Small-bowel obstruction (SBO) requiring adhesiolysis is a frequent and costly problem in the United States. In 2005, ~119 per 100,000 hospitalized patients had adhesiolysis-related disease and although the incidence has remained fairly constant during the past two decades, estimated yearly costs exceeded $2.3 billion in 2005 and continue to increase [1, 2]. Multiple studies have demonstrated that the majority of patients presenting with SBO secondary to adhesions have a history of one or more abdominal or pelvic operations [3–6]. A wide variety of operations, even those most commonly performed laparoscopically, are frequently implicated as the etiology of adhesive SBO, including appendectomy, gynecological procedures, cholecystectomy, colorectal resections, bariatric operations, and other laparotomies [3, 5–8]. Despite the high frequency of SBO, there is still a paucity of data on the best means of treating this long-term postoperative complication. Conservative measures often are employed initially, but studies have documented failure rates from 20 to 73 %, frequently necessitating the need for operative intervention [9].

Currently, open adhesiolysis is accepted as the standard surgical intervention for adhesive SBO; however, since the first successful laparoscopic adhesiolysis was completed in the early 1990s, many small-scale studies have found this to be a feasible, safe, and potentially less morbid operation if performed in the hands of an experienced surgeon [5–7, 10–17]. In a recent systematic review of 29 studies with 2,005 total patients undergoing laparoscopic management of acute SBO, O’Connor et al. [18] found that 64 % of the operations were completed without conversion to an open procedure with a postoperative morbidity of 14.8 % and mortality of 1.5 %. Despite these findings, no randomized, controlled or prospective, clinical trials have compared the open and laparoscopic approaches for adhesiolysis, indicating that more evidence is needed [19]. This current study examined whether laparoscopic adhesiolysis is associated with lower 30-day mortality, major complication rates, and incisional complication rates than traditional open adhesiolysis.

Materials and methods

The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) includes a systematically and voluntarily collected clinical database of surgical cases at hundreds of hospitals across the United States. Further details regarding NSQIPs parameters, outcomes, data abstraction, and sampling strategy have been previously reported [20]. Briefly, a surgical clinical reviewer collects patient demographics, preoperative comorbidities and labs, intraoperative variables, and 30-day outcomes through chart review and written or verbal communication with patients. The case sampling strategy requires that hospitals capture at least 20 % of each subspecialty’s volume and occurs in consecutive 8-day cycles where NSQIP requires the first 40 general surgery cases to be reported [21]. High-volume procedures (herniorrhaphy, lumpectomy, cholecystectomy) are limited to three procedures each per cycle to increase the number of eligible major procedures captured [22].

The 2005–2010 NSQIP database was queried for patients with a diagnosis of intestinal or peritoneal adhesions with obstruction based on International Classification of Disease, Ninth Revision codes (ICD-9 = 560.8) [23]. Patients with this primary diagnosis were cross-referenced with Current Procedural Terminology (CPT) codes to identify patients undergoing open or laparoscopic adhesiolysis for their SBO [24]. Case selection was limited to CPT codes for open adhesiolysis (44005), laparoscopic adhesiolysis (44180), or exploratory laparotomy (49000) with either open or laparoscopic adhesiolysis coded as the other primary procedure. Patients with open (44120) or laparoscopic (44202) small-bowel resection in addition to adhesiolysis were included. Patients with secondary codes for other major gastrointestinal resections (gastrectomy, colectomy, pancreatectomy, hepatectomy) and hernia repair were excluded to keep the population more homogeneous.

Patient demographic characteristics, such as age, sex, and race (white, black, other/unknown), were included in the analysis. Preoperative functional status was related to the patient’s ability to perform activities of daily living in the 30 days before surgery and was defined as either dependent or independent. Patient’s comorbidities were grouped according to organ system (cardiac, pulmonary, renal, hepatic, neurologic) for analysis. Additional factors considered and grouped categorically were diabetes mellitus requiring oral medication or insulin, body mass index (BMI) >30, >10 % body weight lost in the 6 months preceding surgery, steroid use for chronic conditions, smoking within the past year, chemotherapy within 30 days, radiation within 90 days, disseminated cancer, bleeding disorder, preoperative anemia (hematocrit <36), hypoalbuminemia (albumin <3.5 g/dL), and transfusion >4 U of pRBC in the 72 h before surgery.

Patients with an American Society of Anesthesiologists (ASA) class 3, 4, and 5 were compared categorically to ASA class 1 and 2. A wound class of I (clean) and II (clean contaminated) were grouped and compared to a wound class of III (contaminated) and IV (dirty/infected). Patients were considered to have preoperative systemic inflammatory response syndrome (SIRS) if they had two or more of the following at the same time: temperature >38 °C or <36 °C, heart rate >90 beats/min, respiratory rate >20 breaths/min, white blood cell count >12,000 or <4,000, or anion gap acidosis. Sepsis was considered if the patient had SIRS and a documented infection (positive cultures). Septic shock occurred with documented sepsis and organ or circulatory dysfunction.

Operative characteristics included laparoscopic or open surgery and the presence or absence of a small bowel resection as defined by CPT codes. A case was considered emergent when performed as soon as possible and within 12 h of hospital presentation or development of symptoms. Operative time was recorded in minutes and included as a continuous variable. Postoperative length of stay was also continuous and recorded as days from operation to discharge.

Mortality included any death within 30 days. Postoperative complications occurring within 30 days of surgery were classified as either major or incisional and identified by the affected organ system. This methodology for classification of outcomes has been previously described [21]. Major complications included respiratory (pneumonia, reintubation, or failure to wean from ventilator), cardiac (cardiac arrest requiring CPR, myocardial infarction), renal failure (acute or progressive renal failure), venous thromboembolism (deep venous thrombosis or pulmonary embolism), sepsis or septic shock, organ space infection, neurological (coma >24 h, stroke/cerebrovascular accident), or return to the operating room. Incisional complications included any superficial or deep surgical site infection and wound disruption or dehiscence.

Patient and operative characteristics were compared with 30-day mortality, major complications, and incisional complications using Chi square and Student’s t tests as appropriate to the data. Factors with a p value <0.1 on univariate analysis were included in the multivariable logistic regression for each outcome. A two-sided p value <0.05 was considered significant. Assignment to operative approach was nonrandom creating the potential for residual confounding due to case selection. A propensity score was utilized in the analysis to adjust for this potential case selection bias. The propensity score was developed through a forward stepwise regression aimed at maximizing predictive ability for a laparoscopic approach using entrance criteria of p ≤ 0.1 and exit criteria of p ≥ 0.05 and was assessed using a c-statistic. The propensity score had good predictive ability with a c-statistic = 0.74. Factors included in the propensity score were sex, race, ASA class, BMI >30, hypoalbuminemia, emergency case status, smoking status, steroid use, chemotherapy, radiation, or disseminated cancer, cardiac comorbidity, year of surgery, preoperative sepsis, and small-bowel resection. The propensity score was included as a covariate in the multivariable models for the primary endpoints of mortality and major complications. It did not significantly alter the effect estimate for incisional complications and was thus not included in the final multivariable model. All analyses were performed using IBM SPSS Statistics software, Version 19 (© 2010 SPSS, Inc).

Results

A total of 9,619 patients underwent adhesiolysis between 2005 and 2010. Of these cases, 1,434 (14.9 %) were performed laparoscopically. Patient and operative characteristics are shown in Table 1. The laparoscopic group was more likely to be female, younger, independent in functional status, have a BMI >30, and have fewer preoperative comorbidities and septic events. The mean operative time for laparoscopic adhesiolysis was shorter than for open adhesiolysis (77.2 ± 51.9 vs. 94.2 ± 67.8, p < 0.0001). Open operations were more frequently emergent (51.1 vs. 39.3 %, p < 0.0001) and required a higher percentage of small bowel resections (31.5 vs. 8.1 %). Mean postoperative length of stay was significantly shorter in the laparoscopic group compared with the open group (4.7 ± 5.8 vs. 9.9 ± 9.1, p < 0.0001). Despite differences in the overall laparoscopic and open groups, when these groups were subdivided into quintiles based on their propensity for laparoscopic surgery, they were balanced with respect to patient comorbidities and operative characteristics (see Table 3 in Appendix).

Table 1 Patient and operative characteristics by surgical approach
Full size table

There was a total of 3,282 postoperative complications (incisional, major, and death) within 30-days of surgery affecting 2,548 patients (26.5 % of total group). Table 2 summarizes the rates and composition of 30-day postoperative complications and the crude odds ratios (OD) for laparoscopic versus open surgery. On univariate analysis laparoscopic procedures decreased the relative odds of incisional infection by 84 % and the relative odds of major complication by 61 % compared with open procedures. Differences in major complications between the laparoscopic and open groups were primarily driven by decreased respiratory complications, sepsis or septic shock, organ space infection, and venous thromboembolism in the laparoscopic group. The overall 30-day mortality rate for all patients was 4.2 %. On univariate analysis patients undergoing laparoscopic adhesiolysis had 0.26 times the relative odds of mortality within 30 days compared with patients undergoing open adhesiolysis.

Table 2 Thirty-day postoperative complications by operative approach
Full size table

On multivariable analysis, the effect of operative approach on postoperative complications was adjusted for patient and surgical characteristics. Figure 1 displays the adjusted ORs for 30-day mortality, major complications, and incisional complications with the covariates used in each model listed below. The laparoscopic adhesiolysis group had a 49 % reduction in the relative adjusted odds of 30-day mortality compared with the open group (OR = 0.55, 95 % confidence interval (CI) 0.33–0.92, p = 0.024). After adjustment, the laparoscopic approach also was independently associated with a significant decrease in postoperative incisional complications (OR = 0.22, 95 % CI 0.15–0.33, p < 0.0001) and major complications (OR = 0.7, 95 % CI 0.58–0.85, p < 0.0001).

Fig. 1
figure 1

30-day mortality was adjusted for age, sex, race, patient comorbidities (pulmonary, cardiac, renal, neurologic, hepatic), ASA class, diabetes, bleeding disorders, functional status, preoperative sepsis, steroid use, current smoking status, wound class, anemia, preoperative transfusion, >10 % preoperative weight loss, chemotherapy/radiation/disseminated cancer, BMI >30, operative time, emergency case status, small bowel resection, and propensity for laparoscopy Major complications were adjusted for age, sex, race, patient comorbidities (pulmonary, cardiac, renal, neurologic, hepatic), ASA class, diabetes, bleeding disorders, functional status, preoperative sepsis, steroid use, current smoking status, wound class, anemia, preoperative transfusion, >10 % preoperative weight loss, operative time small-bowel resection, and propensity for laparoscopy incisional complications were adjusted for age, race, pulmonary comorbidities, renal insufficiency, wound class, ASA class, RMI >30, smoking status, bleeding disorders, diabetes, steroid use, operative time, and small bowel resection

Full size image

Discussion

This study demonstrated a significant reduction in 30-day mortality, major complications, and incisional complications associated with laparoscopic adhesiolysis compared with open adhesiolysis for the treatment of acute SBO. Whereas laparotomy has long been considered the standard of care in patients with SBO requiring operation, a minimally invasive approach to adhesiolysis that could potentially reduce postoperative pain and recovery time is appealing. Still the concern remains whether laparoscopic trocars can be placed safely and the operation executed efficiently without injury to dilated bowel. Notably, we found that this is not a procedure, open or laparoscopic, with low morbidity and mortality. Our overall rates of major complications and 30-day mortality in patients undergoing surgical adhesiolysis were 20.4 and 4.2 %, respectively. For an operation that is frequently performed across numerous specialties, these numbers likely reflect a notable burden on the healthcare system.

Similarly, Khaikin et al. [7] noted a 45 % postoperative morbidity for patients undergoing laparotomy for acute SBO, and Suter et al. [5] found that conversion from laparoscopic to open adhesiolysis was associated with an increase in postoperative complications from 12.7 to 55.5 % (p < 0.001). We found slightly lower rates of overall postoperative morbidity in the open adhesiolysis group with 22.2 % of patients having major complications and 10.8 % reporting incisional complications. Postoperative morbidity from laparoscopic adhesiolysis has been reported between 12.7 and 19.2 % and varies largely based on the complications included in this figure [5, 7, 10, 14, 15, 18, 25]. Our laparoscopic group’s complication rate was slightly lower than previously reported rates (10 % major complications and 2 % incisional complications) and had 0.7 times the relative adjusted odds of a major complication and 0.22 times the relative adjusted odds of incisional complication compared with the open group.

This observed marked decrease in complications for laparoscopic cases might well account for much of the reduction in observed 30-day mortality. We found an unadjusted 30-day mortality rate of 4.7 % in the open adhesiolysis group compared with 1.3 % in the laparoscopic adhesiolysis group. Early studies on laparoscopic adhesiolysis quote mortality rates at 2.3 %, rising to 5.5 % in patients requiring conversion to an open procedure [5, 10]. Larger more recent studies are consistent with our findings and have shown mortality for laparoscopic adhesiolysis and open adhesiolysis to be 1.5–1.7 and 3.4 % respectively [18, 26]. With the documented stable rates of ~300,000 yearly operations for SBO during the past two decades in the United States [2], our observed 45 % reduction in the relative adjusted odds of 30-day mortality associated with laparoscopic compared to open adhesiolysis could have a significant impact on both patient outcomes and healthcare system utilization.

The use of laparoscopy for adhesion-related disease is becoming more common, but it still has not been widely adopted. Mancini et al. [26] reported 11.4 % of cases were performed laparoscopically in the NIS database, slightly lower than our findings that 14.9 % of overall operations for adhesiolysis were performed laparoscopically. Given the evolution of training and the increasing comfort of many surgeons with laparoscopic techniques, it is not surprising that our data from 2005 to 2010 may have slightly higher rates than those from 2002. In fact, we noted during the course of 6 years within our data that the laparoscopic rate went from 4.8 % in 2005 to 17.3 % in 2010.

It is important to recognize several of the limitations of this study and the NSQIP database. Given that assignment to the laparoscopic group was nonrandom, there is concern that overall healthier individuals are more likely to be selected for laparoscopic surgery than their debilitated, sicker counterparts. We noted that patients undergoing laparoscopic adhesiolysis were more likely to be younger, female, healthier (lower ASA score, fewer comorbidities), and have less preoperative sepsis. A propensity score was created and utilized to address this selection bias. When patients were grouped into quintiles based on their propensity score the groups were similar with respect to patient and operative characteristics. Additionally, the propensity score slightly reduced the multivariable model effect estimates for mortality and major complications further suggesting this study did control for some baseline differences between the laparoscopic and open cases. Regardless, the fact that we cannot control for all selection bias or completely characterize the surgeon’s operative approach algorithm must be acknowledged in this study. NSQIP does not provide surgeon or hospital specific data, so the individual surgeon’s experience, level of ability, and referral or practice patterns are unavailable. It is conceivable that surgeons who are more skilled and comfortable with laparoscopy or institutions with readily available laparoscopic equipment and staff comfortable with the techniques are the ones undertaking laparoscopic adhesiolysis. This could potentially improve the outcomes in this group, but this cannot be quantified by this database. O’Connor et al. [18] report a conversion rate of 29 % in their systematic review, ranging from 6 to 73 % in the literature; however, based solely on NSQIP coding we are unable to identify accurately cases of conversion from laparoscopic to open. Additionally, this study was unable to capture rates of intraoperative and missed enterotomies, as these variables are not available in NSQIP. Finally, the sampling strategy of NSQIP only captures a small proportion of cases at participating hospitals, so these results may not be widely generalizable.

Despite these limitations, this study remains the largest to date examining the differences in postoperative outcomes between patients undergoing open and laparoscopic adhesiolysis. Our rich clinical data on patient comorbidities, operative factors, and 30-day outcomes allows us to risk adjust for numerous factors that have not been previously considered. We observed a significant reduction in postoperative complications, mortality, and postoperative length of stay with the laparoscopic approach. These results align with much of the current literature suggesting that in the appropriately selected patient, laparoscopy can be a safe, feasible, and potentially beneficial way to approach acute adhesion-related SBOs. Given the high cost and widespread, stable prevalence in the United States, more studies should explore specifically when laparoscopy should be used in acute adhesive SBO.