we provide insight into the fac-tors that could influence the timing for a discussion of PEG tube placement and may help to guide the development and implementation of novel patient engagement strategies (eg, decision aids) in practice. Our findings further emphasize the need to diagnose patient preferences to allow for a shared decision that is in the patient’s best interest. Appreciating indi-vidual circumstances and the optimal timing of discussions will help to understand and standardize how providers guide feeding decisions for patients after a stroke.

AcknowledgmentsDrs George and Holloway had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis. Drs George, Kelly, and Holloway are responsible for study concept and design. Drs George and Holloway are respon-sible for acquisition of data. Drs George, Hwang, G.P. Albert, Drs Kelly, and Holloway are responsible for analysis and interpretation of data. Drs George, Hwang, G.P. Albert, Drs Kelly, and Holloway are responsible for critical revision of the article for important intellec-tual content. Dr George and G.P. Albert are responsible for statistical analysis. Dr Holloway is responsible for study supervision.

DisclosuresDr Hwang received funding from the American Brain Foundation, Neurological Care Society, National Institute on Aging, and Apple Pickers Foundation. Dr Holloway is Associate Editor of Neurology Today and consultant for MCG Guidelines. The other authors report no conflicts.

Table 3. Association of Early (1–7 Days) PEG Tube Placement With Patient Outcomes

Outcomes Early PEG (1–7 days) Later PEG (>7 days) Adjusted Odds Ratio (95% CI)*

Survivor length of stay, mean (SD) 10.4 (8.2) 20.5 (14.8) 0.57 (0.56–0.58)

Inpatient mortality, No. (%) 1191 (6.4) 1209 (7.5) 0.92 (0.85–1.01)

Discharge to home, No. (%) 2204 (11.9) 1649 (10.24) 1.15 (1.07–1.24)

Discharge to acute rehab, No. (%)† 2360 (12.7) 1806 (11.2) 1.27 (1.18–1.36)

Discharge to SNF/LTC, No. (%)† 8712 (47.0) 7876 (48.9) 0.83 (0.79–0.87)

The effect of early vs later PEG placement was studied on various outcomes using the outcome as the dependent variable in a multivariable regression analysis adjusting for patient demographics (age, sex, payer, income), comorbidities (Charlson index, atrial fibrillation, tracheostomy placement), and hospital characteristics (teaching status, urban/rural location, stroke volume, intubation use, and mean charges for stroke patients). CI indicates confidence interval; LTC, long-term care facility; PEG, percutaneous endoscopic gastrostomy; SD, standard deviation; and SNF, skilled nursing facility.

*Logistic regression analysis was used, and adjusted odds ratios were reported except for length of stay, where negative binomial regression analysis was used and an incident rate ratio was reported.

†For discharge to acute rehab or SNF/LTC, 16.8% of data were missing because of incomplete reporting of detailed disposition, and therefore, these missing values were excluded from the respective analyses.


Dow

nloaded from



References 1. Martino R, Foley N, Bhogal S, Diamant N, Speechley M, Teasell

R. Dysphagia after stroke: incidence, diagnosis, and pulmo-nary complications. Stroke. 2005;36:2756–2763. doi: 10.1161/01.STR.0000190056.76543.eb.

2. Dennis MS, Lewis SC, Warlow C; FOOD Trial Collaboration. Effect of timing and method of enteral tube feeding for dysphagic stroke patients (FOOD): a multicentre randomised controlled trial. Lancet. 2005;365:764–772. doi: 10.1016/S0140-6736(05)17983-5.

3. Jauch EC, Saver JL, Adams HP Jr, Bruno A, Connors JJ, Demaerschalk BM, et al; American Heart Association Stroke Council; Council on Cardiovascular Nursing; Council on Peripheral Vascular Disease; Council on Clinical Cardiology. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare profession-als from the American Heart Association/American Stroke Association. Stroke. 2013;44:870–947. doi: 10.1161/STR.0b013e318284056a.

4. George BP, Kelly AG, Schneider EB, Holloway RG. Current practices in feeding tube placement for US acute ischemic stroke inpatients. Neurology. 2014;83:874–882. doi: 10.1212/WNL.0000000000000764.

5. Introduction to the nationwide inpatient sample (NIS). Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project Web site. http://www.hcup-us.ahrq.gov/db/nation/nis/NIS_Introduction_ 2011.jsp. Accessed July 21, 2016.

6. Wilmskoetter J, Simpson AN, Simpson KN, Bonilha HS. Practice pat-terns of percutaneous endoscopic gastrostomy tube placement in acute stroke: are the guidelines achievable [published online ahead of print July 27, 2016]? J Stroke Cerebrovasc Dis. doi: 10.1016/j.jstrokecerebrovasdis.2016.07.017.

7. Holloway RG, Arnold RM, Creutzfeldt CJ, Lewis EF, Lutz BJ, McCann RM, et al; American Heart Association Stroke Council, Council on Cardiovascular and Stroke Nursing, and Council on Clinical Cardiology. Palliative and end-of-life care in stroke: a state-ment for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2014;45:1887–1916. doi: 10.1161/STR.0000000000000015.

8. Wilmskoetter J, Simpson KN, Bonilha HS. Hospital readmissions of stroke patients with percutaneous endoscopic gastrostomy feeding tubes. J Stroke Cerebrovasc Dis. 2016;25:2535–2542. doi: 10.1016/j.jstrokecerebrovasdis.2016.06.034.

9. Blackmer J. Tube feeding in stroke patients: a medical and ethical per-spective. Can J Neurol Sci. 2001;28:101–106.

10. Holloway R. Protein-energy supplementation of normal hospital diet did not improve outcomes after recent stroke. ACP J Club. 2005;143:36.

11. Rubin EB, Buehler AE, Halpern SD. States worse than death among hospitalized patients with serious illnesses. JAMA Intern Med. 2016;176:1557–1559. doi: 10.1001/jamainternmed.2016.4362.

12. Norton B, Homer-Ward M, Donnelly MT, Long RG, Holmes GK. A ran-domised prospective comparison of percutaneous endoscopic gastros-tomy and nasogastric tube feeding after acute dysphagic stroke. BMJ. 1996;312:13–16.

13. Park RH, Allison MC, Lang J, Spence E, Morris AJ, Danesh BJ, et al. Randomised comparison of percutaneous endoscopic gastrostomy and nasogastric tube feeding in patients with persisting neurological dyspha-gia. BMJ. 1992;304:1406–1409.

14. Puntis JW. Nutritional support at home and in the community. Arch Dis Child. 2001;84:295–298.

15. Daniel R, Didier P, Hélène P. A 3-month at-home tube feeding in 118 bulimia nervosa patients: a one-year prospective survey in adult patients. Clin Nutr. 2014;33:336–340. doi: 10.1016/j.clnu.2013.05.016.

16. Pearce CB, Collett J, Goggin PM, Duncan HD. Enteral nutrition by naso-jejunal tube in hyperemesis gravidarum. Clin Nutr. 2001;20:461–464. doi: 10.1054/clnu.2001.0484.

17. Creutzfeldt CJ, Holloway RG, Curtis JR. Palliative care: a core compe-tency for stroke neurologists. Stroke. 2015;46:2714–2719. doi: 10.1161/STROKEAHA.115.008224.

18. Sura L, Madhavan A, Carnaby G, Crary MA. Dysphagia in the elderly: management and nutritional considerations. Clin Interv Aging. 2012;7:287–298. doi: 10.2147/CIA.S23404.

19. Mehta AB, Cooke CR, Wiener RS, Walkey AJ. Hospital Variation in Early Tracheostomy in the United States: a population-based study. Crit Care Med. 2016;44:1506–1514. doi: 10.1097/CCM.0000000000001674.

20. Bösel J, Schiller P, Hook Y, Andes M, Neumann JO, Poli S, et al. Stroke-related Early Tracheostomy Versus Prolonged Orotracheal Intubation in Neurocritical Care Trial (SETPOINT): a randomized pilot trial. Stroke. 2013;44:21–28. doi: 10.1161/STROKEAHA.112.669895.

21. Svendsen ML, Ehlers LH, Ingeman A, Johnsen SP. Higher stroke unit volume associated with improved quality of early stroke care and reduced length of stay. Stroke. 2012;43:3041–3045. doi: 10.1161/STROKEAHA.111.645184.

22. Gay EB, Pronovost PJ, Bassett RD, Nelson JE. The intensive care unit family meeting: making it happen. J Crit Care. 2009;24:629.e1–629.e12. doi: 10.1016/j.jcrc.2008.10.003.

23. Cox CE, Wysham NG, Walton B, Jones D, Cass B, Tobin M, et al. Development and usability testing of a Web-based decision aid for families of patients receiving prolonged mechanical ventilation. Ann Intensive Care. 2015;5:6. doi: 10.1186/s13613-015-0045-0.

24. Spatz ES, Krumholz HM, Moulton BW. The new era of informed con-sent: getting to a reasonable-patient standard through shared decision making. JAMA. 2016;315:2063–2064. doi: 10.1001/jama.2016.3070.

25. Agyeman O, Nedeltchev K, Arnold M, Fischer U, Remonda L, Isenegger J, et al. Time to admission in acute ischemic stroke and transient ischemic attack. Stroke. 2006;37:963–966. doi: 10.1161/01.STR.0000206546.76860.6b.


Dow

nloaded from


HollowayBenjamin P. George, Adam G. Kelly, George P. Albert, David Y. Hwang and Robert G.

Observational Study From the US Nationwide Inpatient SampleTiming of Percutaneous Endoscopic Gastrostomy for Acute Ischemic Stroke: An

Print ISSN: 0039-2499. Online ISSN: 1524-4628 Copyright © 2016 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Stroke published online December 13, 2016;Stroke.

http://stroke.ahajournals.org/content/early/2016/12/13/STROKEAHA.116.015119World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://stroke.ahajournals.org/content/suppl/2016/12/13/STROKEAHA.116.015119.DC1Data Supplement (unedited) at:

http://stroke.ahajournals.org//subscriptions/

is online at: Stroke Information about subscribing to Subscriptions:

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

document. Permissions and Rights Question and Answer process is available in the

Request Permissions in the middle column of the Web page under Services. Further information about thisOnce the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.Strokein Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:


Dow

nloaded from

http://stroke.ahajournals.org/content/early/2016/12/13/STROKEAHA.116.015119

http://stroke.ahajournals.org/content/suppl/2016/12/13/STROKEAHA.116.015119.DC1

http://www.ahajournals.org/site/rights/

http://www.lww.com/reprints

http://stroke.ahajournals.org//subscriptions/


Timing of Percutaneous Endoscopic Gastrostomy (PEG) for Acute Ischemic

Stroke: An Observational Study from the US Nationwide Inpatient Sample

Online Supplemental Appendix

Table of Contents

(In order of reference)

Supplemental Methodology page 1 Figure I. Trends in time to PEG placement, 2001-2011 page 3 Table I. Trends in patient and hospital characteristics among ischemic stroke patients receiving a PEG tube, 2001-2011 page 3 Figure II. Proportion of stroke patients with a PEG tube that also received a tracheostomy by age group ... page 4 Table II. Multivariable logistic regression model for predictors of early (1-7 days) PEG tube placement in acute ischemic stroke hospitalizations including race as a predictor .. page 5 Table III. Multivariable logistic regression model for predictors of early (1-7 days) PEG tube placement in acute ischemic stroke hospitalizations (excluding those with PEG tubes placed on day 1 or day 2) ........................................................... page 6 Table IV. Association of early (1-7 days) PEG tube placement with patient outcomes in those age 85 .................................................... page 7

1

Supplemental Methodology

Database Description

The Nationwide Inpatient Sample (NIS) is the largest all-payer nationwide administrative

database of hospital inpatient stays in the US, which is used by researchers and policy makers

to identify, track, and analyze national trends in health care utilization, access, charges, quality,

and outcomes. The NIS contains 100% of admissions from a 20% sample of nonfederal US

hospitals up to 2011. The sample includes discharge-level data from over 8 million

hospitalizations from approximately 1,000 hospitals annually across the United States. The

sampling for the NIS database is stratified based on hospital urban/rural location, teaching

status, bed size, ownership, and US region. In 2012, the NIS was redesigned to capture a 20%

sample of discharges from 100% of HCUP hospitals. Given the net loss of information on the

hospital level during the redesign, we did not perform the analysis using data from 2012-2013

despite its availability.e1

Patient characteristics

Demographic and socioeconomic factors, including age, insurance payer, and patient income,

were identified from the primary dataset. Race had a high degree of missing data compared to

other variables, and therefore, we excluded race from the multivariable analysis. We identified

chronic comorbidities using secondary diagnosis codes included within the NIS and accounted

for 16 conditions within a modified Charlson Comorbidity Index tailored for stroke outcomes

studies.e2 In addition, atrial fibrillation was identified due to its importance for risk of large

cardioembolic stroke. We also identified admissions receiving a tracheostomy (AHRQ Clinical

2

Classification Software code 34) due to its importance in the care for some acutely ill stroke

patients.

Hospital Characteristics

Hospital characteristics within the sample include teaching status and urban/rural location as

reported by the American Hospital Association Annual Survey of Hospitals. We also calculated

annual stroke volume (categories: 30-149, 150-249, 250-349, ≥350 admissions), hospital

intubation use among its stroke patients (stratified into quartiles), and average hospital charges

among stroke admissions (stratified into quartiles). Charges were adjusted using the Medical

Consumer Price Index for all urban consumers prior to averaging to account for inflation.e3 In the

calculation of a hospital’s average charges, the highest and lowest 1% were excluded to avoid

influence of outliers.

Supplemental References

e1. George BP, Schneider EB, Hwang DY. Consequences of the Nationwide Inpatient

Sample Redesign for Studies Examining Between-Hospital Practice Variation. Critical Care

Medicine 2016;44:e1261.

e2. Goldstein LB, Samsa GP, Matchar DB, Horner RD. Charlson Index comorbidity

adjustment for ischemic stroke outcome studies. Stroke 2004;35:1941-1945.

e3. Consumer Price Index [online]. Available at: http://www.bls.gov/cpi/. Accessed July 21,

2016.

3

Figure I. Trends in time to PEG placement, 2001-2011

Table I. Trends in patient and hospital characteristics among ischemic stroke patients receiving a PEG tube, 2001-2011*

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011


Age, mean (sd) 76.3 (11.7) 76.5 (11.8) 75.9 (12.2) 75.6 (12.5) 75.8 (12.5) 74.9 (12.8) 75.4 (12.8) 74.5 (13.1) 74.8 (13.1) 73.8 (13.3) 74.1 (13.3) Female, % 57.7 57.5 57 55.4 54.5 54.7 55.1 53.5 53.6 53.5 54.7 Medicare, % 80.6 81.5 79.8 78.4 79.4 76.5 76.3 73.9 72.7 72.6 73.2 Weighted Charlson Sum, mean (sd) 1.36 (1.36) 1.47 (1.38) 1.39 (1.39) 1.45 (1.39) 1.45 (1.44) 1.66 (1.59) 1.74 (1.62) 1.68 (1.59) 1.88 (1.75) 1.90 (1.81) 2.05 (1.92)

Hospital Characteristics

Teaching hospital, % 38.6 42.5 44.8 43.1 37.1 47.2 44.5 48.8 51.2 53.8 51.5 Urban hospital, % 87.1 88.9 89.2 90.4 89.8 92.1 90.6 91.7 94.1 92.3 93.8 Hospital Stroke Volume, mean (sd) 146 (102) 149 (107) 145 (113) 144 (109) 147 (115) 146 (117) 146 (103) 164 (132) 171 (127) 178 (132) 187 (150) Hospital PEG Volume, mean (sd) 8.2 (7.1) 8.4 (8.6) 8.2 (8.5) 7.8 (7.7) 7.6 (7.9) 7.7 (7.6) 7.5 (7.9) 8.5 (9.4) 8.6 (8.7) 8.9 (9.4) 8.9 (10.2)

* All trends are significant (P-trend<0.01) except hospital PEG volume (P-trend=0.09).

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Percent of stroke patients receiving a PEG tube 1-7 days from admission

4

Figure II. Proportion of stroke patients with a PEG tube that also received a tracheostomy by age group

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

18-54yrs 55-64yrs 65-74yrs 75-84yrs ≥85yrs

PEG and trachestomy

PEG only

5

Table II. Multivariable logistic regression model for predictors of early (1-7 days) PEG tube placement in acute ischemic stroke hospitalizations including race as a predictor*

Patient Characteristics Adjusted

Odds Ratio (95% CI) P value Age (yrs)

18-54 Reference NA 55-64 1.11 (0.99-1.25) 0.07

65-74 1.10 (0.98-1.23) 0.11 75-84 1.23 (1.10-1.38) <0.001 ≥85 1.62 (1.44-1.83) <0.001

Sex Female (vs male) 0.99 (0.94-1.04) 0.57

Race White Reference NA

Black 0.86 (0.80-0.91) <0.001 Hispanic 1.06 (0.97-1.15) 0.22 Other 0.89 (0.80-0.98) 0.02

Charlson Comorbidity Index† Score: 0 Reference NA

Score: 1-2 0.97 (0.92-1.03) 0.50 Score: 3 or more 0.89 (0.82-0.95) <0.01

Insurance Medicare Reference NA

Private 1.09 (1.00-1.19) 0.06 Medicaid 0.79 (0.71-0.88) <0.001 Self Pay, No Charge, or Other 0.85 (0.74-0.97) 0.02

Patient Income Low to middle income Reference NA

High income 0.84 (0.79-0.89) <0.001 Tracheostomy (vs none) 0.27 (0.24-0.30) <0.001 Atrial Fibrillation (vs none) 0.79 (0.75-0.83) <0.001 Hospital Characteristics

Teaching (vs non-teaching) 0.70 (0.67-0.74) <0.001 Urban (vs rural) 0.94 (0.85-1.04) 0.27 Stroke Volume

1st quartile Reference NA 2nd quartile 1.10 (1.03-1.17) <0.01

3rd quartile 1.22 (1.14-1.32) <0.001 4th quartile 1.30 (1.21-1.41) <0.001

Hospital Intubation Use 1st quartile Reference NA

2nd quartile 0.91 (0.82-1.00) 0.05 3rd quartile 0.82 (0.75-0.91) <0.001 4th quartile 0.68 (0.61-0.74) <0.001

Mean Hospital Charges‡ 1st quartile Reference NA

2nd quartile 1.02 (0.93-1.12) 0.62 3rd quartile 0.96 (0.87-1.04) 0.31 4th quartile 0.86 (0.79-0.94) <0.001

* Analysis was conducted on 29,455 admissions with available data on race. The model demonstrates acceptable fit with Hosmer-Lemeshow goodness of fit, P=0.30. C-statistic=0.74. † The Charlson Comorbidity Index was tailored for stroke outcomes studies. ‡ Mean hospital charges represent the average of all hospital charges for stroke admissions within an individual hospital in a given year. Charges were adjusted using the Medical Consumer Price Index for all urban consumers prior to averaging to account for inflation. The highest and lowest 1% were excluded to avoid influence of outliers.

6

Table III. Multivariable logistic regression model for predictors of early (1-7 days) PEG tube placement in acute ischemic stroke hospitalizations (excluding those with PEG tubes placed on day 1 or day 2)

Patient Characteristics Adjusted

Odds Ratio (95% CI) P value Age (yrs)

18-54 Reference NA 55-64 1.09 (0.98-1.25) 0.11

65-74 1.11 (1.00-1.24) 0.05 75-84 1.27 (1.15-1.42) <0.001

85 1.65 (1.48-1.85) <0.001 Sex

Female (vs male) 0.98 (0.94-1.03) 0.44 Charlson Comorbidity Index†

Score: 0 Reference NA Score: 1-2 0.97 (0.91-1.02) 0.24

Score: 3 or more 0.86 (0.80-0.92) <0.001 Insurance

Medicare Reference NA Private 1.08 (0.99-1.17) 0.08

Medicaid 0.78 (0.71-0.87) <0.001 Self Pay, No Charge, or Other 0.86 (0.76-0.98) 0.02

Patient Income Low to middle income Reference NA

High income 0.86 (0.82-0.91) <0.001 Tracheostomy (vs none) 0.27 (0.25-0.30) <0.001 Atrial Fibrillation (vs none) 0.80 (0.76-0.84) <0.001 Hospital Characteristics

Teaching (vs non-teaching) 0.71 (0.68-0.75) <0.001 Urban (vs rural) 0.98 (0.90-1.07) 0.65 Stroke Volume

1st quartile Reference NA 2nd quartile 1.08 (1.02-1.15) 0.01

3rd quartile 1.21 (1.13-1.30) <0.001 4th quartile 1.25 (1.16-1.34) <0.001

Hospital Intubation Use 1st quartile Reference NA

2nd quartile 0.91 (0.83-1.00) 0.04 3rd quartile 0.81 (0.75-0.89) <0.001 4th quartile 0.67 (0.61-0.74) <0.001

Mean Hospital Charges‡ 1st quartile Reference NA

2nd quartile 0.99 (0.93-1.12) 0.79 3rd quartile 0.95 (0.87-1.04) 0.26 4th quartile 0.86 (0.79-0.94) <0.001

* Analysis was conducted on 32,008 admissions with PEG placement between 3 and 30 days from admission. The model demonstrates acceptable fit with Hosmer-Lemeshow goodness of fit, P=0.44. C-statistic=0.73. † The Charlson Comorbidity Index was tailored for stroke outcomes studies. ‡ Mean hospital charges represent the average of all hospital charges for stroke admissions within an individual hospital in a given year. Charges were adjusted using the Medical Consumer Price Index for all urban consumers prior to averaging to account for inflation. The highest and lowest 1% were excluded to avoid influence of outliers.

7

Table IV. Association of early (1-7 days) PEG tube placement with patient outcomes in those age 85

Outcomes Early PEG (1-7 days)

Later PEG (>7 days) Adjusted Odds Ratio

(95% CI)†,‡ Survivor Length of Stay, mean (SD) 9.2 (4.6) 17.3 (10.7) 0.56 (0.55-0.57) Inpatient mortality, No. (%) 384 (7.1) 243 (7.1) 1.05 (0.90-1.27) Discharge to home, No. (%) 586 (10.9) 344 (10.1) 1.05 (0.91-1.22) Discharge to acute rehab, No. (%)§ 375 (8.4) 223 (8.0) 1.12 (0.93-1.34) Discharge to SNF/LTC, No. (%)§ 2,946 (66.3) 1,858 (66.3) 0.96 (0.87-1.07)

Abbreviations: SD=Standard deviation; SNF=Skilled Nursing Facility; LTC=Long-Term Care Facility * The effect of early vs. later PEG placement was studied on various outcomes using the outcome as the dependent variable in a multivariable regression analysis adjusting for patient demographics (age, sex, payer, income), comorbidities (Charlson index, atrial fibrillation, tracheostomy placement), and hospital characteristics (teaching status, urban/rural location, stroke volume, intubation use, and mean charges for stroke patients). ‡ Logistic regression analysis was used and adjusted odds ratios were reported except for the case of length of stay, where negative binomial regression analysis was used and an incident rate ratio was reported. § For discharge to acute rehab or SNF/LTC, 17.6% of data were missing for those age 85 due to incomplete reporting of detailed disposition and therefore these missing values were excluded from the respective analyses.

Documents

Timing of Percutaneous Endoscopic Gastrostomy for Acute ...stroke.ahajournals.org/content/strokeaha/early/2016/12/13/STROKE... · gastrostomy (PEG) feeding compared ... Key Words: