ORIGINAL ARTICLE
Early Complications After Laparoscopic Gastric Bypass Surgery: Results From the Scandinavian Obesity Surgery Registry ˚ Erik Stenberg, MD,∗ † Eva Szabo, MD, PhD,† G¨oran Agren, MD,† Erik N¨aslund, MD, PhD,‡ Lars Boman, MD, PhD,§ Ami Bylund, RN, MSc,¶ Jan Hedenbro, MD, PhD,||∗∗ Anna Laurenius, RD, PhD,†† G¨oran Lundeg˚ardh, MD, PhD,‡‡ Hans L¨onroth, MD, PhD,†† Peter M¨oller, MD,§§ Magnus Sundbom, MD, PhD,¶¶ Johan Ottosson, MD, PhD,∗ † and Ingmar N¨aslund, MD, PhD†; For the Scandinavian Obesity Surgery Registry Study Group
Objective: To identify risk factors for serious and specific early complications of laparoscopic gastric bypass surgery using a large national cohort of patients. Background: Bariatric procedures are among the most common surgical procedures today. There is, however, still a need to identify preoperative and intraoperative risk factors for serious complications. Methods: From the Scandinavian Obesity Surgery Registry database, we identified 26,173 patients undergoing primary laparoscopic gastric bypass operation for morbid obesity between May 1, 2007, and September 30, 2012. Follow-up on day 30 was 95.7%. Preoperative data and data from the operation were analyzed against serious postoperative complications and specific complications. Results: The overall risk of serious postoperative complications was 3.4%. Age (adjusted P = 0.028), other additional operation [odds ratio (OR) = 1.50; confidence interval (CI): 1.04–2.18], intraoperative adverse event (OR = 2.63; 1.89–3.66), and conversion to open surgery (OR = 4.12; CI: 2.47– 6.89) were all risk factors for serious postoperative complications. Annual hospital volume affected the rate of serious postoperative complications. If the hospital was in a learning curve at the time of the operation, the risk for serious postoperative complications was higher (OR = 1.45; CI: 1.22–1.71). The 90-day mortality rate was 0.04%. Conclusions: Intraoperative adverse events and conversion to open surgery are the strongest risk factors for serious complications after laparoscopic gastric bypass surgery. Annual operative volume and total institutional experience
are important for the outcome. Patient related factors, in particular age, also increased the risk but to a lesser extent. Keywords: bariatric surgery, laparoscopic gastric bypass, postoperative complications, risk factors (Ann Surg 2014;260:1040–1047)
O
From the ∗ Department of Surgery, Lindesberg Hospital, Lindesberg, Sweden; ¨ ¨ †Department of Surgery, Orebro University Hospital, Orebro, Sweden; ‡Division of Surgery, Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden; §Department of Surgery, Lycksele Hospital, Lycksele, Sweden; ¶Department of Surgery, Ersta Hospital, Stockholm, Sweden; ||Sk˚ane University Hospital, Lund University, Lund, Sweden; ∗∗ Department of Surgery, Aleris Obesity Sk˚ane, Lund, Sweden; ††Department of Surgery, Sahlgrenska University Hospital, Gastrosurgical Research, Sahlgrenska Academy, University of Gothenburg, Gothenburg, ¨ Simrishamn Hospital, Simrishamn, Sweden; Sweden; ‡‡Osterlenskirurgin, §§Department of Surgery, Kalmar County Hospital, Kalmar, Sweden; and ¶¶Department of Surgical Sciences, Uppsala University, Uppsala, Sweden. ¨ Disclosure: This work was supported by grants from the Orebro County Council. None of the authors have received any grants or funding for this study that would be considered as a conflict of interest. However, Jan Hedenbro has received travel grants from Covidien and Johnson & Johnson and received grants from Lund University funds, Crafoord foundation; Erik N¨aslund received grants from Novo Nordisk, Stockholm County Council, Diabetes Theme center at Karolinska Institutet, Stockholm, Sweden; Magnus Sundbom received expert testimony from the Swedish National Board of Health and Welfare and have received royalties from Studentlitteratur for books for medical students. Ingmar N¨aslund received grants from the Swedish government used to run the Scandinavian Obesity Surgery Registry. For the remaining authors, no relevant potential conflicts of interest are declared. ¨ Reprints: Erik Stenberg, MD, Department of Surgery, Orebro University Hospital, ¨ 70185 Orebro, Sweden. E-mail: [email protected]. C 2013 by Lippincott Williams & Wilkins Copyright ISSN: 0003-4932/13/26006-1040 DOI: 10.1097/SLA.0000000000000431
besity is associated with an increased risk for heart failure,1 diabetes,2 and cancer3 ; reduced life expectancy4 ; and quality of life.5 Surgical treatment is currently the best way to establish longterm weight loss,6 and the benefits of surgery for morbid obesity have been reported in numerous studies.6–10 The number of bariatric procedures performed annually has increased dramatically over the last decade, but now it seems to have reached a plateau.11–13 Gastric bypass is the most frequently used method worldwide,12 and in Sweden 96.8% of all bariatric procedures in 2012 were gastric bypasses.13 Complication rates are low,14,15 but given the large number of procedures performed, even low complication rates will cause serious morbidity for a large number of patients and have a great impact upon health care economy. For this reason, it is important to identify risk factors for postoperative complications. In a recent study from the Michigan Bariatric Surgery Collaborative, previous history of venous thrombosis, mobility limitation, coronary artery disease, age more than 50 years, pulmonary disease, male sex, and smoking history were associated with increased risk for postoperative complications,16 as was the annual number of bariatric procedures performed at each hospital.14 Diabetes has also been described as a risk factor.17 While establishing a new procedure, complication rates are high.18,19 The learning curve of the individual surgeon has been described as 50 to 100 cases.18–21 Adverse intraoperative events have also been associated with a higher risk for postoperative complications.22 Whether or not this increase in risk varies depending on the nature of the specific events, however, has not been studied. The Obesity Surgery Mortality Risk Score (OS-MRS) is an accepted score for risk stratification of mortality after bariatric surgery.23 As the mortality rate is low and declining, it would probably be more suitable to have a risk classification score predicting serious complications. The OS-MRS has not been tested for serious complications. Recently, other scores focusing on serious postoperative complications have been proposed.15–17,24 The Scandinavian Obesity Surgery Registry (SOReg) was created to monitor the expanding practice of obesity surgery, its complications, and the effects of treatment. SOReg now covers 98% of all bariatric procedures performed in Sweden,13 providing a unique possibility for analyses. The objective of this study was to review early complications in a prospective material from the SOReg database and to identify preoperative and intraoperative factors predisposing to postoperative complications.
1040 | www.annalsofsurgery.com
Annals of Surgery r Volume 260, Number 6, December 2014
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
METHODS Study Cohort SOReg is a national quality and research register, which started in May 2007. Since 2010, it has covered virtually all bariatric procedures in Sweden and all bariatric centers report to the register. At present, approximately 8000 bariatric procedures (85/100,000 inhabitants) are performed annually at 44 centers. Data are prospectively collected at baseline about 1 month before surgery; during surgery; and after 6 weeks (for days 0–30) and 1, 2, and 5 years of the surgery. The database is compared annually with the Swedish Population Register so as to cover all mortality. The database is validated in several ways. Different variables are systematically searched for nonlogic or improbable values and in such cases compared to the patient’s record. Departments are reminded if follow-up data are missing and the most important variables are mandatory. Regular audit is performed by randomly comparing selected variables with patient records. So far, the database has been shown to include less than 3% incorrect values.13 From the SOReg database, we identified a total of 29,288 patients who underwent a bariatric procedure between May 1, 2007, and September 30, 2012. We excluded patients undergoing bariatric procedures other than laparoscopic gastric bypass. We also excluded revision surgery and planned open gastric bypass surgery, but not
Early Complications After LRYGBP
conversion from laparoscopic to open surgery. After exclusion, a total of 26,173 patients were available for the study. Of these 25,038 had follow-up data from day 30 (95.7%) (Fig. 1). The surgical technique is well-established in Sweden. Of the laparoscopic gastric bypass operations in 2012, 99% had an antecolic, antegastric laparoscopic Roux-en-Y gastric bypass as described by L¨onroth and Olbers.25 The gastrojejunostomy was constructed using 1 linear stapler combined with handsewn closure of the remaining defect in 98.5% of all cases.13 All patients received prophylaxis against deep vein thrombosis.
Comorbidity Comorbidity is defined as a condition needing active pharmacological or continuous positive airway pressure treatment, registered as hypertension, diabetes, dyslipidemia, diarrhea, depression, sleep apnea, or other condition. History of venous thromboembolism and history of smoking were registered from May 1, 2010. OS-MRS was also calculated for patients after this date.
Hospital Volume and Learning Curve Annual surgical volume for each respective hospital at the time of a specific operation was calculated by number of operations in the
FIGURE 1. CONSORT diagram. C 2013 Lippincott Williams & Wilkins
www.annalsofsurgery.com | 1041
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Stenberg et al
6-month period when the procedure took place plus the preceding 12-months period, divided by 1.5. Surgical volumes for the period before the register existed were based on data from the Swedish Board of Health and Welfare. This procedure enabled us to take into account the increasing experience of the departments included. Each case was also categorized depending on the preceding experience at the institution at the time of the operation.
Outcomes Outcomes were measured in 2 ways. First, specific predefined complications and any other complication arising were registered for all patients. The specific complications were leakage or deep intra-abdominal infection, bleeding requiring intervention, gastrointestinal obstruction/ileus, port-related complications, wound dehiscence, other wound complication (mainly superficial wound infections), anastomotic stricture, stomal ulcer, cardiovascular event, pulmonary complication (other than pulmonary embolism), venous thromboembolism, urinary tract infection, and other (in this case specified) complication. Second, the severity of the complication was assessed using the Clavien-Dindo classification.26 This was included in the register 2010 and data are available for all patients operated after January 1, 2010 (n = 18,737). For this group, the main outcome was a serious complication defined as Clavien ≥ 3b (A complication requiring an invasive procedure under general anesthesia, failure of one or several organ systems requiring treatment on an intensive care facility or death).
Statistical Analysis Operative time and hospital stay were analyzed with unadjusted linear regression. Data on overall postoperative complications, serious complications, and specific complications were analyzed with unadjusted logistic regression. P < 0.05 was considered statistically significant. Patient-specific risk factors with a P < 0.10 were entered into a multivariate analysis using stepwise, forward, logistic regression. Significant adjusted risk factors were then entered into a multiple regression analysis with significant surgical risk factors. P < 0.05 was considered statistically significant. Data were analyzed
using IBM SPSS Statistics version 20 (IBM Corporation, Armonk, NY). For odds ratios (ORs), the 95% confidence interval (CI) was also reported.
Ethical Considerations The study was conducted in accordance with the ethical standards of the Helsinki declaration sixth revision and was approved by ¨ the Uppsala/Orebro Regional Ethic Committee. Trial registration is available at clinicaltrials.gov identifier: NCT01862159.
RESULTS Of the 25,038 patients in the study, 76.0% were women. Mean body mass index (BMI) was 42.7 ± 5.43 kg/m2 . Comorbidity, as defined earlier, was present in 48.0% of all patients. Thirty-one percent had a history of smoking (15.8% active smokers, 15.2% previous history of smoking). Baseline, demographic data are presented in more detail in Table 1. In all cases, included surgery began with intention to complete laparoscopically. The conversion rate was 1.1%. The main reasons for converting to open surgery were difficult anatomic conditions (32.1% of conversions) and adhesions (30.9%). Other reasons were intraoperative complications (14.7%), instrument failure (6.4%), problems with ports (4.5%), and other reasons (11.3%). An additional procedure, such as cholecystectomy, hernia repair, and gynecological procedure, was performed at the same time as the gastric bypass in 3.5% of all patients. An intraoperative adverse event detected and managed during the primary procedure occurred in 2.8%. The most common adverse event was unintentional bowel perforation (52.5% of all events), followed by bleeding (20.8%), submucosal stapling or malrotation (10.0%), instrument failure (5.6%), stapling of the nasogastric tube (5.2%), and other reasons (5.9%). A postoperative complication occurred in 8.7% of all patients between day 0 and day 30 postoperatively or, if the primary hospital stay exceeded 30 days, during the course of the stay. Grading of complications was available for all patients operated after January 1, 2010. For patients operated after this date, a postoperative complication was seen in 8.3% of all cases. Specific complications are
TABLE 1. Baseline Characteristics Operated After May 1, 2007 (All) No. individuals, n Sex, n (%) Male Female Age at operation (mean ± SD), yrs Coexisting conditions with ongoing treatment, n (%) Sleepapnea, n (%) Hypertension, n (%) Diabetes, n (%) Dyslipidemia, n (%) Depression, n (%) Diarrhea, n (%) Other condition,∗ † n (%) Previous venous thromboembolism,† n (%) Smoking,† n (%) Active smoking, n (%) Previous history of smoking, n (%) BMI (mean ± SD), Kg/m2 Waist circumference (mean ± SD), cm
Missing Data, n
25,038 6021 (24.0) 19,017 (76.0) 41.1 ± 10.97 12,026 (48.0) 2445 (9.8) 6321 (25.2) 3760 (15.0) 2540 (10.1) 3453 (13.8) 310 (1.2) 2881 (12.7) 422 (1.7) 4691 (31.0) 2386 (15.8) 2305 (15.2) 42.7 ± 5.43 127.2 ± 13.82
Operated After Jan 1, 2010
Missing Data, n
18,737 0 0 0 0 0 0 0 0 0 0 2302 9449 9906 0 4670
4529 (24.2) 14,208 (75.8) 41.2 ± 11.03 9684 (51.7) 1987 (10.6) 5018 (26.8) 2933 (15.7) 2055 (11.0) 2674 (14.3) 289 (1.5) 2801 (15.2) 407 (2.7) 4588 (30.9) 2316 (15.6) 2272 (15.3) 42.5 ± 5.32 127.1 ± 13.65
0 0 0 0 0 0 0 0 0 0 310 3593 3907 0 3170
∗ Other coexisting conditions are (in percentage of the total material) cardiovascular disease (0.6%), pain or mobility limitations (8.6%), psychiatric disorder other than affective disorder (0.1%), systemic disease (0.5%), pulmonary disease (0.7%), and other (0.8%). †Not mandatory parameter until May 1, 2010.
1042 | www.annalsofsurgery.com
C 2013 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
listed in Table 2. Eleven patients died during the follow-up, 7 due to septicemia, 3 due to cardiovascular events, and 1 from pulmonary embolism. There were no further deaths during the first 90 days after surgery. The 90-day mortality was thus 0.04% (11/25,038). Sleep apnea (OR = 5.28; CI: 1.55–18.07; P = 0.008), dyslipidemia (OR = 7.39; CI: 2.25–24.24; P = 0.001), and age more than 60 years (OR = 7.43; CI: 1.97–28.03; P = 0.003) were associated with an increased risk for death within 90 days after the operation.
Patient-specific Risk Factors
Age (P = 0.028), sleep apnea (P = 0.046), and depression (P = 0.048) increased the risk for serious complications. Adjusted analysis on patient-specific variables resulted in significant results for age (adjusted OR 30–40 years = 1.08; CI: 0.82–1.41, P = 0.585; 40–50 years = 1.34; CI: 1.04–1.72, P = 0.025; 50–60 years = 1.40; CI: 1.06–1.84, P = 0.016; and >60 years = 1.48; CI: 0.99–2.20, P = 0.054). BMI did not affect the risk for serious or specific complications. The greater the waist circumference was, the higher the risk for leakage/deep infection (P = 0.005) and other wound complications (P = 0.032). Age specifically increased the risk for leakage/deep
Early Complications After LRYGBP
infection (P < 0.001), bleeding (P < 0.001), strictures (P = 0.003), and other wound complication (P < 0.001). Previous venous thromboembolism, sleep apnea, hypertension, diabetes, dyslipidemia, depression, and diarrhea all increased the risk for postoperative complication. Sleep apnea specifically increased the risk for leakage/deep infection (OR = 1.67; CI: 1.28–2.17; P < 0.001), bleeding (OR = 1.49; CI: 1.16–1.92; P = 0.002), venous thromboembolism (OR = 2.64; CI: 1.07–6.56; P = 0.036), cardiovascular complications (OR = 2.53; CI: 1.33–4.79; P = 0.004), stomal ulcer (OR = 1.82; CI: 1.11–2.98; P = 0.018), and other wound complications (OR = 1.45; CI: 1.02–2.07; P = 0.038). Hypertension increased the risk for leakage/deep infection (OR = 1.32; CI: 1.08–1.62; P = 0.007), bleeding (OR = 1.78; CI: 1.48–2.12; P < 0.001), cardiovascular complications (OR = 4.27; CI: 2.50–7.27; P < 0.001), and pulmonary complications (OR = 1.60; CI: 1.18– 2.17; P = 0.003) but was associated with a lower risk for gastrointestinal obstruction/ileus (OR = 0.72; CI: 0.52–0.98; P = 0.040). Diabetes increased the risk for leakage/deep infection (OR = 1.64; CI: 1.31–2.05; P < 0.001), bleeding (OR = 1.71; CI: 1.39–2.11; P < 0.001), cardiovascular complications (OR = 3.41; CI: 1.98–5.86; P < 0.001), pulmonary complications (OR = 1.80; CI: 1.28–2.53;
TABLE 2. Perioperative Data
Access Laparoscopic, n (%) Converted to open surgery, n (%) Additional procedure, n (%)∗ Length of operation, (mean ± SD), min Length of hospital stay (mean ± SD), d Intraoperative adverse event, n (%) Bleeding, n (%) Unintentional bowel injury, n (%) Instrument failure, n (%) Submucosal stapling or malrotation, n (%)† Stapling of nasogastric tube, n (%) Other complication, n (%) Complication within 30 d, n (%) Leakage or abscesses, n (%) Bleeding, n (%) Wound dehiscence, n (%) Port-related complication, n (%) Other wound complication, n (%) Small bowel obstruction, n (%) Stricture, n (%) Stomal ulcer, n (%) Venous thromboembolism, n (%) Cardiovascular complication, n (%) Pulmonary complication, n (%) Urinary tract infection, n (%) Severity of complication,‡ n (%) No complication I II IIIa IIIb IVa IVb V
Operated After May 1, 2007 (All)
Missing Data, n
Operated After Jan 1, 2010
Missing Data, n
24,773 (98.9) 265 (1.1) 881 (3.5) 77.2 ± 38.10 2.3 ± 3.17 712 (2.8) 148 (0.6) 374 (1.5) 40 (0.2) 71 (0.3) 37 (0.1) 42 (0.2) 2180 (8.7) 449 (1.8) 522 (2.1) 15 (0.1) 140 (0.6) 266 (1.1) 241 (1.0) 59 (0.2) 116 (0.5) 27 (0.1) 56 (0.2) 183 (0.7) 107 (0.4)
0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
18,599 (99.3) 138 (0.7) 633 (3.4) 72.4 ± 34.99 2.0 ± 2.90 508 (2.7) 105 (0.6) 268 (1.4) 32 (0.2) 47 (0.3) 24 (0.1) 32 (0.2) 1546 (8.3) 298 (1.6) 371 (2.0) 9 (0.0) 92 (0.5) 169 (0.9) 197 (1.1) 48 (0.3) 91 (0.5) 18 (0.1) 40 (0.2) 119 (0.6) 77 (0.4)
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
17,191 (91.7) 197 (1.1) 579 (3.1) 140 (0.7) 598 (3.2) 16 (0.1) 8 (0.0) 8 (0.0)
0 0 0 0 0 0 0 0
∗ Other procedure performed at the same time as the bariatric procedure. †Submucosal stapling, malrotation of bowel. ‡Classification according to Clavien-Dindo. Grade I is a complication without need for pharmacological, surgical, endoscopic, or radiological intervention; grade II requires pharmacological treatment; grade IIIa intervention not under general anesthesia; grade IIIb intervention under general anesthesia; grade IVa singe organ dysfunction; grade IVb multiorgan dysfunction; grade V death.
C 2013 Lippincott Williams & Wilkins
www.annalsofsurgery.com | 1043
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Stenberg et al
P = 0.001), stomal ulcer (OR = 1.81; CI: 1.18–2.77; P = 0.007), and other wound complication (OR = 1.59; CI: 1.18–2.13; P = 0.002) but was associated with a lower risk for gastrointestinal obstruction/ileus (OR = 0.65; CI: 0.43–0.99; P = 0.044). Dyslipidemia increased the risk for bleeding (OR = 1.82; CI: 1.45–2.30; P < 0.001), venous thromboembolism (OR = 2.53; CI: 1.02–6.28; P = 0.045), cardiovascular complications (OR = 2.17; CI: 1.12–4.20; P = 0.021), and pulmonary complications (OR = 1.54; CI: 1.02–2.32; P = 0.040). Diarrhea increased the risk for bleeding (OR = 1.91; CI: 1.07–3.43; P = 0.030), venous thromboembolism (OR = 6.42; CI: 1.51–27.21; P = 0.012), and anastomotic stricture (OR = 5.86; CI: 2.11–16.28; P = 0.006). Depression increased the risk for leakage/deep infection (OR = 1.30; CI: 1.01–1.67; P = 0.038), bleeding (OR = 1.38; CI: 1.10–1.73; P = 0.005), pulmonary complications (OR = 1.81; CI: 1.28–2.57; P = 0.001), and stomal ulcer (OR = 1.72; CI: 1.10–2.69; P = 0.016). Previous history of venous thromboembolism increased the risk for postoperative venous thromboembolism (OR = 11.24; CI: 3.65–34.63; P < 0.001) and port-related complications (OR = 3.33; CI: 1.43–7.72; P = 0.005). History of smoking increased the risk for pulmonary complications (OR = 1.53; CI: 1.01–2.30; P = 0.042), other wound complications (OR = 1.47; CI: 1.02–2.12; P = 0.040), and urinary tract infections (OR = 1.92; CI: 1.19–3.11; P = 0.007). Male sex was associated with a higher total incidence of postoperative complications but did not increase the risk for serious complications. Male sex increased the risk for leakage/deep infection (OR = 1.60; CI: 1.31–1.95; P < 0.001), bleeding (OR = 1.60; CI: 1.33–1.92; P < 0.001), and port-related complications (OR = 1.65; CI: 1.09–2.21; P = 0.015) but had a lower risk for gastrointestinal obstruction/ileus (OR = 0.54; CI: 0.37–0.76; P = 0.001). OS-MRS did not predict serious complications. It did, however, predict some specific complications. Higher scores on the OS-MRS were associated with an increased risk for leakage/deep infection (P = 0.004), bleeding (P < 0.001), cardiovascular complications (P = 0.012), port-related complications (P = 0.001), and other wound complications (P = 0.002)
Surgery-specific Risk Factors Conversion to open surgery was associated with an increased risk for serious postoperative complications (P < 0.001), and specifically increased the risk for leakage/deep infection (OR = 5.98; CI: 3.92–9.13; P < 0.001), bleeding (OR = 2.06; CI: 1.12–3.78; P = 0.021), venous thromboembolism (OR = 11.81; CI: 3.53–39.45; P < 0.001), and other wound complications, mainly wound infections (OR = 15.57; CI: 10.62–22.82; P < 0.001). There was no statistically significant difference between the reasons for conversion to open surgery with regards to serious postoperative complications. Conversion for problems with ports (OR = 11.49; CI: 2.51–52.58; P = 0.002), difficult anatomic conditions (OR = 7.66; CI: 3.93–14.92; P < 0.001), intraoperative complications (OR = 6.56; CI: 2.32–18.55; P < 0.001), and adherences (OR = 4.53; CI: 1.96–10.47; P < 0.001) specifically increased the risk for leakage/deep infection. Conversion due to intraoperative complications also increased the risk for postoperative bleeding (OR = 5.43; CI: 1.92–15.32; P = 0.001). The risk for other wound complications was raised for all reasons for conversion. Suffering from an intraoperative adverse event increased the risk for a serious postoperative complication, specifically leakage/deep infection (OR = 3.79; CI: 2.74–5.23; P < 0.001), bleeding (OR = 2.05; CI: 1.40–3.01; P < 0.001), pulmonary complications (OR = 3.56; CI: 2.15–5.90; P < 0.001), stomal ulcer (OR = 2.55; CI: 1.24–5.25; P = 0.011), port-related complications (OR = 2.08; CI: 1.02–4.27; P = 0.045), and other wound complications (OR = 2.83; CI: 1.78–4.49; P < 0.001). Submucosal stapling or unintentional malrotation (10.6% suffered from a serious complication) resulted 1044 | www.annalsofsurgery.com
in the highest risk of a serious complication followed by unintentional bowel injury (9.0%), bleeding (7.6%), and instrument failure (6.2%). For stapling of the nasogastric tube (4.2%) and other reasons (3.1%), the risk was not significantly increased. Performance of an additional procedure at the time of the operation increased the risk for other wound complications (OR = 2.13; CI: 1.33–3.42; P = 0.002) (Table 3). Multiple regression analyses of age- and surgery-specific risk factors resulted in significant results for intraoperative complications (adjusted OR = 2.31; CI: 1.65–3.24; P < 0.001), conversion (adjusted OR = 3.12; CI: 1.83–5.30; P < 0.001), and age (adjusted OR 30– 40 years = 1.06; CI: 0.81–1.67; P = 0.664; 40–50 years = 1.30; CI: 1.00–1.67; P = 0.046; 50–60 years = 1.33; CI: 1.01–1.75; P = 0.044; and >60 years = 1.33; CI: 0.89–1.99; P = 0.159) whereas other procedure at the same time was not significant (adjusted OR = 1.31; CI: 0.90–1.91; P = 0.152).
Institutional Risk Factors Annual hospital volumes were inversely correlated to the risk for serious complications (P < 0.001). Increasing annual hospital volumes decreased the risk for leakage/deep infection (P < 0.001) and pulmonary complications (P = 0.023). The incidence of serious postoperative complications was significantly higher with volumes up to 200 cases per year. Length of operating time and hospital stay were reduced with volumes up to 300 cases per year (Table 4). The risk for serious postoperative complications was higher during the first 400 operations at a specific hospital (P < 0.001). For specific complications, the risk for leakage/deep infection (OR = 1.41; CI: 1.17–1.70; P < 0.001), pulmonary complications (OR = 1.48; CI: 1.10–1.98; P = 0.009), and port-related complications (OR = 1.62; CI: 1.16–2.25; P = 0.005) was increased (Table 5).
DISCUSSION This study is one of the largest studies published on postoperative complications after bariatric surgery, and the follow-up rate was very high (95.7%) and mortality registration was 100%. Compared with other comparable studies, this study is unique in that data are based on patient-specific follow-up and not on records from selected hospitals. All centers in the study used a similar standardized surgical technique and reported complications using a uniform grading system. The data were prospectively collected nationwide in a country with a high bariatric surgery rate. The definition of comorbidity in the register is strictly defined as a condition requiring active treatment. This differs from other studies where the definition has been wider and more vague. Overall complication rates are often difficult to compare between different studies as the definition of complications varies. We have chosen a commonly used and broad definition according to the Clavien-Dindo scale26 with all deviations from a normal postoperative course being counted as a complication. Despite this liberal definition, only 8.7% suffered from a postoperative complication within the first 30 days after surgery. Serious complications were defined as complications resulting in an intervention under general anesthesia, single or multiorgan failure, or death (Clavien 3b or more). These complications are easier to compare with other studies and these complication rates were well in line or lower than those of other comparable studies.15,16 Death within the first 90 days after surgery was 0.04% (11/25,038) which is low compared with the rates described in other studies.15–17,22,24 There are probably several reasons for this low mortality rate. Nearly all patients received prophylaxis for deep venous thrombosis and preoperative antibiotic prophylaxis. The postoperative hospital stay was short due to early ambulation. The operative method and technique were standard at all centers included, and the care received at Swedish intensive care units is of uniformly C 2013 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Early Complications After LRYGBP
TABLE 3. Preoperative Data and Complications Within 30 Days Any Complication
Sex Female Male Age, yrs 60 Coexisting medical condition Sleepapnea Hypertension Diabetes Dyslipidemia Depression Diarrhea Previous venous thromboembolism Smoking No history of smoking History of smoking Body mass index, Kg/m2 60 Waist circumference, cm 140 OS-MRS 0 1 2 3 4 5 Access Laparoscopic Converted Intraoperative Complication No complication Complication Other operation at the same time No Yes
Serious Complications
n
n (%)
OR (CI)
P
19,017 6021
1616 (8.5%) 564 (9.4%)
Reference 1.11 (1.01–1.23)
n
n (%)
OR (CI)
P
0.037
14,208 4529
472 (3.3%) 158 (3.5%)
Reference 1.05 (0.88–1.26)
0.588
4063 7023 8058 4688 1206
296 (7.3%) 554 (7.9%) 698 (8.7%) 502 (10.7%) 130 (10.8%)
Reference 1.09 (0.94–1.26) 1.21 (1.05–1.39) 1.53 (1.31–1.77) 1.54 (1.24–1.91)
0.250 0.009
Early Complications After Laparoscopic Gastric Bypass Surgery: Results From the Scandinavian Obesity Surgery Registry ˚ Erik Stenberg, MD,∗ † Eva Szabo, MD, PhD,† G¨oran Agren, MD,† Erik N¨aslund, MD, PhD,‡ Lars Boman, MD, PhD,§ Ami Bylund, RN, MSc,¶ Jan Hedenbro, MD, PhD,||∗∗ Anna Laurenius, RD, PhD,†† G¨oran Lundeg˚ardh, MD, PhD,‡‡ Hans L¨onroth, MD, PhD,†† Peter M¨oller, MD,§§ Magnus Sundbom, MD, PhD,¶¶ Johan Ottosson, MD, PhD,∗ † and Ingmar N¨aslund, MD, PhD†; For the Scandinavian Obesity Surgery Registry Study Group
Objective: To identify risk factors for serious and specific early complications of laparoscopic gastric bypass surgery using a large national cohort of patients. Background: Bariatric procedures are among the most common surgical procedures today. There is, however, still a need to identify preoperative and intraoperative risk factors for serious complications. Methods: From the Scandinavian Obesity Surgery Registry database, we identified 26,173 patients undergoing primary laparoscopic gastric bypass operation for morbid obesity between May 1, 2007, and September 30, 2012. Follow-up on day 30 was 95.7%. Preoperative data and data from the operation were analyzed against serious postoperative complications and specific complications. Results: The overall risk of serious postoperative complications was 3.4%. Age (adjusted P = 0.028), other additional operation [odds ratio (OR) = 1.50; confidence interval (CI): 1.04–2.18], intraoperative adverse event (OR = 2.63; 1.89–3.66), and conversion to open surgery (OR = 4.12; CI: 2.47– 6.89) were all risk factors for serious postoperative complications. Annual hospital volume affected the rate of serious postoperative complications. If the hospital was in a learning curve at the time of the operation, the risk for serious postoperative complications was higher (OR = 1.45; CI: 1.22–1.71). The 90-day mortality rate was 0.04%. Conclusions: Intraoperative adverse events and conversion to open surgery are the strongest risk factors for serious complications after laparoscopic gastric bypass surgery. Annual operative volume and total institutional experience
are important for the outcome. Patient related factors, in particular age, also increased the risk but to a lesser extent. Keywords: bariatric surgery, laparoscopic gastric bypass, postoperative complications, risk factors (Ann Surg 2014;260:1040–1047)
O
From the ∗ Department of Surgery, Lindesberg Hospital, Lindesberg, Sweden; ¨ ¨ †Department of Surgery, Orebro University Hospital, Orebro, Sweden; ‡Division of Surgery, Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden; §Department of Surgery, Lycksele Hospital, Lycksele, Sweden; ¶Department of Surgery, Ersta Hospital, Stockholm, Sweden; ||Sk˚ane University Hospital, Lund University, Lund, Sweden; ∗∗ Department of Surgery, Aleris Obesity Sk˚ane, Lund, Sweden; ††Department of Surgery, Sahlgrenska University Hospital, Gastrosurgical Research, Sahlgrenska Academy, University of Gothenburg, Gothenburg, ¨ Simrishamn Hospital, Simrishamn, Sweden; Sweden; ‡‡Osterlenskirurgin, §§Department of Surgery, Kalmar County Hospital, Kalmar, Sweden; and ¶¶Department of Surgical Sciences, Uppsala University, Uppsala, Sweden. ¨ Disclosure: This work was supported by grants from the Orebro County Council. None of the authors have received any grants or funding for this study that would be considered as a conflict of interest. However, Jan Hedenbro has received travel grants from Covidien and Johnson & Johnson and received grants from Lund University funds, Crafoord foundation; Erik N¨aslund received grants from Novo Nordisk, Stockholm County Council, Diabetes Theme center at Karolinska Institutet, Stockholm, Sweden; Magnus Sundbom received expert testimony from the Swedish National Board of Health and Welfare and have received royalties from Studentlitteratur for books for medical students. Ingmar N¨aslund received grants from the Swedish government used to run the Scandinavian Obesity Surgery Registry. For the remaining authors, no relevant potential conflicts of interest are declared. ¨ Reprints: Erik Stenberg, MD, Department of Surgery, Orebro University Hospital, ¨ 70185 Orebro, Sweden. E-mail: [email protected]. C 2013 by Lippincott Williams & Wilkins Copyright ISSN: 0003-4932/13/26006-1040 DOI: 10.1097/SLA.0000000000000431
besity is associated with an increased risk for heart failure,1 diabetes,2 and cancer3 ; reduced life expectancy4 ; and quality of life.5 Surgical treatment is currently the best way to establish longterm weight loss,6 and the benefits of surgery for morbid obesity have been reported in numerous studies.6–10 The number of bariatric procedures performed annually has increased dramatically over the last decade, but now it seems to have reached a plateau.11–13 Gastric bypass is the most frequently used method worldwide,12 and in Sweden 96.8% of all bariatric procedures in 2012 were gastric bypasses.13 Complication rates are low,14,15 but given the large number of procedures performed, even low complication rates will cause serious morbidity for a large number of patients and have a great impact upon health care economy. For this reason, it is important to identify risk factors for postoperative complications. In a recent study from the Michigan Bariatric Surgery Collaborative, previous history of venous thrombosis, mobility limitation, coronary artery disease, age more than 50 years, pulmonary disease, male sex, and smoking history were associated with increased risk for postoperative complications,16 as was the annual number of bariatric procedures performed at each hospital.14 Diabetes has also been described as a risk factor.17 While establishing a new procedure, complication rates are high.18,19 The learning curve of the individual surgeon has been described as 50 to 100 cases.18–21 Adverse intraoperative events have also been associated with a higher risk for postoperative complications.22 Whether or not this increase in risk varies depending on the nature of the specific events, however, has not been studied. The Obesity Surgery Mortality Risk Score (OS-MRS) is an accepted score for risk stratification of mortality after bariatric surgery.23 As the mortality rate is low and declining, it would probably be more suitable to have a risk classification score predicting serious complications. The OS-MRS has not been tested for serious complications. Recently, other scores focusing on serious postoperative complications have been proposed.15–17,24 The Scandinavian Obesity Surgery Registry (SOReg) was created to monitor the expanding practice of obesity surgery, its complications, and the effects of treatment. SOReg now covers 98% of all bariatric procedures performed in Sweden,13 providing a unique possibility for analyses. The objective of this study was to review early complications in a prospective material from the SOReg database and to identify preoperative and intraoperative factors predisposing to postoperative complications.
1040 | www.annalsofsurgery.com
Annals of Surgery r Volume 260, Number 6, December 2014
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
METHODS Study Cohort SOReg is a national quality and research register, which started in May 2007. Since 2010, it has covered virtually all bariatric procedures in Sweden and all bariatric centers report to the register. At present, approximately 8000 bariatric procedures (85/100,000 inhabitants) are performed annually at 44 centers. Data are prospectively collected at baseline about 1 month before surgery; during surgery; and after 6 weeks (for days 0–30) and 1, 2, and 5 years of the surgery. The database is compared annually with the Swedish Population Register so as to cover all mortality. The database is validated in several ways. Different variables are systematically searched for nonlogic or improbable values and in such cases compared to the patient’s record. Departments are reminded if follow-up data are missing and the most important variables are mandatory. Regular audit is performed by randomly comparing selected variables with patient records. So far, the database has been shown to include less than 3% incorrect values.13 From the SOReg database, we identified a total of 29,288 patients who underwent a bariatric procedure between May 1, 2007, and September 30, 2012. We excluded patients undergoing bariatric procedures other than laparoscopic gastric bypass. We also excluded revision surgery and planned open gastric bypass surgery, but not
Early Complications After LRYGBP
conversion from laparoscopic to open surgery. After exclusion, a total of 26,173 patients were available for the study. Of these 25,038 had follow-up data from day 30 (95.7%) (Fig. 1). The surgical technique is well-established in Sweden. Of the laparoscopic gastric bypass operations in 2012, 99% had an antecolic, antegastric laparoscopic Roux-en-Y gastric bypass as described by L¨onroth and Olbers.25 The gastrojejunostomy was constructed using 1 linear stapler combined with handsewn closure of the remaining defect in 98.5% of all cases.13 All patients received prophylaxis against deep vein thrombosis.
Comorbidity Comorbidity is defined as a condition needing active pharmacological or continuous positive airway pressure treatment, registered as hypertension, diabetes, dyslipidemia, diarrhea, depression, sleep apnea, or other condition. History of venous thromboembolism and history of smoking were registered from May 1, 2010. OS-MRS was also calculated for patients after this date.
Hospital Volume and Learning Curve Annual surgical volume for each respective hospital at the time of a specific operation was calculated by number of operations in the
FIGURE 1. CONSORT diagram. C 2013 Lippincott Williams & Wilkins
www.annalsofsurgery.com | 1041
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Stenberg et al
6-month period when the procedure took place plus the preceding 12-months period, divided by 1.5. Surgical volumes for the period before the register existed were based on data from the Swedish Board of Health and Welfare. This procedure enabled us to take into account the increasing experience of the departments included. Each case was also categorized depending on the preceding experience at the institution at the time of the operation.
Outcomes Outcomes were measured in 2 ways. First, specific predefined complications and any other complication arising were registered for all patients. The specific complications were leakage or deep intra-abdominal infection, bleeding requiring intervention, gastrointestinal obstruction/ileus, port-related complications, wound dehiscence, other wound complication (mainly superficial wound infections), anastomotic stricture, stomal ulcer, cardiovascular event, pulmonary complication (other than pulmonary embolism), venous thromboembolism, urinary tract infection, and other (in this case specified) complication. Second, the severity of the complication was assessed using the Clavien-Dindo classification.26 This was included in the register 2010 and data are available for all patients operated after January 1, 2010 (n = 18,737). For this group, the main outcome was a serious complication defined as Clavien ≥ 3b (A complication requiring an invasive procedure under general anesthesia, failure of one or several organ systems requiring treatment on an intensive care facility or death).
Statistical Analysis Operative time and hospital stay were analyzed with unadjusted linear regression. Data on overall postoperative complications, serious complications, and specific complications were analyzed with unadjusted logistic regression. P < 0.05 was considered statistically significant. Patient-specific risk factors with a P < 0.10 were entered into a multivariate analysis using stepwise, forward, logistic regression. Significant adjusted risk factors were then entered into a multiple regression analysis with significant surgical risk factors. P < 0.05 was considered statistically significant. Data were analyzed
using IBM SPSS Statistics version 20 (IBM Corporation, Armonk, NY). For odds ratios (ORs), the 95% confidence interval (CI) was also reported.
Ethical Considerations The study was conducted in accordance with the ethical standards of the Helsinki declaration sixth revision and was approved by ¨ the Uppsala/Orebro Regional Ethic Committee. Trial registration is available at clinicaltrials.gov identifier: NCT01862159.
RESULTS Of the 25,038 patients in the study, 76.0% were women. Mean body mass index (BMI) was 42.7 ± 5.43 kg/m2 . Comorbidity, as defined earlier, was present in 48.0% of all patients. Thirty-one percent had a history of smoking (15.8% active smokers, 15.2% previous history of smoking). Baseline, demographic data are presented in more detail in Table 1. In all cases, included surgery began with intention to complete laparoscopically. The conversion rate was 1.1%. The main reasons for converting to open surgery were difficult anatomic conditions (32.1% of conversions) and adhesions (30.9%). Other reasons were intraoperative complications (14.7%), instrument failure (6.4%), problems with ports (4.5%), and other reasons (11.3%). An additional procedure, such as cholecystectomy, hernia repair, and gynecological procedure, was performed at the same time as the gastric bypass in 3.5% of all patients. An intraoperative adverse event detected and managed during the primary procedure occurred in 2.8%. The most common adverse event was unintentional bowel perforation (52.5% of all events), followed by bleeding (20.8%), submucosal stapling or malrotation (10.0%), instrument failure (5.6%), stapling of the nasogastric tube (5.2%), and other reasons (5.9%). A postoperative complication occurred in 8.7% of all patients between day 0 and day 30 postoperatively or, if the primary hospital stay exceeded 30 days, during the course of the stay. Grading of complications was available for all patients operated after January 1, 2010. For patients operated after this date, a postoperative complication was seen in 8.3% of all cases. Specific complications are
TABLE 1. Baseline Characteristics Operated After May 1, 2007 (All) No. individuals, n Sex, n (%) Male Female Age at operation (mean ± SD), yrs Coexisting conditions with ongoing treatment, n (%) Sleepapnea, n (%) Hypertension, n (%) Diabetes, n (%) Dyslipidemia, n (%) Depression, n (%) Diarrhea, n (%) Other condition,∗ † n (%) Previous venous thromboembolism,† n (%) Smoking,† n (%) Active smoking, n (%) Previous history of smoking, n (%) BMI (mean ± SD), Kg/m2 Waist circumference (mean ± SD), cm
Missing Data, n
25,038 6021 (24.0) 19,017 (76.0) 41.1 ± 10.97 12,026 (48.0) 2445 (9.8) 6321 (25.2) 3760 (15.0) 2540 (10.1) 3453 (13.8) 310 (1.2) 2881 (12.7) 422 (1.7) 4691 (31.0) 2386 (15.8) 2305 (15.2) 42.7 ± 5.43 127.2 ± 13.82
Operated After Jan 1, 2010
Missing Data, n
18,737 0 0 0 0 0 0 0 0 0 0 2302 9449 9906 0 4670
4529 (24.2) 14,208 (75.8) 41.2 ± 11.03 9684 (51.7) 1987 (10.6) 5018 (26.8) 2933 (15.7) 2055 (11.0) 2674 (14.3) 289 (1.5) 2801 (15.2) 407 (2.7) 4588 (30.9) 2316 (15.6) 2272 (15.3) 42.5 ± 5.32 127.1 ± 13.65
0 0 0 0 0 0 0 0 0 0 310 3593 3907 0 3170
∗ Other coexisting conditions are (in percentage of the total material) cardiovascular disease (0.6%), pain or mobility limitations (8.6%), psychiatric disorder other than affective disorder (0.1%), systemic disease (0.5%), pulmonary disease (0.7%), and other (0.8%). †Not mandatory parameter until May 1, 2010.
1042 | www.annalsofsurgery.com
C 2013 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
listed in Table 2. Eleven patients died during the follow-up, 7 due to septicemia, 3 due to cardiovascular events, and 1 from pulmonary embolism. There were no further deaths during the first 90 days after surgery. The 90-day mortality was thus 0.04% (11/25,038). Sleep apnea (OR = 5.28; CI: 1.55–18.07; P = 0.008), dyslipidemia (OR = 7.39; CI: 2.25–24.24; P = 0.001), and age more than 60 years (OR = 7.43; CI: 1.97–28.03; P = 0.003) were associated with an increased risk for death within 90 days after the operation.
Patient-specific Risk Factors
Age (P = 0.028), sleep apnea (P = 0.046), and depression (P = 0.048) increased the risk for serious complications. Adjusted analysis on patient-specific variables resulted in significant results for age (adjusted OR 30–40 years = 1.08; CI: 0.82–1.41, P = 0.585; 40–50 years = 1.34; CI: 1.04–1.72, P = 0.025; 50–60 years = 1.40; CI: 1.06–1.84, P = 0.016; and >60 years = 1.48; CI: 0.99–2.20, P = 0.054). BMI did not affect the risk for serious or specific complications. The greater the waist circumference was, the higher the risk for leakage/deep infection (P = 0.005) and other wound complications (P = 0.032). Age specifically increased the risk for leakage/deep
Early Complications After LRYGBP
infection (P < 0.001), bleeding (P < 0.001), strictures (P = 0.003), and other wound complication (P < 0.001). Previous venous thromboembolism, sleep apnea, hypertension, diabetes, dyslipidemia, depression, and diarrhea all increased the risk for postoperative complication. Sleep apnea specifically increased the risk for leakage/deep infection (OR = 1.67; CI: 1.28–2.17; P < 0.001), bleeding (OR = 1.49; CI: 1.16–1.92; P = 0.002), venous thromboembolism (OR = 2.64; CI: 1.07–6.56; P = 0.036), cardiovascular complications (OR = 2.53; CI: 1.33–4.79; P = 0.004), stomal ulcer (OR = 1.82; CI: 1.11–2.98; P = 0.018), and other wound complications (OR = 1.45; CI: 1.02–2.07; P = 0.038). Hypertension increased the risk for leakage/deep infection (OR = 1.32; CI: 1.08–1.62; P = 0.007), bleeding (OR = 1.78; CI: 1.48–2.12; P < 0.001), cardiovascular complications (OR = 4.27; CI: 2.50–7.27; P < 0.001), and pulmonary complications (OR = 1.60; CI: 1.18– 2.17; P = 0.003) but was associated with a lower risk for gastrointestinal obstruction/ileus (OR = 0.72; CI: 0.52–0.98; P = 0.040). Diabetes increased the risk for leakage/deep infection (OR = 1.64; CI: 1.31–2.05; P < 0.001), bleeding (OR = 1.71; CI: 1.39–2.11; P < 0.001), cardiovascular complications (OR = 3.41; CI: 1.98–5.86; P < 0.001), pulmonary complications (OR = 1.80; CI: 1.28–2.53;
TABLE 2. Perioperative Data
Access Laparoscopic, n (%) Converted to open surgery, n (%) Additional procedure, n (%)∗ Length of operation, (mean ± SD), min Length of hospital stay (mean ± SD), d Intraoperative adverse event, n (%) Bleeding, n (%) Unintentional bowel injury, n (%) Instrument failure, n (%) Submucosal stapling or malrotation, n (%)† Stapling of nasogastric tube, n (%) Other complication, n (%) Complication within 30 d, n (%) Leakage or abscesses, n (%) Bleeding, n (%) Wound dehiscence, n (%) Port-related complication, n (%) Other wound complication, n (%) Small bowel obstruction, n (%) Stricture, n (%) Stomal ulcer, n (%) Venous thromboembolism, n (%) Cardiovascular complication, n (%) Pulmonary complication, n (%) Urinary tract infection, n (%) Severity of complication,‡ n (%) No complication I II IIIa IIIb IVa IVb V
Operated After May 1, 2007 (All)
Missing Data, n
Operated After Jan 1, 2010
Missing Data, n
24,773 (98.9) 265 (1.1) 881 (3.5) 77.2 ± 38.10 2.3 ± 3.17 712 (2.8) 148 (0.6) 374 (1.5) 40 (0.2) 71 (0.3) 37 (0.1) 42 (0.2) 2180 (8.7) 449 (1.8) 522 (2.1) 15 (0.1) 140 (0.6) 266 (1.1) 241 (1.0) 59 (0.2) 116 (0.5) 27 (0.1) 56 (0.2) 183 (0.7) 107 (0.4)
0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
18,599 (99.3) 138 (0.7) 633 (3.4) 72.4 ± 34.99 2.0 ± 2.90 508 (2.7) 105 (0.6) 268 (1.4) 32 (0.2) 47 (0.3) 24 (0.1) 32 (0.2) 1546 (8.3) 298 (1.6) 371 (2.0) 9 (0.0) 92 (0.5) 169 (0.9) 197 (1.1) 48 (0.3) 91 (0.5) 18 (0.1) 40 (0.2) 119 (0.6) 77 (0.4)
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
17,191 (91.7) 197 (1.1) 579 (3.1) 140 (0.7) 598 (3.2) 16 (0.1) 8 (0.0) 8 (0.0)
0 0 0 0 0 0 0 0
∗ Other procedure performed at the same time as the bariatric procedure. †Submucosal stapling, malrotation of bowel. ‡Classification according to Clavien-Dindo. Grade I is a complication without need for pharmacological, surgical, endoscopic, or radiological intervention; grade II requires pharmacological treatment; grade IIIa intervention not under general anesthesia; grade IIIb intervention under general anesthesia; grade IVa singe organ dysfunction; grade IVb multiorgan dysfunction; grade V death.
C 2013 Lippincott Williams & Wilkins
www.annalsofsurgery.com | 1043
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Stenberg et al
P = 0.001), stomal ulcer (OR = 1.81; CI: 1.18–2.77; P = 0.007), and other wound complication (OR = 1.59; CI: 1.18–2.13; P = 0.002) but was associated with a lower risk for gastrointestinal obstruction/ileus (OR = 0.65; CI: 0.43–0.99; P = 0.044). Dyslipidemia increased the risk for bleeding (OR = 1.82; CI: 1.45–2.30; P < 0.001), venous thromboembolism (OR = 2.53; CI: 1.02–6.28; P = 0.045), cardiovascular complications (OR = 2.17; CI: 1.12–4.20; P = 0.021), and pulmonary complications (OR = 1.54; CI: 1.02–2.32; P = 0.040). Diarrhea increased the risk for bleeding (OR = 1.91; CI: 1.07–3.43; P = 0.030), venous thromboembolism (OR = 6.42; CI: 1.51–27.21; P = 0.012), and anastomotic stricture (OR = 5.86; CI: 2.11–16.28; P = 0.006). Depression increased the risk for leakage/deep infection (OR = 1.30; CI: 1.01–1.67; P = 0.038), bleeding (OR = 1.38; CI: 1.10–1.73; P = 0.005), pulmonary complications (OR = 1.81; CI: 1.28–2.57; P = 0.001), and stomal ulcer (OR = 1.72; CI: 1.10–2.69; P = 0.016). Previous history of venous thromboembolism increased the risk for postoperative venous thromboembolism (OR = 11.24; CI: 3.65–34.63; P < 0.001) and port-related complications (OR = 3.33; CI: 1.43–7.72; P = 0.005). History of smoking increased the risk for pulmonary complications (OR = 1.53; CI: 1.01–2.30; P = 0.042), other wound complications (OR = 1.47; CI: 1.02–2.12; P = 0.040), and urinary tract infections (OR = 1.92; CI: 1.19–3.11; P = 0.007). Male sex was associated with a higher total incidence of postoperative complications but did not increase the risk for serious complications. Male sex increased the risk for leakage/deep infection (OR = 1.60; CI: 1.31–1.95; P < 0.001), bleeding (OR = 1.60; CI: 1.33–1.92; P < 0.001), and port-related complications (OR = 1.65; CI: 1.09–2.21; P = 0.015) but had a lower risk for gastrointestinal obstruction/ileus (OR = 0.54; CI: 0.37–0.76; P = 0.001). OS-MRS did not predict serious complications. It did, however, predict some specific complications. Higher scores on the OS-MRS were associated with an increased risk for leakage/deep infection (P = 0.004), bleeding (P < 0.001), cardiovascular complications (P = 0.012), port-related complications (P = 0.001), and other wound complications (P = 0.002)
Surgery-specific Risk Factors Conversion to open surgery was associated with an increased risk for serious postoperative complications (P < 0.001), and specifically increased the risk for leakage/deep infection (OR = 5.98; CI: 3.92–9.13; P < 0.001), bleeding (OR = 2.06; CI: 1.12–3.78; P = 0.021), venous thromboembolism (OR = 11.81; CI: 3.53–39.45; P < 0.001), and other wound complications, mainly wound infections (OR = 15.57; CI: 10.62–22.82; P < 0.001). There was no statistically significant difference between the reasons for conversion to open surgery with regards to serious postoperative complications. Conversion for problems with ports (OR = 11.49; CI: 2.51–52.58; P = 0.002), difficult anatomic conditions (OR = 7.66; CI: 3.93–14.92; P < 0.001), intraoperative complications (OR = 6.56; CI: 2.32–18.55; P < 0.001), and adherences (OR = 4.53; CI: 1.96–10.47; P < 0.001) specifically increased the risk for leakage/deep infection. Conversion due to intraoperative complications also increased the risk for postoperative bleeding (OR = 5.43; CI: 1.92–15.32; P = 0.001). The risk for other wound complications was raised for all reasons for conversion. Suffering from an intraoperative adverse event increased the risk for a serious postoperative complication, specifically leakage/deep infection (OR = 3.79; CI: 2.74–5.23; P < 0.001), bleeding (OR = 2.05; CI: 1.40–3.01; P < 0.001), pulmonary complications (OR = 3.56; CI: 2.15–5.90; P < 0.001), stomal ulcer (OR = 2.55; CI: 1.24–5.25; P = 0.011), port-related complications (OR = 2.08; CI: 1.02–4.27; P = 0.045), and other wound complications (OR = 2.83; CI: 1.78–4.49; P < 0.001). Submucosal stapling or unintentional malrotation (10.6% suffered from a serious complication) resulted 1044 | www.annalsofsurgery.com
in the highest risk of a serious complication followed by unintentional bowel injury (9.0%), bleeding (7.6%), and instrument failure (6.2%). For stapling of the nasogastric tube (4.2%) and other reasons (3.1%), the risk was not significantly increased. Performance of an additional procedure at the time of the operation increased the risk for other wound complications (OR = 2.13; CI: 1.33–3.42; P = 0.002) (Table 3). Multiple regression analyses of age- and surgery-specific risk factors resulted in significant results for intraoperative complications (adjusted OR = 2.31; CI: 1.65–3.24; P < 0.001), conversion (adjusted OR = 3.12; CI: 1.83–5.30; P < 0.001), and age (adjusted OR 30– 40 years = 1.06; CI: 0.81–1.67; P = 0.664; 40–50 years = 1.30; CI: 1.00–1.67; P = 0.046; 50–60 years = 1.33; CI: 1.01–1.75; P = 0.044; and >60 years = 1.33; CI: 0.89–1.99; P = 0.159) whereas other procedure at the same time was not significant (adjusted OR = 1.31; CI: 0.90–1.91; P = 0.152).
Institutional Risk Factors Annual hospital volumes were inversely correlated to the risk for serious complications (P < 0.001). Increasing annual hospital volumes decreased the risk for leakage/deep infection (P < 0.001) and pulmonary complications (P = 0.023). The incidence of serious postoperative complications was significantly higher with volumes up to 200 cases per year. Length of operating time and hospital stay were reduced with volumes up to 300 cases per year (Table 4). The risk for serious postoperative complications was higher during the first 400 operations at a specific hospital (P < 0.001). For specific complications, the risk for leakage/deep infection (OR = 1.41; CI: 1.17–1.70; P < 0.001), pulmonary complications (OR = 1.48; CI: 1.10–1.98; P = 0.009), and port-related complications (OR = 1.62; CI: 1.16–2.25; P = 0.005) was increased (Table 5).
DISCUSSION This study is one of the largest studies published on postoperative complications after bariatric surgery, and the follow-up rate was very high (95.7%) and mortality registration was 100%. Compared with other comparable studies, this study is unique in that data are based on patient-specific follow-up and not on records from selected hospitals. All centers in the study used a similar standardized surgical technique and reported complications using a uniform grading system. The data were prospectively collected nationwide in a country with a high bariatric surgery rate. The definition of comorbidity in the register is strictly defined as a condition requiring active treatment. This differs from other studies where the definition has been wider and more vague. Overall complication rates are often difficult to compare between different studies as the definition of complications varies. We have chosen a commonly used and broad definition according to the Clavien-Dindo scale26 with all deviations from a normal postoperative course being counted as a complication. Despite this liberal definition, only 8.7% suffered from a postoperative complication within the first 30 days after surgery. Serious complications were defined as complications resulting in an intervention under general anesthesia, single or multiorgan failure, or death (Clavien 3b or more). These complications are easier to compare with other studies and these complication rates were well in line or lower than those of other comparable studies.15,16 Death within the first 90 days after surgery was 0.04% (11/25,038) which is low compared with the rates described in other studies.15–17,22,24 There are probably several reasons for this low mortality rate. Nearly all patients received prophylaxis for deep venous thrombosis and preoperative antibiotic prophylaxis. The postoperative hospital stay was short due to early ambulation. The operative method and technique were standard at all centers included, and the care received at Swedish intensive care units is of uniformly C 2013 Lippincott Williams & Wilkins
Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Annals of Surgery r Volume 260, Number 6, December 2014
Early Complications After LRYGBP
TABLE 3. Preoperative Data and Complications Within 30 Days Any Complication
Sex Female Male Age, yrs 60 Coexisting medical condition Sleepapnea Hypertension Diabetes Dyslipidemia Depression Diarrhea Previous venous thromboembolism Smoking No history of smoking History of smoking Body mass index, Kg/m2 60 Waist circumference, cm 140 OS-MRS 0 1 2 3 4 5 Access Laparoscopic Converted Intraoperative Complication No complication Complication Other operation at the same time No Yes
Serious Complications
n
n (%)
OR (CI)
P
19,017 6021
1616 (8.5%) 564 (9.4%)
Reference 1.11 (1.01–1.23)
n
n (%)
OR (CI)
P
0.037
14,208 4529
472 (3.3%) 158 (3.5%)
Reference 1.05 (0.88–1.26)
0.588
4063 7023 8058 4688 1206
296 (7.3%) 554 (7.9%) 698 (8.7%) 502 (10.7%) 130 (10.8%)
Reference 1.09 (0.94–1.26) 1.21 (1.05–1.39) 1.53 (1.31–1.77) 1.54 (1.24–1.91)
0.250 0.009
