Is Ambulatory Laparoscopic Roux-En-Y Gastric Bypass Associated With Higher Adverse Events?

John M. Morton, MD, MPH; Deborah Winegar, PhD; Robin Blackstone, MD; Bruce Wolfe, MD

Annals of Surgery. 2014;259(2):286-292.

Abstract and Introduction

Abstract

Objective: To determine the impact of length of stay upon 30-day outcomes.

Background: It has been recommended the goal length of stay (LOS) after laparoscopic Roux-en-Y gastric bypass (LRYGB) should be 1 day to improve resource utilization. This study's aim was to assess LRYGB outcomes by LOS.

Methods: Data were obtained from the BOLD (Bariatric Outcomes Longitudinal Database) for 51,788 laparoscopic gastric bypass (LRYGB) procedures performed between 2007 and 2010. Logistic regression models were used to evaluate age, sex, race, body mass index, insurance status, comorbidities, and LOS as predictors for 30-day mortality, serious complications, and readmissions.

Results: Overall patient demographics were as follows: median age, 45 years; median body mass index, 46.3 kg/m²; % female, 78.6; % white, 77.8; % private insurance, 86.2; and % comorbidities more than 5 (39.1%). Overall, 30-day outcomes included mortality, 0.1%; serious complications, 0.5%; and readmissions, 3.8%. median LOS was 2 days, and the distribution of LOS was as follows [n (%)]: 0 (1.0), 1 (18.4), 2 (59.0), 3 (17.5), and 4 (4.1). Using the median LOS 2 days as reference, the logistic regression analysis revealed that ambulatory LOS of was significantly associated with increased risk of 30-day mortality (odds ratio: 13.02; P < 0.0001) as was LOS 1 day (odds ratio: 2.02; P < 0.0552). For LOS of 0 day, there was a trend toward an increase in the rate of 30-day serious complications (odds ratio: 1.9; P < 0.16). There was no significant trend between LOS status and 30-day readmission rates.

Conclusions: In this large, prospective, clinical database, LOS of 1 day or less for LRYGB patients was significantly associated with an increased risk of 30-day mortality and a trend toward increased risk of 30-day serious complications.

Introduction

Bariatric surgery remains the only effective and enduring treatment of morbid obesity. The Centers for Medicare & Medicaid Services (CMS) National Coverage Determination for Bariatric Surgery provided for receipt of surgical care for the morbidly obese patient along with recognition of American Society for Metabolic and Bariatric Surgery (ASMBS) Bariatric Surgery Centers of Excellence (BSCOE) and the American College of Surgeons Bariatric Surgery Center Network.^[1] Since the recognition of bariatric COEs by the CMS, bariatric surgery has undergone dramatic advances in the safe delivery of care for the obese patient. The cornerstone of the CMS-accredited bariatric COE program is demonstration of clinical experience by surgical volume, which has been previously shown to decrease complications.^[2,3] Another key tenet to the CMS National Coverage Decision was the emphasis on the availability of clinical services, which is meant to provide comprehensive care and specialized consultation if needed for a population known to have increased risk of postoperative complications. In addition to the COE movement, other advances in bariatric surgical care have occurred most notably the clear advent of laparoscopy. Level 1 data exist for the laparoscopic approach superiority for decreased morbidity.^[4] All of these factors have converged upon bariatric surgery to decidedly improve patient outcomes, particularly for laparoscopic Roux-en-Y gastric bypass (LRYGB).

A considerable amount of contemporary evidence displays the current safety profile of LRYGB with a clear evolution from prior performance.^[5] In a University Health Consortium study, Nguyen et al^[6] has shown a 33% reduction rate in gastric bypass mortality over time with an in-patient mortality rate of 0.2%. The Longitudinal Assessment of Bariatric Surgery study also showed an exceptionally low 30-day mortality rate of 0.2% for LRYGB.^[7] The Michigan Bariatric Surgery Collaborative study recapitulates these results with a 30-day mortality rate of 0.14%.^[8] Encinosa et al^[9] further demonstrated substantial improvement in bariatric surgery for all payers, with 21% and 31% reduction rates in complications and readmissions, respectively.

Although patient safety has remained a paramount concern since the Institute of Medicine publication To Err Is Human,^[10] another concern from policy makers are burgeoning health care costs. Many cost-containment strategies been proposed including nonreimbursement for certain complications deemed preventable such as hospital-acquired conditions and 30-day readmissions.^[11,12] Another potential cost-saving measure involves shifting episodes of care from an in-patient setting to an ambulatory model.

The combination of increased bariatric surgical patient safety and emphasis on cost-consciousness has emboldened some to advocate for ambulatory bariatric surgery. Although data exist for safe ambulatory laparoscopic gastric banding,^[13,14] little is known regarding the safe provision of ambulatory laparoscopic Roux-en-Y gastric bypass. Currently, there are few publications advocating for ambulatory LRYGB.^[15,16] One study demonstrated the potential for ambulatory LRYGB with 0.1% mortality, 4.3% serious complication, and 1.7% readmission rate. Despite the lack of abundant literature to support short-stay LRYGB, a health care consulting firm has issued a guideline indicating that a goal length of stay (LOS) for LRYGB should be 1 day or less.^[17] Given the lack of population-based data, the purpose of this study was to analyze outcomes on the basis of LOS.

1 2 3 4 5 6

Next Section

Table 1. Summary Patient Demographics and Outcomes
Laparoscopic Roux-en-Y Gastric Bypass (N = 51,788)
Age at surgery, yr
Mean (SD)	45.2 (11.45)
Median (range)	45.0 (18–75)
Age at surgery: ≥50 yr	19,287 (37.2%)
Sex: Female	40,711 (78.6%)
Race: White	40,274 (77.8%)
BMI, kg/m²
Mean (SD)	47.8 (7.92)
Median (range)	46.3 (35–143)
BMI, kg/m²
<50	4,994 (67.6%)
50–70	15,949 (30.8%)
≥70	845 (1.6%)
Insurance status: Private	44,627 (86.2%)
No. comorbidity: ≥5	20,245 (39.1%)
Length of hospital stay, d
Mean (SD)	2.1 (0.75)
Median (range)	2.0 (0–4)
Length of hospital stay, d
0	507 (1.0%)
1	9,513 (18.4%)
0/1	10,020 (19.3%)
2	30,592 (59.1%)
3	9,047 (17.5%)
4	2,129 (4.1%)
30-d mortality	53 (0.1%)
30-d serious complications	397 (0.8%)
30-d readmission	1,951 (3.8%)

Table 2. Patient Demographics by LOS
Baseline Characteristics	LOS = 0 (n = 507)	LOS = 1 (n = 9,513)	LOS = 0/1 (n = 10,020)	LOS = 2 (n = 30,592)	LOS = 3 (n = 9,047)	LOS = 4 (n = 2,129)	P*
Age at surgery, mean (SD), yr	44.2 (11.28)	44.4 (11.50)	44.3 (11.48)	45.0 (11.34)	46.1 (11.61)	48.0 (11.65)	<0.0001
BMI, mean (SD), kg/m²	47.2 (7.18)	47.4 (7.69)	47.4 (7.66)	47.7 (7.74)	48.4 (8.39)	49.2 (9.23)	<0.0001
Sex: Female	398 (78.5%)	7,447 (78.3%)	7,845 (78.3%)	24,146 (78.9%)	7,124 (78.7%)	1,596 (75.0%)	0.0007
Race: White	390 (76.9%)	7,113 (74.8%)	7,503 (74.9%)	24,202 (79.1%)	6,986 (77.2%)	1,583 (74.4%)	<0.0001
Insurance: Private	447 (88.2%)	8,425 (88.6%)	8,872 (88.5%)	26,569 (86.8%)	7,512 (83.0%)	1,674 (78.6%)	<0.0001
No. comorbidities: ≥5	178 (35.1%)	3,430 (36.1%)	3,608 (36.0%)	11,853 (38.7%)	3,746 (41.4%)	1,038 (48.8%)	<0.0001

*P values for the comparison between LOS groups were calculated using the χ ² test for categorical variables and the Kruskal-Wallis test for continuous variables.

Table 3. Thirty-Day Outcomes by LOS
Baseline Characteristics	LOS = 0 (n = 507)	LOS = 1 (n = 9,513)	LOS = 2 (n = 30,592)	LOS = 3 (n = 9,047)	LOS = 4 (n = 2,129)	P*
Mortality	4 (0.8%)	12 (0.1%)	20 (0.1%)	7 (0.1%)	10 (0.5%)	<0.0001
Serious complications	5 (1.0%)	45 (0.5%)	166 (0.5%)	110 (1.2%)	71 (3.3%)	<0.0001
Readmissions	12 (2.4%)	291 (3.1%)	1,032 (3.4%)	479 (5.3%)	137 (6.4%)	<0.0001

*P values for the comparison between LOS groups were calculated using the χ ² test.

Table 4. Thirty-Day Mortality: Logistic Regression Analysis
	Unadjusted		Adjusted, MLR*		Adjusted, MLR†
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Age at surgery, yr
<50	1.00		Reference		1.00
≥50	3.28 (1.86–5.79)	<0.0001	2.71 (1.48–4.96)	0.0013	2.63 (1.45–4.79)	0.0015
Sex
Female	1.00		Reference		1.00
Male	2.42 (1.39–4.19)	0.0017	2.12 (1.22–3.69)	0.0079	2.06 (1.18–3.58)	0.0107
Race
White	1.00		Reference		1.00
Nonwhite	1.26 (0.68–2.31)	0.4645	1.63 (0.87–3.06)	0.1278
BMI, kg/m²
<50		1.00		Reference		1.00
50–70	1.80 (1.04–3.13)	0.0362	1.76 (1.01–3.07)	0.0478	1.80 (1.03–3.14)	0.0393
≥70	2.96 (0.70–12.46)	0.1383	2.89 (0.67–12.48)	0.1561	3.15 (0.73–13.54)	0.1232
Insurance status
Private	1.00		Reference
Nonprivate	2.03 (1.08–3.79)	0.0269	1.33 (0.70–2.54)	0.3802
No. comorbidity
<5	1.00		Reference		1.00
≥5	3.03 (1.72–5.36)	0.0001	2.22 (1.21–4.06)	0.0099	2.17 (1.20–3.93)	0.0102
LOS, d
2	1.00		Reference		1.00
0	12.16 (4.14–35.69)	<0.0001	12.97 (4.40–38.25)	<0.0001	13.02 (4.42–38.39)	<0.0001
1	1.93 (0.94–3.95)	0.0718	1.97 (0.96–4.05)	0.0632	2.02 (0.98–4.13)	0.0552
0/1	2.44 (1.27–4.72)	0.0077	2.51 (1.30–4.85)	0.0063	2.56 (1.32–4.94)	0.0052
3	1.18 (0.50–2.80)	0.7011	1.04 (0.44–2.46)	0.9329	1.07 (0.45–2.54)	0.8765
4	7.21 (3.37–15.43)	<0.0001	5.33 (2.47–11.49)	<0.0001	5.61 (2.61–12.05)	<0.0001

*Adjusted for all covariates.
†Adjusted for significant predictors (P < 0.10).
CI indicates confidence interval; MLR multivariate logistic regression.

Table 5. Thirty-Day Serious Complications: Logistic Regression Analysis
	Unadjusted		Adjusted, MLR*		Adjusted, MLR†
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Age at surgery, yr
<50	1.00		1.00		1.00
≥50	1.84 (1.51–2.25)	<0.0001	1.41 (1.14–1.74)	0.0014	1.41 (1.15–1.74)	0.0012
Sex
Female		1.00		1.00		1.00
Male	1.95 (1.58–2.40)	<0.0001	1.77 (1.44–2.19)	<0.0001	1.78 (1.44–2.19)	<0.0001
Race
White	1.00		1.00
Nonwhite	0.84 (0.65–1.08)	0.1729	0.96 (0.75–1.25)	0.7821
BMI, kg/m²
<50	1.00		1.00		1.00
50–70	1.55 (1.26–1.90)	<0.0001	1.42 (1.15–1.74)	0.0009	1.42 (1.15–1.74)	0.0010
≥70	2.22 (1.24–3.98)	0.0076	1.71 (0.94–3.09)	0.0789	1.70 (0.94–3.08)	0.0809
Insurance status
Private	1.00		1.00		1.00
Nonprivate	1.81 (1.43–2.30)	<0.0001	1.28 (1.00–1.64)	0.0475	1.28 (1.00–1.64)	0.0485
No. comorbidity
<5	1.00		1.00		1.00
≥5	2.30 (1.88–2.81)	<0.0001	1.82 (1.47–2.26)	<0.0001	1.83 (1.48–2.27)	<0.0001
LOS, d
2	1.00		1.00		1.00
0	1.83 (0.75–4.46)	0.1870	1.90 (0.78–4.65)	0.1599	1.90 (0.78–4.65)	0.1603
1	0.87 (0.63–1.21)	0.4129	0.90 (0.65–1.25)	0.5246	0.90 (0.64–1.25)	0.5180
0/1	0.92 (0.67–1.26)	0.6024	0.95 (0.69–1.30)	0.7429	0.95 (0.69–1.30)	0.7351
3	2.26 (1.77–2.87)	<0.0001	2.12 (1.67–2.71)	<0.0001	2.12 (1.66–2.71)	<0.0001
4	6.33 (4.77–8.38)	<0.0001	5.35 (4.03–7.12)	<0.0001	5.35 (4.02–7.11)	<0.0001

*Adjusted for all covariates.
† Adjusted for significant predictors (P < 0.10).
CI indicates confidence interval; MLR, multivariate logistic regression.

Table 6. Thirty-Day Readmissions: Logistic Regression Analysis
	Unadjusted		Adjusted, MLR*		Adjusted, MLR†
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
Age at surgery, yr
<50	1.00		1.00		1.00
≥50	0.97 (0.88–1.06)	0.4877	0.86 (0.77–0.94)	0.0021	0.85 (0.77–0.94)	0.0013
Sex
Female	1.00		1.00
Male	0.95 (0.85–1.06)	0.3892	0.95 (0.85–1.06)	0.3854
Race: Nonwhite	1.11 (1.00–1.23)	0.0575	1.13 (1.01–1.25)	0.0327	1.13 (1.01–1.26)	0.0259
BMI, kg/m²
<50	1.00		1.00
50–70	1.12 (1.02–1.23)	0.0225	1.06 (0.96–1.17)	0.2532
≥70	1.18 (0.84–1.66)	0.3338	0.99 (0.70–1.39)	0.9479
Insurance status
Private	1.00		1.00		1.00
Nonprivate	1.50 (1.34–1.68)	<0.0001	1.35 (1.20–1.52)	<0.0001	1.36 (1.20–1.53)	<0.0001
No. comorbidity
<5	1.00		1.00		1.00
≥5	1.37 (1.25–1.50)	<0.0001	1.35 (1.23–1.49)	<0.0001	1.36 (1.23–1.50)	<0.0001
LOS, d
2	1.00		1.00		1.00
0	0.69 (0.39–1.24)	0.2146	0.70 (0.40–1.25)	0.2314	0.70 (0.39–1.25)	0.2294
1	0.90 (0.79–1.03)	0.1338	0.91 (0.80–1.04)	0.1657	0.91 (0.80–1.04)	0.1595
0/1	0.89 (0.78–1.02)	0.0888	0.90 (0.79–1.03)	0.1134	0.90 (0.79–1.02)	0.1087
3	1.60 (1.43–1.79)	<0.0001	1.57 (1.41–1.76)	<0.0001	1.58 (1.41–1.76)	<0.0001
4	1.97 (1.64–2.37)	<0.0001	1.88 (1.56–2.26)	<0.0001	1.88 (1.56–2.26)	<0.0001

*Adjusted for all covariates.
†Adjusted for significant predictors (P < 0.10).
CI indicates confidence interval; MLR, multivariate logistic regression.

Table 7. Causes of 30-Day Mortality by LOS
Cause of Death	Overall (n = 51,788)	LOS = 0 (n = 507)	LOS = 1 (n = 9,513)	LOS = 0/1 (n = 10,020)	LOS = 2 (n = 30,592)	LOS = 3 (n = 9,047)	LOS = 4 (n = 2,129)
Sepsis	13	0	1	1	6	3	3
Bowel obstruction	1	0	0	0	0	0	1
Cerebrovascular accident (stroke)	3	0	1	1	0	1	1
Myocardial infarction	6	3	1	4	1	0	1
Other cause	5	0	2	2	2	0	1
Respiratory failure, including ARDS	2	0	1	1	1	0	0
Cardiac failure	6	1	0	1	4	1	0
Pulmonary embolus	6	0	3	3	2	1	0
Undetermined	11	0	3	3	4	1	3