Long-term outcomes after transoral incisionless fundoplication in patients with GERD and LPR symptoms

Thirty-four consecutive patients underwent TIF using the EsophyX-2 device at our institution between May 2008 and June 2010. Patients undergoing the same procedure at our institution after June 2010 were enrolled in a prospective study (TIF Registry) and were therefore not included. All thirty-four patients were asked to give their permission to gather their baseline, TIF, and follow-up data through a retrospective chart review and were asked to complete a detailed mail-in questionnaire. Follow-up phone calls were made and reminder notices were sent when needed. The data from the 28 patients (82%) who responded were analyzed. Three patients failed to respond, two others could not be reached (having moved out of state), and one patient was terminally ill from ovarian cancer diagnosed in the interval.

Patients considered for surgery had persistent GERD and/or LPR symptoms, which were not controlled or only partially controlled on antisecretory medications, and who were either dissatisfied with their current therapy or unwilling to continue taking medications indefinitely. All potential candidates for antireflux surgery were subjected to our routine diagnostic protocol for surgical fundoplication; a complete history and physical examination, symptom assessment, and other relevant tests. All patients were required to have undergone the following tests: (1) a recent EGD to confirm the diagnosis of GERD and rule out the presence of other esophagogastric pathology; (2) an upper gastrointestinal series (UGI) to better delineate the GE junction anatomy and its measurements; and (3) gastroesophageal high-resolution manometry (HRM) to rule out unsuspected esophageal motility abnormalities that may contraindicate surgery. Twenty-four-hour pH-metry (with either the Bravo wireless system or a catheter-based combined pH/impedance monitoring system) was administered whenever the diagnosis of GERD was uncertain, especially in patients with atypical manifestations. The TIF procedure was considered appropriate and was offered to the patients as an alternative to laparoscopic fundoplication when measurement of the axial height of the hiatal hernia did not exceed 2 cm.

All procedures were performed using the EsophyX device and according to the TIF-2 technique described in a white paper (Bell et al. 2009), which we coauthored. Procedures were performed in the operating room with the authors acting as cosurgeons and with the patients under general orotracheal anesthesia.

The patient is positioned in a left lateral decubitus and in a slight reverse Trendelenburg position. A preprocedure gastroscopy is performed to assess the dimensions of the hiatus, assign a Hill grade to the tightness of the GE valve, and confirm the absence of retained intragastric contents. The EsophyX-2 device is then gently introduced into the stomach transorally over the flexible endoscope while maintaining full endoscopic visualization of the lumen throughout the insertion phase. The stomach is distended using air insufflation from the gastroscope and CO₂ insufflation through the working channel of the scope at a maximum pressure of 12–15 mmHg. A retroflex view is used to observe the tissue mold, elbow, and distal portion of the chassis crossing the GE junction. The back of the mold is then aligned to the lesser curvature. The scope is withdrawn inside the device, allowing for partial closure of the mold at the elbow, then readvanced into the stomach and placed in a retroflex position. The usual orientation landmarks are recognized: the lesser curve, the greater curve, and the anterior and posterior margins of the valve [18]. Three areas (and six locations) are identified for plications: the posterior corner of the valve (the 10 and 11 o’clock locations, to the left of the screen), the anterior corner of the valve (the 1 and 2 o’clock locations, to the right of the screen), and the greater curve (the 5 and 7 o’clock locations at the center of the screen). The deployment of polypropylene H-shaped fastener sets (two fasteners per set) follows the same steps at each location (Fig. 1): the helical retractor is placed into the gastric mucosa within 5 mm of the Z-line, released from the tissue mold, and used to gently retract the tissue caudad. Suction is occasionally applied to the tissue invaginator at this juncture to assist in reducing a small hiatal hernia (if present) and to insure proper positioning of the level of the fundoplication in relation to the Z-line. With gentle tension on the helical retractor, the captured tissue is manipulated into the tissue mold as the mold is closed while desufflating the stomach. The tissue mold and helical retractor are then locked into position. Stylets are then slowly advanced followed by transmural esophagogastric deployment of fasteners over the stylets, one at a time and under direct visualization. We thereby use at least 12 (range = 12–16) plications to create a 240–270° valve by building the anterior and posterior corners of the reconstructed valve and achieving 3–4 cm of vertical length by fixing the gastric fundus to the esophagus at the greater curve. If deemed necessary by visual inspection, we use up to four additional fasteners to reinforce or bolster the fundoplication at the appropriate locations.

Patients received perioperatively a combination of intravenous analgesics, antiemetics, and anticholinergics. Patients were discharged home the day after surgery, were given a prescription for oral narcotics, and asked to continue PPI therapy for 2 weeks. They were allowed to return to work and drive within 3–5 days. Patients were advised to follow a postoperative diet during the first 6 weeks that consisted of liquids (2 weeks), soft/pureed foods (2 weeks), and a soft low residue diet (2 weeks). Patients were asked to avoid lifting anything heavier than 10–15 lb. during the first 3–5 weeks and not to engage in vigorous sports and other strenuous activities for the first 6 weeks.

All patients were seen back in our office approximately 10 days, 6 weeks, and 3 months after surgery and every 6 months thereafter. Patients were asked to contact us upon the return of any symptom. A request to participate in this retrospective study was mailed to all patients along with detailed, disease-specific questionnaires. Data were collected by reviewing the patients’ charts using paper case report forms and then entered into Excel spreadsheets. All data entered was monitored for accuracy and completeness against the source documents in patient charts. Continuous variables were summarized as means and standard deviations or medians and ranges.

Age, weight, height, gender, previous medical history, use of GERD medication, types of GERD symptoms, and duration of GERD were recorded through review of medical charts. GERD Health-related Quality of Life (GERD-HRQL), Gastroesophageal Reflux Symptom Score (GERSS), and Reflux Symptom Index (RSI) are validated questionnaires routinely used in clinical practice for assessing typical and atypical GERD symptoms [13]. All 28 patients completed the GERD-HRQL before TIF prospectively at the time of their physical evaluation. However, all patients were asked to complete preoperative RSI and GERSS questionnaires by recall. The strength of the relationship between the retrospective responses and the prospective GERD-HRQL responses across all domains of the questionnaires was estimated by the Pearson product-moment correlation coefficient. There was a very strong positive correlation between the regurgitation questions (r = 0.90, P < 0.0001); a strong correlation between the abdominal distension and coughing questions (r = 0.77, r = 0.75, P < 0.001 in both cases); a moderate positive correlation between the heartburn questions (r = 0.68, P < 0.001); and a weak, not statistically significant correlation between the dysphagia questions (r = 0.26, P = 0.19). GERD medication use was recorded as “none” if medication was not taken, “occasionally” if any dose was taken 1–3 days per week, and “daily” if any dose was taken 4–7 days per week. Patient satisfaction with overall health condition was assessed as a part of the GERD-HRQL questionnaire and was recorded as “satisfied,” “neutral,” or “dissatisfied.”

The primary clinical effectiveness measure was GERD symptom elimination at follow-up based on score normalization. Typical symptoms were evaluated using GERD-HRQL. The GERD-HRQL is a validated disease-specific questionnaire measuring ten items (6 related to heartburn, 2 to dysphagia, 1 to bloating, and 1 to the impact of medication on daily life) on the VAS scale from 0 (no symptoms) to 5 (worst symptoms) [19, 20]. The scores were indicative of rare or eliminated symptoms if none of the abnormal scores at baseline was >2. The same six questions as those evaluating heartburn were used to evaluate regurgitation scores. The GERSS questionnaire was developed and validated to measure both typical and atypical symptoms associated with GERD: heartburn, regurgitation, abdominal distention, dysphagia, and cough [21, 22]. Each of the five symptoms was scored as a product of severity (from 0 = not at all to 3 = severely) and frequency (0 = never to 4 = daily). The item scores varied from 0 to 12 and the total scores from 0 to 60. Patients with controlled reflux symptoms by either medical or surgical therapy are expected to have a total symptom score of <18 [22]. Atypical symptoms were evaluated using RSI scores. The RSI is a nine-item questionnaire that was developed and validated to measure symptoms associated with LPR, such as hoarseness, throat clearing, excess throat mucus, dysphagia, and cough [23]. The scale for each individual item ranges from 0 (no problem) to 5 (severe problem), with a maximum total score of 45 and a normality threshold at ≤13. The follow-up consisted of symptom evaluation by questionnaires. Scores of ≤2 to each question in the GERD-HRQL, GERSS, and RSI questionnaires were indicative of eliminated symptoms. We considered a 50% or more reduction in scores at follow-up as a clinically significant improvement.

Secondary clinical effectiveness measures were PPI discontinuation and incidence of any unanticipated serious and nonserious adverse events, as a measure of safety. The use of PPIs and other GERD medications such as H2RA and antacids was recorded. A discontinuation of daily PPI use, defined as any dose taken ≤3 days per week, was considered clinically significant.

Patients were stratified into two groups based on their primary symptoms and the responses to the GERD-HRQL questionnaire: a “typical” group that included patients whose predominant symptoms were heartburn and regurgitation, and an “atypical” group, which included patients whose predominant symptoms were extraesophageal in nature, representative of LPR symptomatology.

Statistical analysis was performed using Minitab 15 statistical software (Minitab Inc., State College, PA). Aggregate data, including patient demographics, baseline characteristics, efficacy, safety, and patient satisfaction results, were summarized by descriptive statistics. Mean and standard deviation were generally reported for continuous variables. Median and range were reported for data with skewed distribution. P values for changes at follow-up compared to those at baseline were calculated using the Mann–Whitney U test and the paired t test. Fisher’s exact test was used to compare frequencies. Values with P < 0.05 were considered significant. Univariate and multiple logistic regression analyses were performed to identify various preoperative predictors of success and failure. Various factors included in the regression model, such as BMI (≤30 vs. >30), typical and atypical symptom scores, and presence of esophagitis, were analyzed.

From the 34 patients treated, 28 patients (82%) gave their permission to access their data and were included in the study. All patients had documented chronic GERD for a median of 5 (range = 1–20) years. Median age was 57 years (range = 23–77), BMI ranged from 18.3 to 36.4 kg/m², and only one patient was morbidly obese. Esophagitis was present in 21% of patients, and 14% of patients had short-segment Barrett’s esophagus, confirmed by biopsy. Demographic characteristics are presented in Table 1 and show an even distribution between males and females. The majority of the patients available to follow-up (82%) were younger than 65 years.

The population of GERD sufferers consisted of patients with predominant typical GERD symptoms (n = 15, 54%) and atypical symptoms (n = 13, 46%). Among patients with predominantly atypical symptoms, 92% (12/13) complained about constantly clearing their throat, 77% (10/13) experienced globus sensation, and 69% (9/13) complained about postnasal drip. However, 10/13 (77%) reported troublesome heartburn and regurgitation as a secondary complaint. All patients in the atypical subgroup had abnormal RSI scores (Fig. 2), with inadequate or partial symptom control despite being on daily (12/13, 92%) or occasional (1/13, 8%) PPI therapy. The majority of the patients were dissatisfied with their current health condition (12/13, 92%).

The patients stratified in the typical subgroup suffered from troublesome heartburn and regurgitation. However, 53% of the typical GERD patients reported severe or moderate atypical symptoms as a secondary complaint. Fourteen of 15 (93%) typical GERD patients were dissatisfied with their current health condition despite daily (13/15, 87%) or occasional (2/15, 13%) PPI therapy (Fig. 3).

It appears that patients with predominant typical symptoms experienced marginally better clinical outcomes. However, this conclusion did not reach statistical significance.

All TIF procedures were completed successfully and resulted in creating full-thickness esophagogastric fundoplications with a median of a 270° (range = 240–300°) wrap around the esophagus, and a length of 3 (range = 2–4) cm above the Z-line. The average length of time from introduction of the EsophyX device to post-TIF endoscopy was 55 min. As confirmed endoscopically after TIF, the small reducible hiatal hernia (≤2 cm) present in 75% (21/28) of the patients was completely reduced in all 21 patients. Moderately deteriorated gastroesophageal junctions (Hill grades II and III) were typically corrected to Hill grade I after TIF. There were no perioperative complications, and no instances of hospital readmission or blood transfusion related to use of the EsophyX device. All 28 patients were discharged 1 day after surgery.