Introduction

Gastroesophageal reflux disease (GERD) is one of the most common diseases encountered by gastroenterologists, surgeons and primary care physicians [1]. The global prevalence of GERD is 13.98%; in Europe, the prevalence is 14.12% and in North America 19.55% [2]. The estimated number of persons with GERD is 1.03 billion.

According to the World Gastroenterology Organization Global Guidelines, the core principles of GERD management entail lifestyle interventions and reduction of esophageal luminal acid either by local acid neutralization or by suppression of gastric acid secretion using medical treatment, or, rarely, antireflux surgery [3]. Surgical intervention may be considered in large hiatal hernia causing volume-related reflux symptoms with evidence of aspiration or cardia dysfunction, noncompliance with medical treatment, side effects of medical therapy, esophagitis refractory to medical treatment or persistent symptoms caused by refractory GERD [3].

In the surgical guidelines for management of GERD, SAGES recommends either partial or complete fundoplication [4]. If reflux rather than dysphagia is the patient’s main symptom, complete fundoplication should be carried out [4]. If the risk of dysphagia is to be kept to a minimum, partial fundoplication should be given preference [4].

At a multi-society consensus conference, a conditional recommendation with moderate certainty of evidence stated that a patient with GERD would be more likely to benefit from partial rather than complete fundoplication [5]. That recommendation is based on six meta-analyses [7,8,9,10,11,12].

The maximum number of available RCTs is 13 studies with 1564 patients [8]. In that meta-analysis, no significant difference was found in the perioperative complications or the recurrent reflux rates. Often, however, only relatively small subgroups are available to answer individual questions [7]. There is also a paucity of routine comparative clinical studies.

This analysis of prospectively recorded data retrospectively compared Nissen with Toupet fundoplication using propensity score matching. Only patients with axial hiatal hernia and reflux disease and for whom no mesh was used for hiatoplasty reinforcement were included in the present analysis. The main focus of analysis was on the perioperative and 1-year follow-up outcome.

Methods

Registry design

Herniamed is an Internet-based hernia registry in which hospitals and independent surgeons in Germany, Austria and Switzerland can voluntarily document their routine hernia operations. A contract is made with every responsible surgeon where the latter commits to ensuring that data on the hernia operations carried out in the hospital/surgical clinics are completely and correctly entered into the Herniamed Registry. As a prerequisite for documentation in the Herniamed Registry, all patients must sign a separate consent form agreeing to their data being processed in the Herniamed Registry and indicating their acceptance of the follow-up procedure. As part of the information provided to patients regarding participation in the Herniamed Registry, they are told to inform the treating hospital or the treating clinic about any problem occurring after hernia repair. If problems occur after the hernia operation, the patient should attend the treating hospital or surgical clinic for examination.

All complications are recorded up to postoperative day 30. To ensure that all postoperative complications are in fact recorded, the patient is asked about them again on 1-year follow-up. On 1-year follow-up, patients and their primary care physicians are sent a questionnaire asking them about pain and recurrent reflux. If such problems are experienced, patients are requested to attend the hospital/medical practice for further diagnostic examination. Studies have revealed that follow-up by means of a questionnaire sent to the patient and/or their primary care physician is effective [13].

The findings reported in the returned questionnaires are entered into a database by the treating hospital/medical practice and are then available for pseudonymous statistical analysis.

The present analysis of prospectively recorded data retrospectively compared the outcomes of patients with axial hiatal hernia who underwent laparoscopic Nissen or Toupet fundoplication without the use of a mesh for hiatoplasty.

Inclusion criteria for the analysis

From the processed data with a total of 1,152,149 cases, patients were selected for analysis based on the following inclusion criteria:

  • Hiatal hernia operations

  • The entry-state key of the patient master data as well as of the operation data shows status complete, i.e., only completely documented cases for obligatory entries

  • Patients at least 16 years old (only valid entries on age)

  • Only elective hernias

  • Only primary procedures

  • Only procedures with axial location (type I)

  • Only patients with reflux

  • Only operations with Nissen or Toupet fundoplication

  • Only laparoscopic procedures

  • Only procedures without a mesh

  • Operation date up to and including 28/02/2022

  • 1-year follow-up with fully documented visits

Methodology

All analyses were performed with the software SAS 9.4 (SAS Institute Inc., Cary, NC, USA) and intentionally calculated to a full significance level of 5%, i.e., they were not corrected in respect of multiple tests, and each p value ≤ 0.05 represents a significant result.

Data processing

Individual outcome and also independent variables (risk factors, complications) were summarized to global variables. A general, intra- or postoperative complication or risk factor was deemed present if at least one single item applied.

Descriptive and unadjusted statistics

Categorical patient data were presented as absolute and relative frequencies.

For continuous data the mean and standard deviation, or for log-transformed data in case of non-normal distribution, the mean and range (because the range is not symmetrical after back transformation) were presented.

Unadjusted analyses were carried out for analysis of the comparison groups in respect of the individual matching parameters. These were used to describe the patient cohort before matching. For categorical parameters, the asymptotic Chi-square test was performed. For continuous parameters, the robust t test (Satterthwaite) was used.

Propensity score matching

Nissen and Toupet fundoplication outcomes were compared using first a propensity score matching to compare patients with balanced co-variable profiles. Propensity scores were generated using a logistic regression model (Nissen vs. Toupet) with the following matching variables:

  • Age in years

  • Defect size in cm2—logarithmized

  • BMI [underweight/normal weight/overweight/obesity]

  • ASA [I/II/III–IV]

  • Preoperative pain [yes/no/unknown]

  • Regurgitation/reflux [yes/no]

  • Dysphagia [yes/no]

  • Pain [yes/no]

  • Anemia [yes/no]

  • Lung disorders [yes/no]

  • Risk factors [yes/no]

Risk factors were deemed to be present if at least one of the following risk factors applied

  • COPD [yes/no]

  • Diabetes mellitus [yes/no]

  • Aortic aneurysm [yes/no]

  • Immunosuppression [yes/no]

  • Corticoids [yes/no]

  • Smoking abuse [yes/no]

  • Coagulopathy [yes/no]

  • Platelet aggregation inhibitors (discontinued less than seven days ago [yes/no]

  • Coumarin derivatives (quick/INR not in normal range) [yes/no]

  • Non-vitamin K antagonist oral anticoagulants (NOAKs, e.g., Xarelto) discontinued < 2 days [yes/no]*

  • Liver cirrhosis [yes/no]*

*since database update (February 2022).

In addition, the following fixed matching variable was specified:

  • Gender [male/female]*

which had to have the same category in the matched pair.

Propensity score matching was performed using greedy algorithm and a caliper of 0.2 standard deviations. The balance of the matched sample was checked using standardized differences (also given for the pre-matched sample) that should not exceed 10% (< 0.1) after matching.

Analysis of outcome parameters after matching

Matched samples were then analyzed via McNemar’s exact test (test for systematic deviation between the paired patients) for the following perioperative and 1-year follow-up outcomes:

  • Intraoperative complications

  • Postoperative complications

  • General complications

  • Complication-related reoperations

  • 1-year follow-up

  • Pain at rest

  • Pain on exertion

  • Pain requiring treatment

  • Recurrence

  • Trocar hernia

  • Secondary hemorrhage

  • Blood effusion

  • Infection

The results obtained are presented as the concordant cases as well as non-diagonal elements (disadvantageous cases) of the 2 × 2 frequency table, the corresponding p values and the odds ratio (OR) estimates for matched samples with corresponding 95% confidence interval.

In addition, the ORs are illustrated in a forest plot for all outcome parameters.

Results

Patient selection was made in accordance with the patient inclusion flowchart presented in Fig. 1. Following patient selection, 5281 patients were finally enrolled in the analysis to study the relation of the operation technique as well as of other potential influencing variables of the outcome parameters. Here, Nissen fundoplication was used in 2560 (48.5%) cases and Toupet fundoplication in 2721 (51.5%) cases.

Fig. 1
figure 1

Patient inclusion flowchart

Full size image

Descriptive statistics and unadjusted analyses before matching

In this section, differences between the operation techniques for the matching variables were analyzed. Table 1 presents the descriptive statistics as well as the unadjusted test results for the continuous variables age and defect size. Table 2 shows the corresponding values for the categorical variables. Unadjusted analysis of the relationship between the comparison groups and the intraoperative, general and postoperative complications, complication-related reoperations, recurrences, pain at rest, pain on exertion and pain requiring treatment as well as trocar hernias, secondary hemorrhage, blood effusions and infections on 1-year follow-up before matching is shown in Table 3.

Table 1 Presentation of ranges and of unadjusted analysis results for homogeneity between Nissen and Toupet fundoplication for age and defect size
Full size table
Table 2 Presentation of descriptive statistics and unadjusted analysis results for homogeneity between Nissen and Toupet fundoplication for categorical matching variables
Full size table
Table 3 Presentation of descriptive and unadjusted analysis results for homogeneity between Nissen and Toupet fundoplication for outcome variables
Full size table

Matching

Propensity score matching of the 2560 cases treated with Nissen technique with the 2721 cases with the Toupet technique was performed for n = 2290 (89.5%) patients. The standardized differences of the matching variables both before (original sample) and after (matched sample) matching are shown in Fig. 2.

Fig. 2
figure 2

Scatter plot of standardized differences before and after matching

Full size image

For all the matching variables, the standardized difference after matching is less than 10%, indicating good balancing of the variables included in the model.

Statistical analyses

Table 4 gives a summary of the analysis results of the comparison groups for the various outcome parameters. Here, analysis focused on whether there were systematic differences in the outcomes of the Nissen versus Toupet fundoplication operations. No systematic discrepancy was found between the comparison groups for any of the outcome parameters (Table 4, Fig. 3). Also the postoperative dysphagia rate showed with 5.34% after LNF and with 4.64% after LTF no significant difference (p = 0.289).

Table 4 Analysis of systematic differences in the perioperative and 1-year follow-up outcomes between Nissen and Toupet fundoplication (matched pair analysis including odds ratios and corresponding 95% confidence intervals)
Full size table
Fig. 3
figure 3

Forest plot of adjusted odds ratio (OR) including confidence intervals for perioperative and 1-year follow-up outcomes comparing Nissen versus Toupet fundoplication

Full size image

Recurrent reflux following LNF (n = 150) was diagnosed only radiologically in 58 cases (38.7%), in 28 cases (18.7%) radiologically and endoscopically, in 57 cases (38.0% only endoscopically and in 7 cases (4.7%) only clinically. In the patient population following LTF (n = 122), recurrent reflux was found by endoscopy only in 61 cases (50.0%), by endoscopy and radiology in 21 cases (17.2%) and clinically in 21 (17.2%) cases.

Standardized differences for patients with and without follow-up-Information

Figure 4 shows the results of the standardized differences for patients with (n = 5281) and without (n = 2804) follow-up information. Standardized differences above a value of 10% were found only for age. The patients in the analysis population were on average 3.6 years older. Standardized differences of less than 0.1 were found for all other variables, including the perioperative complication rates, and are thus not suggestive of any bias in patient selection.

Fig. 4
figure 4

Scatter plot of standardized differences between patients with and without follow-up information

Full size image

Discussion

Analysis after propensity score matching using prospectively recorded data from the Herniamed Registry demonstrates that for comparative patient cohorts, there was no significant difference between Nissen and Toupet fundoplication with respect to the perioperative and 1-year follow-up outcomes. But this applies only for patients with type I axial hiatal hernias and reflux disease and without mesh augmentation of hiatoplasty.

The perioperative complication rate for both operation techniques was around 6% and additionally 5% for postoperative dysphagia. The complication-related reoperation rate was around 1%. The recurrent reflux rate, in the majority of cases proven by radiology or endoscopy, was around 6% after 1 year for both techniques. Around 8% of patients still reported pain requiring treatment after Nissen or Toupet fundoplication. Pain on exertion was even over 11%. A recent meta-analysis also found a reflux recurrence rate of 11.5% for the Nissen fundoplication and 9.6% for the Toupet fundoplication with no significant difference in the short term [7]. In the long-term follow-up, the reflux recurrence rate also showed no significant difference (LNF 17.3% vs LTF 13.2%) [7]. We do not conclude equality from our data, but that there was no significant difference. There is no prediction on 5-year results from our 1-year follow-up data.

Likewise, the rate of chest pain in the meta-analysis was over 12% [7]. This highlights the fact that surgical treatment of GERD by means of laparoscopic Nissen or Toupet fundoplication is associated with a not inconsiderable postoperative complication rate as well as recurrence and pain rate. Accordingly, surgery should be indicated for GERD on the basis of the international guidelines [1, 2, 4,5,6]. Similarly, the operation technique should be selected in accordance with the recommendations. If the main focus is on treatment of the reflux, preference should be given to Nissen fundoplication, and if the risk of dysphagia is to be kept to a minimum, the recommendation is more for Toupet fundoplication [4]. Due to a lack of evidence, the role of robotic fundoplication cannot yet be evaluated at present [4]. According to the guidelines, surgical repair is as effective as conservative treatment provided that the patients are carefully selected for the operation and the procedure is carried out by an experienced surgeon [14]. The learning curve in antireflux laparoscopic surgery is considered as particularly steep [15]. Registry analyses have weaknesses. First, there is the issue of data accuracy and completeness. All participating institutions have signed a contract with Herniamed stating that the surgeon is responsible for data accuracy and completeness. At the time of audits for certification of hernia centers, the auditor can check data entries in Herniamed.

The Herniamed Registry has been validated against data from the statutory health insurance funds, thus attesting to the necessary accuracy. Since 1-year follow-up information is not available for one-third of patients, it is necessary to rule out selection bias. On analysis of the standardized differences for patients with and without follow-up, values below 0.1 were routinely found, also for the perioperative complication rates, which argue against selection bias.

In summary, it can be stated that for axial hiatal hernia with GERD and appropriate surgical indication based on fundus manometry test results, no systematic differences between the Nissen and Toupet operation could be found. Since the perioperative complications as well as the recurrence and pain rates are of relevance, these operations should be performed by experienced surgeons.