ORIGINAL ARTICLE ANZJSurg.com
Is laparoscopic colectomy as cost beneficial as open colectomy? Asim Shabbir,* April C. Roslani,* Kutt-Sing Wong,* Charles B. S. Tsang,* Hwee-Bee Wong† and Wai-Kit Cheong* *Division of Colorectal Surgery, Department of Surgery, National University Hospital, and †Clinical Trials and Epidemiology Research Unit, Singapore, Singapore
Key words colectomy, costs and cost analysis, devices, laparoscopy, treatment outcome. Abbreviations LC, laparoscopic colectomy; OC, open colectomy. Correspondence Dr Kutt-Sing Wong, Raffles Surgery Centre, 585 North Bridge Road, # 01-00, Raffles Hospital, Singapore 188770, Singapore. Email:
[email protected] A. Shabbir MRCS, MMed; A. C. Roslani MRCSEd, MS Malaya; K.-S. Wong FRCS Glasgow, FRCS Edinburgh; C. B. S. Tsang FRCS Edinburgh, FICS; H.-B. Wong MSc; W.-K. Cheong FRCS Glasgow, FRCS Edinburgh. Accepted for publication 19 May 2008. doi: 10.1111/j.1445-2197.2009.04857.x
Abstract Background: Laparoscopic colectomy has yet to gain widespread acceptance in costconscious health-care institutions. The aim of the present study was to define the cost– benefit relationship of laparoscopic versus open colectomy. Methods: Thirty-two consecutive patients undergoing elective laparoscopic colectomy (LC) by a single colorectal surgeon between August 2004 and September 2005 were reviewed. Cases were matched with a historical cohort undergoing elective open colectomy (OC) between June 2003 and July 2004. Demography, perioperative data, histopathology and cost were compared. Results: Both groups had similar demographics. Most resections (90.6%) were for cancer. Operative time was significantly longer for LC compared to OC (180 min vs 110 min, P < 0.001). Four patients (12.5%) in the LC group required conversion. LC patients, however, had lower median pain scores (3, 2 and 1 vs 6, 4 and 2 at 24, 48 and 72 h postoperatively, P < 0.001), faster resolution of ileus (3 vs 4 days, P < 0.001) and earlier discharge (6 vs 9 days, P < 0.001) compared to the OC group. As a result, overall hospital cost for both procedures was not significantly different (US$7943 vs US$7253, P = 0.41). Conclusion: Laparoscopic colectomy is as cost-beneficial in the short term as open colectomy.
Introduction Laparoscopic colectomy (LC) is becoming an integral component of the colorectal surgeon’s armamentarium. This is due to its shortterm benefits over the open approach.1 Owing to its perceived greater cost, however, this technique has yet to gain widespread acceptance in cost-conscious health-care institutions. Although there are a number of studies pointing to higher operative costs in LC, very few address total costs (including cost of readmission for complications). For those that do, results are conflicting.2–4 In addition, perioperative management may not have been standardized, particularly in retrospective studies, which could have skewed results. A large randomized controlled trial has demonstrated higher costs in the laparoscopic group but could not show
ª 2009 The Authors Journal compilation ª 2009 Royal Australasian College of Surgeons
any clinical benefit for the latter other than a reduction in wound infections.2 The aim of the present study was to determine whether LC can be performed at a comparable cost to open colectomy (OC), while demonstrating a clinical benefit.
Methods Ethics approval was obtained from the Institutional Review Board of National University Hospital, Singapore. Patients undergoing elective LC by a single colorectal surgeon, trained in the laparoscopic procedure, between August 2004 and September 2005 were reviewed. Informed, written consent was obtained from all patients. Cases were matched for age, sex, comorbidities (diabetes, hypertension,
ANZ J Surg 79 (2009) 265–270
266
ischaemic heart disease), surgical procedure, pathology and Duke’s staging (in cases of cancer), with a historical cohort undergoing elective OC by experienced colorectal surgeons within the unit between June 2003 and July 2004. Prior to August 2004 all colectomies were performed using the open technique, and were chosen for comparison in order to avoid possible selection bias. Perioperative data and total cost were compared between the LC and OC groups. Patients with acute intestinal obstruction, preoperative radiological evidence of locally advanced disease and those with contraindications to pneumoperitoneum were excluded. Diagnoses had been established using colonoscopy, computed tomography of the abdomen and pelvis, and histological confirmation from biopsies in cases of cancer. Previous intra-abdominal surgery was not considered a contraindication for laparoscopic surgery. All patients were managed using the same standard colorectal clinical pathway, which has been in use at National University Hospital since 2001.5 This pathway included pre-defined criteria for resuming feeds and discharge, allowing objective assessment and decision-making by medical staff not directly involved in the study. Patients in the OC group underwent mechanical bowel preparation with 4 L of polyethyleneglycol prior to surgery; those in the LC group did not, owing to a difference in practice among the surgeons. Intravenous antibiotic prophylaxis was administered to all patients (1 g ceftriaxone and 500 mg metronidazole). For cancer surgery, oncological principles were strictly adhered to. The technique of LC has been previously described.6 In general, a lateral-to-medial approach with extra-corporeal vascular division and stapled anastomosis was performed for right-sided resections while a medial-to-lateral approach with intra-corporeal vascular division and double-stapled anastomosis was used in left-sided resections. Hand-assist devices were not used. A four-port technique was used for anterior resection, whereas hemicolectomy was performed using either three or four ports. Intra-corporeal division of lymphovascular pedicles and distal bowel was accomplished using endoscopic staplers (Endopath ETC Endoscopic Linear Cutter, ATW 45; Ethicon Endo-Surgery, Cincinnati, OH, USA). Specimen extraction was performed after release of pneumoperitoneum and the extraction wound protected with a sterile plastic bag. Specimen removal was completed after extra-corporeal division of proximal bowel and marginal vessels. For both open and laparoscopic anterior resections, anastomoses were achieved using a double-stapling technique (Proximate Linear Stapler, TX 30 or Endopath ETC Endoscopic Linear Cutter, ATW 45, and Proximate ILS Curved Intraluminal Stapler, CDH29 or 33; Ethicon Endo-Surgery). This required closure of the extraction site and re-establishment of the pneumoperitoneum in laparoscopic cases. Extra-corporeal functional end-to-end stapled anastomoses (Proximate Linear Cutters, TLC75 or 100; Ethicon Endo-Surgery) were performed for the remaining patients. Polydioxanone sutures (PDS II; Ethicon, Somerville, NJ, USA) were used for en masse closures of extraction sites or midline incisions. Conversions were defined as the need for an abdominal incision >7 cm.2,7
Shabbir et al.
Postoperative pain was recorded using a visual analogue scale of 0–10 and dosage of analgesics was adjusted accordingly. Return of bowel function was measured by the number of postoperative days to tolerance of soft diet. Prolonged ileus was defined as return of bowel function more than 72 h postoperatively. Discharge criteria were as follows: pain score £3 with or without oral analgesics; tolerance of two consecutive oral diets; passage of flatus; and absence of pyrexia for at least 12 h. Total length of stay was standardized to begin from the day of surgery to discharge from hospital. Total cost included all expenses from the day of surgery to discharge, as well as costs resulting from readmissions for complications. These expenses included those arising from laboratory and radiological investigations, room charges, treatment fees, medication, consumables, operating room charges, procedural costs, therapy and counselling. Charges were standardized to current, non-subsidized rates. Complications were assessed up to 30 days postoperatively. Infectious complications were culture-proven.
Statistical analysis All statistical analyses were carried out using SPSS version 14.0 (SPSS, Chicago, IL, USA). Demographics, perioperative and histopathology data between LC and OC were compared. Chi-square or Fisher’s exact tests were used for categorical variables and the Mann–Whitney U-test was used for continuous variables. Because the cost data were skewed and the sample size was small, the recommended bootstrap statistical method8 was used to compare the mean costs between LC and OC. This method makes no assumption on the normality of data and on the equality of the variances or the shape of the distributions. A total of 10 000 simulations have been carried out for each calculation. P < 0.05 were taken to indicate statistical significance.
Results Thirty-two patients underwent LC, and were matched to a similar number of patients undergoing OC. Both groups had similar demographics (Table 1). Age ranged from 29 to 89 years, with a mean of 65 – 13 years. Thirty-five (55%) were male. Twenty-nine patients (90.6%) in each group underwent surgery for cancer, while three (9.4%) needed colectomies for diverticulosis. Low anterior resection was the most frequently performed operation in this series, with 11 cases (34.4%) in each group, followed by high anterior resection (n = 9, 28.1%) and right hemicolectomy (n = 7, 21.8%). For cancer patients there was no difference between groups in terms of the length of bowel resected, tumour size, total number of lymph nodes harvested and resection margins (Table 2). Proximal and distal margins were histologically clear for all patients in both groups. Four patients (12.5%) in the laparoscopic arm were converted to open surgery. Reasons for conversion were intolerance of pneumoperitoneum (one patient), failure to identify the left ureter (one ª 2009 The Authors Journal compilation ª 2009 Royal Australasian College of Surgeons
Cost–benefit of laparoscopic colectomy
267
Table 1 Demographics Parameters
LC (n = 32) n (%)
Age (years) Median (range) 69 (32–89) Sex (ratio) Male : Female 17 (53):15 (47) Co-morbidities Hypertension 14 (44) Diabetes mellitus 8 (25) Ischaemic heart 5 (16) disease Previous abdominal 9 (28) surgery
OC (n = 32) n (%)
67 (29–84)
P
NS
18 (56):14 (47) NS 14 (44) 10 (31) 8 (13)
NS NS NS
7 (22)
NS
LC, laparoscopic colectomy; NS, not significant; OC, open colectomy.
patient) and tumour invasion of surrounding structures (two patients). Although the operative time in LC was significantly longer (180 min vs 110 min, P < 0.001), patients in this group were on average able to tolerate soft diet 1 day earlier, and were discharged 3 days ahead of the OC group. Median pain scores at 24, 48 and 72 h postoperatively were significantly lower for LC compared to OC (3, 2 and 1 compared to 6, 4 and 2; P < 0.001). This was despite a trend (albeit not attaining statistical significance) towards higher analgesic requirements in the OC group (Table 3). There was no significant difference between the two groups in terms of overall postoperative complications (LC, n = 12, 37.5%; OC, n = 10, 31.3%; Table 4). The majority of complications were minor. Surprisingly, the incidence of wound-related complications was higher in LC (6 vs 3), but only two cases were culture-proven wound infections, with one patient requiring readmission within 24 h. This was the sole readmission in the LC group (3.1%). The only case of wound dehiscence occurred in the OC group. Table 2 Histopathology of cancer cases Parameters
LC (n = 29)
Duke’s stage, n (%) A 2 (7) B 12 (41) C 15 (52) Length of resected bowel (cm) Median (range) 16 (5–41) Tumour size (cm) Median (range) 4.0 (0.8–8.3) Total no. lymph nodes harvested Median (range) 14 (4–40)
OC (n = 29)
P
2 (7) 12 (41) 15 (52)
NS NS NS
19 (9–33)
NS
4.5 (0.9–9.0)
NS
16 (5–44)
NS
LC, laparoscopic colectomy; NS, not significant; OC, open colectomy.
ª 2009 The Authors Journal compilation ª 2009 Royal Australasian College of Surgeons
There was no difference in 30-day readmission rates, with only one patient (3.1%) in each group. Mortality was confined to one case in each group (3.1%), both from acute myocardial infarctions. The overall costs for LC and OC were comparable (US$7943 vs US$7253, P = 0.41). Predictably, LC incurred significantly higher operative room charges (US$3595 vs US$3136, P = 0.03) in addition to greater costs from consumables (US$1375 vs US$602, P < 0.0001). Although the additional expenses incurred by the OC group because of longer in-hospital stay and pharmaceutical charges were not statistically significant, they may have been enough to offset the lower procedure charges, hence the similar overall costs. No differences were noted with respect to costs arising from auxiliary services (i.e. nursing, rehabilitation, dietician and counselling services) or in-house laboratory and radiological investigations (Table 5).
Discussion Increasing health-care costs are not only of concern to policy makers but have a significant direct bearing on patient’s choice of treatment and decision-making. Newer procedures should not only be weighed on their technical merits and demerits, but also be scrutinized for their economic feasibility. A true cost–benefit analysis must measure both direct procedure-related costs, as well as indirect costs, including cost to society at large.
Table 3 Perioperative data Variable
LC (n = 32)†
OC (n = 32)
Operative time (min) Median (range) 180 (90–330) 100 (30–195) Maximum pain score: median (range) 24 h 3 (1–7) 6 (2–9) 48 h 2 (0–5) 4 (2–8) 72 h 1 (0–3) 2 (1–7) Analgesia: median (range) Parenteral 40 (5–140) 60 (6–232) narcotic (mg) Oral 8 (2–17) 10 (2–16) acetoaminophen (g) Oral 275 (50–550) 400 (100–800) narcotic (mg) Time to resume diet (days) Median (range) 3 (1–7) 4 (3–7) Length of stay (days) Median (range) 6 (3–18) 9 (6–23)
P
