Quality of life (visual analog scale) before surgery and 3 months and 5 years after conventional Nissen fundoplication (CNF) and laparoscopic Nissen fundoplication in the 2000 randomized controlled trial (LNFI) and the cohort study (LNFII). * P < .05 vs preoperative results for the 3 groups.
Percentage of patients using acid-suppressing drugs before surgery and 5 years after conventional Nissen fundoplication (CNF) and laparoscopic Nissen fundoplication in the 2000 randomized controlled trial (LNFI) and the cohort study (LNFII). * P < .001 vs preoperative results for the 3 groups.
Percentage of patients with esophagitis before surgery and 3 months after conventional Nissen fundoplication (CNF) and laparoscopic Nissen fundoplication in the 2000 randomized controlled trial (LNFI) and the cohort study (LNFII). * P < .05 vs preoperative results for the 3 groups.
Total esophageal acid exposure time before surgery and 3 months after conventional Nissen fundoplication (CNF) and laparoscopic Nissen fundoplication in the 2000 randomized controlled trial (LNFI) and the cohort study (LNFII). * P < .001 vs preoperative results for the 3 groups.
Time curve of percentage of patients free from reintervention after conventional Nissen fundoplication (CNF) and laparoscopic Nissen fundoplication in the 2000 randomized controlled trial (LNFI) and the cohort study (LNFII).
Joris A. J. L. Broeders, Werner A. Draaisma, Jan J. B. van Lanschot, Ivo A. M. J. Broeders, Hein G. Gooszen. Impact of Surgeon Experience on 5-Year Outcome of Laparoscopic Nissen Fundoplication. Arch Surg. 2011;146(3):340–346. doi:10.1001/archsurg.2011.32
To investigate the 5-year effect of surgeon experience with laparoscopic Nissen fundoplication (LNF). In 2000, a randomized controlled trial (RCT) was prematurely terminated because LNF for gastroesophageal reflux disease was associated with a higher risk to develop dysphagia than conventional Nissen fundoplication (CNF). Criticism focused on alleged bias caused by the relative lack of experience with the laparoscopic approach of the participating surgeons.
Multicenter RCT and prospective cohort study.
University medical centers and tertiary teaching hospitals.
In the RCT, 74 patients underwent CNF and 93 patients underwent LNF (LNFI). The complete setup of the cohort study (LNFII) (n = 121) mirrored the RCT, except that surgeon experience increased from more than 5 to more than 30 LNFs per surgeon.
Conventional Nissen fundoplication, LNFI, and LNFII.
Main Outcome Measures
Intraoperative and in-hospital characteristics, objective reflux control, and clinical outcome.
In LNFII, operating time (110 vs 165 minutes; P < .001), dysphagia (2.5% vs 12.3%; P = .008), dilatations for dysphagia (0.8% vs 7.0%; P = .02), and conversions (3.5% vs 7.7%; P = .19) were reduced compared with LNFI. Moreover, in LNFII, hospitalization (4.2 vs 5.6 days; P = .07 and 4.2 vs 7.6 days; P < .001) and in-hospital complications (5.1% vs 13.5%; P = .046 and 5.1% vs 19.3%; P = .005) were reduced compared with LNFI and CNF, respectively. In LNFII, the 6-month reintervention rate was reduced compared with LNFI (0.8% vs 10.1%; P = .002). Esophagitis and esophageal acid exposure at 3 months and reflux symptoms, proton-pump inhibitor use, and quality of life at 5 years improved similarly.
Operating time, complications, hospitalization, early dysphagia, dilatations for dysphagia, and reintervention rate after LNF improve significantly when surgeon experience increases from more than 5 to more than 30 LNFs. In contrast, short-term objective reflux control and 5-year clinical outcome do not improve with experience. In experienced hands, LNF reduces in-hospital complications and hospitalization compared with CNF, with similar 5-year effectiveness and reoperation rate.
Laparoscopic Nissen fundoplication (LNF) rapidly replaced conventional Nissen fundoplication (CNF) as surgical therapy for gastroesophageal reflux disease (GERD) after its first report in 1991.1 As with the introduction of laparoscopic cholecystectomy, evidence from randomized studies that proved equivalence of the new approach was lacking at that time. Therefore, in 1997, a multicenter randomized clinical trial (RCT) was initiated in the Netherlands to compare the effectiveness of LNF (LNFI) with CNF. Interim analysis demonstrated that LNF carries a substantially higher risk to develop dysphagia needing dilatation or reoperation (LNFI, 7 vs CNF, 0; P = .02) than its open counterpart, and consequently, the trial was prematurely terminated.2 The harsh criticism in response to the publication of these results3- 8 focused on alleged bias caused by the relative lack of experience with the laparoscopic approach of the participating surgeons.
We responded by initiating a second cohort study on LNF only (LNFII) to investigate the effect of experience on long-term outcome. In the cohort study, the complete setup was the same as in the RCT, except that surgeon experience had grown from a minimum of 5 LNFs9 to more than 30 LNFs per surgeon in accordance with new insights.10- 13 Participating surgeons, indications for surgery, techniques, and patient management were identical to those in the RCT. As a result, the effect of surgical skill was the only new variable in this second study. The Achilles' heel of studies that have so far evaluated the effect of surgeon experience on the outcome of LNF is that surgical techniques have been changed during the course of the study.9,14- 17 These studies are biased because improvement of surgical technique, as well as increased experience, accounts for the better outcome in the latter patients. The purpose of this prospective study was thus to determine the isolated effect of surgeon experience on intraoperative and in-hospital characteristics, short-term objective reflux control, and 5-year outcome of LNF. These results were compared with the outcome after the initial LNF experience and CNF.
From 1997 to 1999, 177 patients were included in a multicenter RCT to undergo Nissen fundoplication for refractory GERD. Surgical treatment was proposed to patients with heartburn and regurgitation insufficiently responding to at least 40 mg of omeprazole daily. If 40 mg of omeprazole was insufficient to suppress symptoms, the dosage was raised to 40 mg 2 or 3 times daily. If symptoms recurred after return to 40 mg daily, patients were classified as having proton pump inhibitor–refractory GERD. At the time of the interim analysis, 46 patients underwent CNF and 57 patients, LNF. Four patients were never operated on; 6 patients withdrew; and, in the period needed for conducting and discussing the interim analysis, another 64 patients were randomized. Consequently, 74 patients underwent CNF and 93 patients underwent laparoscopic 360° fundoplication (LNFI). All participating surgeons had completed the learning curve of more than 5 LNFs for experienced laparoscopic surgeons and more than 20 for less experienced laparoscopic surgeons, defined by Watson et al.9 The short-2 and long-term18 results of the RCT have been published previously. The operations of the second cohort study were performed by surgeons who participated in the initial RCT, after extension of the learning curve to a minimum of 30 LNFs in accordance with new insights.10- 13 The surgical techniques in the RCT included standardized mobilization of the esophagus over 5 to 7 cm in length, division of the short gastric vessels, posterior crural repair, and creation of a floppy 360° fundoplication of 3 to 3.5 cm and were identical for the cohort study. The cranial fundoplication suture was also stitched to the diaphragm and the caudal suture was fixed to the esophagus at the level of the lower esophageal sphincter (“3-point fixation”). A total of 121 patients were included in the cohort study (LNFII).
Surgical techniques, operating time, blood loss, conversions, splenectomy rate, in-hospital complications, and hospitalization were registered on case record forms. All patients were asked to fill out a questionnaire on quality of life, GERD symptoms, and use of acid-suppressing drugs before surgery and at 3 months and 5 years after surgery. In addition, all patients were asked to rate the change in their reflux symptoms (Visick score) and general health status compared with the preoperative state. At 5 years, patients were asked whether they would choose surgery again in retrospect. Surgical reinterventions, indications for reintervention, and surgical procedures performed were recorded up to 5 years after initial surgery. For objective evaluation, all patients were asked to undergo upper endoscopy and 24-hour pH monitoring both before and at 3 months after surgery. These data and subjective outcome data were collected prospectively in consecutive patients.
The Visick score, the only “overall score” available and validated, correlates well with the most prominent symptom of GERD (heartburn)19 and with a validated questionnaire20 for reflux symptoms21 and was used in this study to score the effect of surgery on symptoms in 4 grades: complete resolution (grade I), improvement (grade II), no effect of surgery (grade III), or deterioration (grade IV), always in comparison with their preoperative state. In addition, the frequency of heartburn, regurgitation, and dysphagia were scored from 0 to 5. The severity of these symptoms was scored from 0 to 3. For the final assessment of these symptoms, a combined frequency and severity grading system was used resulting in a grade from 0 (symptom absent) to 3 (symptom frequent and severe).22 The impact on quality of life was measured using a visual analog scale validated for quality of life assessment after esophageal surgery.23 The scale of this instrument ranged from worst possible health to perfect health (0-100).24
Experienced gastroenterologists evaluated the presence of esophagitis and hiatal hernia size during endoscopy. Initially, the Savary-Miller classification25 was used to grade esophagitis, and later, the Los Angeles classification26 for esophagitis was applied.
Ambulatory 24-hour pH monitoring was performed according to previously described methods,27 after suspending medication that could affect the results at least 7 days beforehand.
For parameters that were scored before and after surgery, only those patients were included for whom both the preoperative data and the results at 3 months or 5 years after surgery were available. The number of patients for whom data were available was determined for each parameter. Continuous variables were expressed as mean and standard error of the mean and were analyzed for differences between the 3 groups using a 1-way analysis of variance and Bonferroni post hoc test. To determine significant effects of surgery on quantitive data, the paired-samples t test was used. The χ2 test was used to compare nominal variables between groups, and effects of surgery on these variables were analyzed using the McNemar-Bowker test. Differences were considered statistically significant at a P value less than .05. Statistical analysis was performed using SPSS version 15.0 (SPSS, Chicago, Illinois).
The study population comprised 288 patients (74 CNF, 93 LNFI, and 121 LNFII). Patient characteristics and mean hiatal hernia size were similar at baseline (Table 1).
All but 3 surgical procedures included mobilization of the fundus with division of the short gastric vessels. Mean length of the fundoplication and the percentage of patients with crural repair and 3-point fixation of the fundoplication were similar in CNF, LNFI, and CNFII (Table 2). In the second LNF cohort, skin-to-skin time was reduced compared with the first LNF group (mean [SEM], 110 [3.7] vs 165 [5.3] minutes; P < .001) and approached the operating time of the open procedure (mean [SEM], 98 [4.3] minutes; P = .22). Blood loss was reduced in the LNFI and LNFII groups compared with the CNF group (mean [SEM], 92  vs 207  mL; P = .04 and 107  vs 207  mL; P = .07). There were no differences in splenectomy rate between LNFI and LNFII (0.8% vs 1.8%; P = .58). In LNFII, the conversion rate was substantially reduced compared with LNFI, but this difference did not reach statistical significance (3.5% vs 7.7%; P = .19). There was a nonsignificant trend toward a lower splenectomy rate in the LNF II cohort compared with the CNF cohort as well (0.8% vs 4.3%; P = .13) (Table 3).
Mean (SEM) hospital stay was reduced from 7.6 (0.7) days after CNF to 5.6 (0.4) days after LNFI (P = .009). Hospitalization decreased even more after LNFII to mean (SEM) 4.2 (0.4) days, both compared with CNF (P < .001) and LNFI (P = .07). In-hospital complications were significantly lower after LNFII, compared with CNF (5.1% vs 19.3%; P = .005) and LNFI (5.1% vs 13.5%; P = .046) (Table 4).
The prevalence of dysphagia was higher after LNFI, compared with both CNF (12.3% vs 0%; P = .01) and LNFII (12.3% vs 2.5%; P = .008). Compared with CNF and LNFII, the percentage of patients who underwent dilatation for dysphagia was higher after LNFI as well (0% vs 7.0%; P = .07 and 0.8% vs 7.0%; P = .02) (Table 4). Quality of life and the percentage of patients using acid-suppressing drugs improved significantly at 3 months compared with preoperatively, with no differences between the groups (Figure 1 and Figure 2).
The effectiveness of surgery in terms of controlling esophagitis and the esophageal acid exposure was high (all P < .001 vs preoperatively), without divergences in any of the 3 groups before or after surgery (Figure 3 and Figure 4).
The self-rated effect of surgery on reflux symptoms was similar in the 3 groups. At 5 years, the 3 groups had similar heartburn, regurgitation, and dysphagia grades. The self-rated change in general health and the percentage of patients who would choose surgery again in retrospect was similar as well (Table 5). Quality of life and the percentage of patients using acid-suppressing drugs remained significantly improved at 5 years compared with the preoperative values, without significant differences between the 3 groups (Figure 1 and Figure 2). However, the use of acid-suppressing drugs increased with time in the 3 groups.
After LNFII, the high early reintervention rate that was observed in LNFI was no longer shown (Figure 5). At 6 months, the reintervention rate was significantly higher after LNFI (10.1%), compared with CNF (1.4%; P = .03) and LNFII (0.8%; P = .002). In the LNFI group, most patients underwent reoperation for persistent dysphagia. The percentage of patients free from reintervention stabilized after 12 months in the LNFI group, while in the CNF and LNFII groups, this took 2.5 years. As a result, the difference in reintervention rate after LNFI (15.2%) was reduced and lost significance at 5 years compared with CNF (11.6%; P = .52) and LNFII (11.6%; P = .46). The shift of the reintervention curve from LNFI to LNFII illustrates the effect of surgeon experience on outcome (Figure 5). The indications for surgical reintervention are presented in Table 6.
In 2000, an RCT was prematurely terminated at interim analysis because LNF for GERD was associated with a higher risk to develop dysphagia than CNF.2 The indication for surgery, construction of the fundoplication, and potential bias caused by relative lack of laparoscopic experience were criticized after the publication of the trial.3- 8 The current study was designed to evaluate whether the unfavorable outcome after LNF was the result of relative inexperience with the laparoscopic approach. A long-term cohort study was initiated, mirroring the methods of the RCT, except for the laparoscopic experience that had grown to more than 30 LNFs per surgeon. Previous studies that evaluate the impact of surgeon experience on outcome are biased because surgical techniques changed during the study. In contrast, the present study further expands on the data currently available in the literature by showing the isolated effect of surgeon experience on long-term outcome of LNF. Operating time, conversion rate, and complications are the most commonly used parameters to define the effect of experience on outcome of antireflux surgery.9,14- 16 However, objective reflux control and long-term subjective outcome would be more relevant in clinical practice. To our knowledge, the current study is the first to objectively document the effect of surgeon experience on 5-year outcome of LNF.
Studies that evaluate the impact of individual surgeon experience are scarce, but several studies have investigated the effect of institutional experience on outcome of LNF. In line with the results of the present study, these reports have demonstrated that experience reduces operating time,9,11- 13,15,28- 32 conversions,9- 13,28,29,32 intraoperative complications,12,30 in-hospital complications,11,30,32 hospital stay,13,31 early dysphagia,10,11,15 and reinterventions.9 In agreement with the present study, the only study that evaluated the effect on acid exposure during postoperative 24-hour pH monitoring did not identify differences between the first 40 and subsequent institutional cases.29 The current study did not focus on the effect of institutional expert experience, but on individual surgeon experience in a pragmatic trial with “all-day practice” applicability. It demonstrates significant improvement of highly comparable parameters.
The results of the current study should be compared with those of 6 earlier studies that report on the impact of individual surgeon experience.9,14- 17,33 The results of the studies that have previously addressed this issue are biased because changes in surgical techniques were introduced in the course of the study.9,14- 17 As a result, the effect of the increase in experience cannot be separated from the impact of changes in surgical technique. In the present study, surgical technique was deliberately left unchanged during the course of the study. The only previous study free of modifications of surgical technique violated the intention-to-treat principle because patients with major complications were excluded from follow-up and this percentage was twice as high in the learning group.33 Despite these sources of bias, the results of the previous studies are in line with the present report and demonstrate that individual surgeon experience reduces operating time,9,14,15,17 conversions,9,14,17 in-hospital complications,15,17 hospital stay,14 early dysphagia,15,17 dilatations for dysphagia,14,16,17 and surgical reinterventions9,14 of LNF. In contrast, the current study and earlier reports show that objective reflux control33 and long-term clinical outcome17 are not correlated with experience. Once a good result is obtained, the effect lasts at least 5 years after surgery, as shown by postoperative reflux symptoms, proton pump inhibitor use, and quality of life.
The current analysis confirms the criticism expressed in 2000 by showing that early outcome of LNF improves with experience increasing from 5 to a minimum of 30 procedures. Moreover, it shows a reduction of blood loss, splenectomy rate, in-hospital complications, and duration of hospitalization, with similar long-term effectiveness and reoperation rate compared with CNF. Two recent meta-analyses comparing LNF and CNF confirm that the most important benefits of LNF are reduction of in-hospital complications and hospital stay, with similar efficacy compared with CNF.34,35 The 10-year results of 2 RCTs comparing LNF and CNF have demonstrated that laparoscopic fundoplication reduces the incidence of incisional hernias36 and incisional hernia corrections.37 In contrast, the current study reports only 1 incisional hernia correction in the first laparoscopic cohort and none in the open group and the second laparoscopic cohort at 5 years. Our group has previously demonstrated that the higher rate of incisional hernias after open surgery is not present at 5 years18 and becomes clinically relevant between 5 and 10 years of follow-up.36,37
In conclusion, this analysis clearly illustrates the effect of experience on the early outcome of laparoscopic antireflux surgery and represents a plea for centralization or concentration of expertise, even more so because the outcome after reoperation is inferior to that of the primary operation.38,39
Correspondence: Hein G. Gooszen, MD, PhD, Department of Surgery, HP G04.228, University Medical Center Utrecht, PO Box 85500, 3508 GA Utrecht, the Netherlands (firstname.lastname@example.org).
Accepted for Publication: March 10, 2010.
Author Contributions: Dr J. A. J. L. Broeders had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors have seen and approved the final version of the manuscript. Study concept and design: van Lanschot and Gooszen. Acquisition of data: Draaisma and Rijnhart–de Jong. Analysis and interpretation of data: J. A. J. L. Broeders, Smout, and I. A. M. J. Broeders. Drafting of the manuscript: J. A. J. L. Broeders and Draaisma. Critical revision of the manuscript for important intellectual content: Rijnhart–de Jong, Smout, van Lanschot, I. A. M. J. Broeders, and Gooszen. Statistical analysis: J. A. J. L. Broeders. Administrative, technical, and material support: Rijnhart–de Jong. Study supervision: Gooszen.
Financial Disclosure: None reported.
Funding/Support: Dr J. A. J. L. Broeders is supported by a University Medical Center Utrecht Alexandre Suerman MD/PhD grant.
Previous Presentation: Oral presentation at the annual meeting of the European Association for Endoscopic Surgery; June 13, 2008; Stockholm, Sweden.