Laparoscopic Versus Open Colorectal Surgery

Abstract

Summary Background Data:

Studies comparing the costs of colorectal resection by laparoscopic (LPS) and open approaches are small sized or not randomized. The main purpose of this study is to compare the hospital costs of LPS and open colorectal surgery in a large series of randomized patients.

Methods:

A total of 517 patients with colorectal disease were randomly assigned to LPS (n = 258) or open (n = 259) resection. The following costs were calculated: surgical instruments, operative room (OR) occupation, routine care, postoperative morbidity, and length of hospital stay (LOS). Follow-up for postoperative morbidity was carried out for 30 days after hospital discharge.

Results:

Operative time was 37 minutes longer in the LPS group. Overall morbidity rate was 18.2% (47 of 258) in the LPS versus 34.7% (90 of 259) in the open group (P = 0.0005). The mean LOS was 9.9 (2.6) days in the LPS group and 12.4 (3.9) days in the open group (P < 0.0001). The additional OR charge in the LPS group was €1171 per patient randomized (€864 due to surgical instruments and €307 due to longer time). The saving in the LPS group was €1046 per patient randomized (€401 due to shorter LOS and €645 due to the lower cost of postoperative complications). The net balance resulted in €125 extra cost per patient allocated to the LPS group.

Conclusions:

The present cost-benefit analysis showed a slight additional cost in the LPS group. The better postoperative short-term outcome in patients receiving LPS had a key role to nearly balance the operative room charges due to laparoscopy.

Hospital costs of laparoscopic versus open colectomy were assessed in 517 randomized patients. Cost-benefit analysis showed a slight additional cost in the LPS group. The better postoperative short-term outcome in patients receiving LPS had a key role to nearly balance the operative room charges due to laparoscopy.

Laparoscopic colorectal resection (LPS) was associated with a better immune and inflammatory response, a better postoperative outcome, and a shorter hospital stay compared with open surgery.^1–3 Recent trials reported similar cancer recurrence rate and long-term patient survival comparing LPS and open surgery; thus, concerns about the oncologic adequacy of laparoscopy should be abandoned.^4–6

A major drawback to indicate LPS as a routine technique is the high cost. Studies comparing costs of laparoscopic and open colectomy reported conflicting results.^2,7–21 The large majority of these studies were small-sized or not randomized, and they did not assess the costs of postoperative morbidity in detail. A precise quantification of the healthcare resources consumed by postoperative morbidity is a key point to address correctly the issue of cost-benefit of LPS.

The aim of the present randomized trial is to compare the hospital costs of LPS and open colorectal surgery in a single-center series of 517 randomized patients.

METHODS

Clinical Trial

From February 2000 to December 2003, adult patients admitted to our department for colorectal disease were assessed for study eligibility. The present series is a significant extension of our previously published trials in which the primary endpoint was focused on either short-term postoperative outcome or risk factors for postoperative morbidity.^3,22 Inclusion criteria were age ≥18 years and suitability to elective surgery. Exclusion criteria were cancer infiltrating adjacent organs, organ dysfunctions, ongoing infection, and neutropenia.³ The protocol was approved by the Ethical Committee of the San Raffaele Hospital. The potential participants had the study design explained, and then they were required to sign a written informed consent before randomization.

Eligible participating patients were randomly allocated to LPS or open colorectal surgery. Randomization lists according to the site of the lesion (right colon, left colon, sigmoid colon, rectum) were generated by a computer program.³

All patients were treated on a strictly controlled protocol with regard to bowel preparation, antibiotic prophylaxis, homologous blood transfusion criteria, analgesic administration, feeding, and postoperative care.³ All the operations were performed by the same well-trained surgical team (M.B., A.V., W.Z.).²³ Conversion to open surgery was defined as the need to perform an abdominal incision longer than 7 cm.

Four trained members of the surgical staff who were not involved in the study registered postoperative complications according to a priori definition.²⁴ They also decided the first day of solid food recovery and the day of hospital discharge. Microbiologic analysis and positive culture proved all infectious complications. Patients were discharged after meeting the following criteria: occurrence of bowel movement and full recovery of both deambulation and oral food intake. Follow-up for infectious and noninfectious complications was carried out for 30 days after hospital discharge by weekly office visits.

Economic Calculations

The following costs were calculated: the costs of surgery, the mean in-hospital-related costs of routine surgical care per patient, and the costs of treating postoperative infectious and noninfectious complications.

The cost of surgical instruments has been reported according to the current price in Italy. The cost of operative room (OR) in our hospital is currently €502 per hour.

Resources that were definitively used in the treatment of complications were recorded on a specific electronic record form. The evaluation was solely based on data documented in the patient records. The following items were assessed: complication type and duration in days; laboratory and microbiology analysis; medical, technical, and diagnostic services; surgical and therapeutic interventions; medications; and ambulatory follow-up consultations. Complication costs were calculated separately for the 2 groups, as complication costs may vary under the influence of the type of surgery. In patients who developed multiple complications, resources used to treat each complication were recorded separately.

For all resources used to treat complications, their costs per unit (per day, per service, per analysis, etc.) were gathered. Diagnostic and therapeutic services, pharmaceuticals, and devices were valued according to the National List of Sanitary Costs by the Italian Ministry of Health. Intensive Care Unit (ICU) stay was valued at a flat rate per day (€930), which covers the average daily ICU cost. No patient was scheduled for routine postoperative monitoring in ICU.

The cost of prolonged length of stay (LOS) due to complications covers the cost of board, lodging, routine medical supervision, and nursing. The mean LOS of uncomplicated patients was the basis to calculate the prolonged LOS in each patient with complication. Indirect costs, eg, loss of productivity by the patient, were not taken into account.

Sensitivity analysis was used to evaluate the effect on results when certain underlying parameters are being changed. In this context, we analyzed whether the exclusion of nonsurviving patients (censoring for death) had a major impact on the cost-benefit analysis.

Statistical Analysis

All patients were analyzed on an intention-to-treat basis. Descriptive data are reported as mean (standard deviation), 95% confidence interval (CI), median and range, or number of patients and percentage. Comparison between groups for discrete variables was made by the χ² test or the Fisher exact test when appropriate. Student's t test and analysis of variance for repeated measures were used to compare normally distributed variables. In other situations, nonparametric analysis was performed: the Mann-Whitney U test was used to test differences between groups, whereas the Wilcoxon signed ranks test and the Friedman test were used for comparisons between and among paired data. Post hoc multiple comparisons were adjusted by using the Bonferroni's correction. The P values <0.05 were considered to indicate statistical significance (2-tailed test).

The nonparametric bootstrapping method was applied since traditional statistical methods to analyze the difference in costs are considered inappropriate.²⁵ Because costs are typically not normally distributed (Kolmogorov Smirnov test, P < 0.01), Student's t statistics cannot be used to compare the cost means. A nonparametric approach based on Mann-Whitney U statistics has the disadvantage that assumes that the populations may differ only in their location. Thus, they are not able to adjust for unequal variances, which is very likely to be the case in costs distributions.

No restricting assumptions are needed for the application of the bootstrapping method for its ability to deal with skewness. A bootstrap approach is a resampling technique that provides estimates of the standard error, confidence intervals, and distributions for any statistic. The observed data for costs are treated as an empirical probability distribution that is resampled with replacement many times. Each resampling is used to provide an estimate of cost. The repeated estimates are used to establish an empirical distribution with a sufficient sample size from which tests of hypothesis are constructed. We used 30,000 resamples for the calculations. All costs are expressed in euros (€).

RESULTS

Clinical Outcome

Figure 1 shows the diagram of the trial according to the CONSORT statement.²⁶ Of the 517 patients randomized, 258 were assigned to the LPS group and 259 to the open group. The 2 groups were well balanced for demographics and preoperative parameters (Table 1). No difference between groups was found with respect to Dukes' stage and distribution of benign diseases (data not shown).

FIGURE 1. Study design according to the CONSORT statement. LPS, laparoscopy.

TABLE 1. Demographics and Clinical Characteristics of Patients

Table 2 shows the types of operation carried out and the surgical variables. The mean operative time was shorter in the open than in the LPS group (95% CI, 24.8–49.2, P < 0.0001). The mean operative blood loss was lower in the LPS group (median, 100 mL; range, 50–1600 mL) than in the open group (median, 150 mL; range, 50–1550 mL) (95% CI, 36.5–119.5, P = 0.0003). This translated into a significant reduction of the homologous blood transfusion rate in the LPS group (95% CI, 3.3%–14.3%, P = 0.004). Thirteen (5.0%) patients in the LPS group needed conversion to open surgery (lack of progress due to either adhesions, n = 7 or narrow pelvis, n = 5; development of hypercapnia, n = 1). These patients remained in the LPS arm for data analysis. All patients with rectal cancer had restorative curative R0 resection. A primary diverting stoma was fashioned in 21 patients in the LPS group and in 19 patients in the open group.

TABLE 2. Surgical Details

Three patients died in the postoperative course: 2 in the LPS group (1 for a massive upper gastrointestinal bleeding and 1 for septic shock) and 1 in the open group (septic shock). The overall morbidity rate was lower in the LPS group (47 of 258, 18.2%) than in the open group (90 of 259, 34.7%) (95% CI, 9.1%–24.0%, P = 0.0005). Fewer patients had infectious complications in the LPS group (24 of 258, 9.3%) compared with the open group (53 of 259, 20.5%) (95% CI, 5.1%–17.3%, P = 0.0006). Reoperation was necessary in 15 (5.8%) patients in the LPS group (9 anastomotic leak, 3 adhesion, 2 bleeding, 1 bowel herniation through a 12-mm trocar site) and in 24 (9.3%) patients in the open group (15 anastomotic leak, 6 adhesion, 3 bleeding) (P = 0.19).

Table 3 reports postoperative complications in detail. The LPS group had a lower wound infection rate than the open group (95% CI, 2.2%–12.4%, P = 0.009). In both groups, about one third of postoperative infections occurred after discharge.

TABLE 3. Postoperative Complications in Detail

The mean (SD) length of hospital stay was 9.9 (2.6) days (median, 8 days; range, 5–83 days) in the LPS group and 12.4 (3.9) days (median, 10 days; range, 6–80 days) in the open group (95% CI, 1.9–3.1, P < 0.0001). Recovery of oral food intake occurred after 3.7 (1.3) days in the LPS and 5.0 (2.0) days in the open group (95% CI, 1.1–1.6, P < 0.0001).

Cost Analysis

The extra charge of LPS due to surgical instruments for all surgical procedures was €222,966 (€864 per patient randomized) (Table 4). The longer operative time for LPS had €307 additional charge for OR occupancy per patient randomized. Therefore, the additional OR charge in the LPS group was €1171 per patient randomized.

TABLE 4. Extra Charge of LPS due to Surgical Instruments

A total of 380 patients had an uneventful postoperative course (LPS, n = 211; open, n = 169). In these patients, the mean cost of routine care was the same in both groups (€288/day), whereas the mean LOS was 7.8 (1.5) days in the LPS group and 9.5 (2.4) days in the open group (95% CI, 1.3–2.1, P = 0.0001). This translated in €490 savings for each patient without complication in the LPS group (€288 multiplied by 1.7, which represents the LOS difference between LPS and open groups). Thus, the savings in the LPS group due to 211 patients without complications was €103,390 (€401 per patient randomized).

Table 5 shows the cost of postoperative complications. The additional cost due to patients who developed postoperative complications was €111,516 in the LPS group and €278,004 in the open group, with €166,488 saved in the LPS group (€645 per patient randomized).

TABLE 5. Cost of Postoperative Complications

The overall saving per patient randomized in the LPS group due to both shorter LOS in patients without complications (€401) and lower cost of complications (€645) was €1046. Considering €1171 additional OR charge in the LPS group, there was €125 extra cost per patient randomly allocated to the LPS group. Censoring for death (sensitivity analysis), the extra cost per patient randomly allocated to the LPS group remained substantially unchanged (€119).

The mean cost of postoperative complications was lower in the LPS group than in the open group, but the difference was not significant (95% CI, €−1932 to €670). The determinants of the costs of complications were the resources consumed to treat complications and the additional LOS due to complications. Both were not statistically different comparing the 2 groups. The mean cost (SD) of the resources consumed was €981 (€3812) in the LPS group and €1255 (€5057) in the open group. The mean cost (SD) of additional LOS was €1021 (€1218) in the LPS group and €1369 (€2528) in the open group. When each type of complication was considered separately, we found that the mean cost of wound infection was significantly lower in the LPS than in the open group (95% CI, €−4377 to €−213, P = 0.05). Comparing the 2 groups, no significant difference was found in the mean cost of the other complications.

DISCUSSION

In the current trial, the LPS group had a lower postoperative complication rate and a shorter LOS when compared with the open group. The postoperative infection rate found in the LPS group was similar to the pooled rate calculated by Chapman et al in a systematic review.²⁷ Yet, the postoperative infection rate found in the open group was consistent with intent-to-treat analysis from our previous trial.²⁸ The shorter LOS in the LPS group may reflect the faster postoperative recovery of bowel function, oral food intake, and physical activity.³

A major drawback of LPS is the high cost due to OR charges. In view of the worldwide increasing concerns over exploding costs in medical care, the decision process for adopting new routine treatments should not only weigh clinical benefits and risks but also consider whether these benefits are worth the health resources used. This decision-making process should be informed by cost-benefit analyses of clinical trials. Recently, Nelson et al concluded that the results from major trials provide support to conduct comprehensive cost-effectiveness analyses of laparoscopic colorectal resection.⁵

Sixteen cost-benefit analyses comparing LPS and open surgery were carried out so far, with contrasting results.^2,7–21 Seven nonrandomized studies reported a lower cost for LPS compared with open surgery.^7–13 Of these, 3 were case-matched series,^9,10,13 4 included only either right colectomy^10,12 or sigmoid resection,^9,11 and all but one¹³ were small sized. Four small-sized studies reported lower costs for open surgery than LPS.^2,14–16 One of these was a randomized series of 39 patients who underwent sigmoid resection for polyps.² In the other nonrandomized trials, only patients who underwent right colectomy were included^14,15 or cost-analysis was simply based on OR costs.^15,16 Five studies, 4 of them not randomized, reported similar costs comparing LPS and open surgery.^17–21 One was a case-matched series,¹⁷ 2 were retrospective reviews,^18,19 and in one prospective study the assignment to laparoscopic or open groups was based upon patient request after informed consent.²⁰ In the randomized study,²¹ patients included in the cost analysis were a subset of the entire population of a larger multicenter trial.

To evaluate the hospital costs of both laparoscopy and open surgery in a reliable manner, a precise assessment of health care resources consumed to treat postoperative complications must be performed. In our series, the cost of postoperative complications differed according to the type of complication occurred. The most expensive complications were sepsis, anastomotic leak, abdominal abscess, intestinal obstruction, and wound infection. No significant difference in the mean cost of each postoperative complication was found by comparing the 2 groups, with the exception of wound infection, which was significantly more expensive in the open group. The overall cost of postoperative morbidity was higher in the open group because of the greater number of complication events.

The present cost-benefit analysis showed a slight additional cost in the LPS group. This result was influenced by the better postoperative short-term outcome in patients who received LPS. Indeed, the savings due to both shorter LOS and lower postoperative morbidity rate in the LPS group covered about 90% of the OR extra charge due to laparoscopic technique. Different economic analyses could derive from other trials, which did not show a better postoperative short-term outcome in patients who underwent laparoscopic colorectal resection. In particular, no reduction in postoperative morbidity has been reported in large multicenter trials in which the 20% conversion rate reduced the potential benefits of laparoscopic approach.5,21,29,30 Single-center trials may magnify the advantages of LPS because well-experienced surgical teams perform all operations.^3,4 Conversely, multicenter trials well reflect current surgical practice.5,21,29,30

The present cost-benefit analysis was focused on the hospital costs. Janson et al, who assessed costs due to productivity loss, reported €398 additional cost in the open group compared with the LPS group.²¹ The assessment of community-associated costs, including sick leave, rehabilitation, and full recovery of physical and social performance, would probably favor laparoscopic approach in our series.

Some general limitations about both transferability and reproducibility of the present clinical and economic data should be pointed out. The economic parameter that we used for the present analysis may differ from country to country based on the type of health care and reimbursement system. Moreover, comparable results might be achieved in hospitals where the same type of operations is performed on a similar scale and complication rate.

CONCLUSION

The present cost-benefit analysis showed a slight additional cost in the LPS group compared with the open group. The better postoperative short-term outcome in patients receiving LPS had a key role to nearly balance the operative room charges due to laparoscopy.

Discussions

Dr. Haglund: First, I want to congratulate the authors for a very interesting paper. The laparoscopic technique in colorectal surgery has been around for more than 10 years, but the surgical community has still not decided whether this technique should be a standard one and, if so, when. This decision depends on several different factors, such as costs, time in the OR, hospital stay, postoperative morbidity and mortality, and, not the least important, if the technique has an impact on the long-term effects or not.

I want to stress the importance of the clinical outcome among these factors. You said in your introduction that there are no adverse effects on cancer cure with the laparoscopic technique. How scientifically solid is the evidence for this statement?

You concluded in your abstract that, despite the extra costs, you recommend laparoscopic colorectal surgery because of clinical advantages, which in your study was limited to less risk for wound infections. My second question is then: what type of evidence is there in the literature to support the expectations of less wound infections with the laparoscopic technique in colorectal surgery compared with the open one?

In the March 2005 issue of Annals of Surgery, the group from Copenhagen presented their data on functional recovery after open versus laparoscopic colonic resection from a randomized controlled trial. All patients in their series had enforced multimodal rehabilitation, so-called fast-track surgery. The median hospital stay for both the open and laparoscopy groups of patients in that study was 2 days (range, 2–5 and 2–20 days, respectively). There was no significant difference between the groups in length of hospital stay, postoperative morbidity, mortality, or readmission rate. In your patients, the average length of stay was 7.8 days in the laparoscopic compared with 9.5 days in the open group. This difference was of significant importance on your economic evaluation. However, it seems likely from the Danish data that hospital stay and costs could be reduced significantly compared with the data presented in your study and, furthermore, be very similar in open and laparoscopic surgery. If so, the economic difference would be much greater, more advantageous for open surgery.

Have you made any effort with enforced multimodal rehabilitation in your patients? What factors decided the length of hospital stay in your groups? The Danish study was observer and patient blinded. I understand your study was not; and this is, of course, a very important difference.

The fact that your study was open may have influenced the length of hospital stay by, for instance, prejudgments and expectations that laparoscopic patients should be sent home a couple of days earlier than those having open surgery. I would like you to comment on this issue.

Finally, I want to thank the association for the opportunity to discuss this paper.

Dr. Braga: Thank you, Dr. Haglund, for your comments. I think you raised the most important issues about laparoscopic surgery.

Your first point is the scientific evidence about oncologic adequacy of laparoscopic colorectal surgery. So far, 4 large randomized trials showed no difference in long-term survival when LPS was compared with open surgery. I think that whether the same oncologic principles were respected intraoperatively, there is no reason to sustain that patients undergoing laparoscopic surgery could have a worse prognosis.

In our series, laparoscopy significantly reduced postoperative complications, in particular, wound infections. In general, single-center trials showed significant advantages in patients who underwent laparoscopy, whereas multicenter trials did not probably because of the high conversion rate, the low surgical volume, and the limited training in the majority of surgical groups. Multicenter trials minimize the potential benefits of laparoscopy; however, they truly reflect the current practice nationwide.

Your third point concerns postoperative recovery and length of stay. In my opinion, this is a key point. We were not able to reproduce the results reported by Kehlet et al who also had a high early readmission rate. In our series, length of hospital stay is comparable with other European trials and longer than in U.S. trials for several reasons: in Italy, the homecare service is not optimally organized, we have not outpatient guesthouses, and about one third of our patients are from the south of Italy. All these circumstances make us very careful before discharging patients. Anyway, these factors had the same impact in both groups, and the difference in length of hospital stay was similar when compared with U.S. trials.

Our study was not blind. In my opinion, it is very difficult to blind such a study, especially when follow-up for postoperative complications is carried out for a 30-day period after discharge.

Dr. Jeekel: Concerning your presentation, I think this is an important randomized study, but I missed data of the patients. Last year we completed a large, more than 1000 patients, multicenter study. We found shorter length of hospital stay. Concerning the cancer, I think there is enough experimental evidence that there will be less cancer growth, and I hope we will find less cancer growths in laparoscopic surgery.

If you do the study, then you should know whether the surgeons are able to do it well; in our second study with rectal cancer on which we are embarking now, we do a feasibility study.

What about your participants: a feasibility study where you really show in movies that you are doing these operations well, I think that is a very important issue to do.

About fast track: the recent publications on fast track, I think fast track is not really proven so far; there has been no randomized study to show that fast track is better.

Dr. Braga: I think that the learning curve is a very important issue. Before starting our randomized trial, all the surgical team was well trained in all laparoscopic colorectal procedures.

I agree with you about fast track. The Danish group reported 25% early readmission rate, I think this would not be acceptable in our country. However, to shorten hospital stay will be a more and more important point in the future.

Dr. Senninger: How did you manage pain control after the operation, especially in the open group, because we have seen by giving epidural anesthesia that we could reduce postoperative length of stay to less than 7 days in the average in the open group without a problem? The second thing is: did you have discharge criteria that your staff would follow to let the patient go home?

Dr. Braga: Postoperative pain was managed with both epidural analgesia and intravenous patient-controlled administration of analgesics. In a previous trial, we demonstrated that the control of pain was similar in both groups, but the consumption of analgesic drugs was significantly less after laparoscopic surgery.

The discharge criteria were full recovery of the following: ambulation ability, oral feeding, and bowel function.

Dr. Heberer: I have another brief question regarding the interpretation of the economic analysis. You performed the economic analysis, found that the laparoscopic group is more expensive, and concluded that the laparoscopic procedure should still be preferred. An alternative interpretation could be to ask the question how cost in the laparoscopic group could be driven down. I wonder whether you would also consider this option.

Dr. Braga: I concluded that laparoscopy should be preferred to open surgery because there was a minimum extra cost and the short-term postoperative outcome was better. To reduce costs, the main option could be using nondisposable devices when possible and to push the companies toward a policy of instrument cost reduction.