Typical symptoms and not positive reflux-cough correlation predict cure of gastroesophageal reflux disease related chronic cough after laparoscopic fundoplication: a retrospective study

Gastroesophageal reflux disease (GERD) is a condition which develops when the reflux of stomach contents causes troublesome symptoms and/or complications [1, 2]. The typical symptoms of GERD are heartburn and regurgitation, while the atypical symptoms of GERD include non-cardiac chest pain and extraesophageal symptoms such as chronic cough, chronic asthma, chronic laryngitis and dental erosions [3‐5]. The prevalence of GERD in East Asia is about 2.5–7.8% [6]. Chronic cough defined as cough that persisits for > 8 weeks, affects 11–20% of the adult population and significantly impairs health-related quality of life [7]. GERD symptom response rates for anti-reflux surgery varies based on 1) evidence for GERD, such as ambulatory pH monitoring metrics, PPI response and GERD phenotypes, and 2) the particular symptom [3]. while laproscopic fundoplication (LF) can control typical symptoms in about 90% of patients [8‐13], the resolution of chronic cough is less predictive (about 51–96%) [9, 10, 14, 15]. Impedance-pH monitoring is widely used in the diagnosis of GERD, and its ability to simultaneously record symptoms allows evaluation of reflux-symptom correlation [16]. The aim of this study was to investgate if positive reflux-cough correlation predicts resolution of chronic cough after antireflux surgery.

Between January 2013 and December 2017, a total of 1636 patients underwent laparoscopic fundoplication, among which 97 patients meet the criteria for further follow up (Fig. 1). Fourteen patients couldn’t be reached and 4 patients died (1 for leukemia, 1 for pulmonary fibrosis and two unclear), thus 79 patients were included in this study. Among which 30 patients underwent Nissen-Rossetti procedure, 34 underwent Nissen procedure, 14 under Toupet procedure, and 1 underwent Dor procedure.

The demographics factors and symptomatic outcomes among the 79 patients were listed in Table 1 and Table 2, respectively. At a median follow-up time of 35 months after surgery (range 7–65 months), symptomatic outcome of chronic cough was rated cured in 47 (59.5%) and excellent in 10 (12.7%) of the 79 patients. Heartburn was rated cured in 47 (79.7%) and excellent in 5 (8.5%) of the 59 patients. Regurgitation was rated cured in 46 (74.2%) and excellent in 7 (11.3%) of the 62 patients.

At the univariate level (Table 3), three factors were significantly associated with the cure rate of chronic cough. These included typical symptoms, Number of Reflux episodes (impedance) and Ineffective esophageal motility. At the multivariate level, all the factors listed in Table 3 together with BMI (overweight, obesity vs normal) and type of surgery (Nissen-Rossetti, Nissen vs Toupet) were offered to the regression procedure. Present of typical symptoms (OR = 6.435, 95%CI = 1.427–29.032, p = 0.015) and number of Reflux episodes (impedance) ≥73 (OR = 0.306, 95%CI = 0.107–0.874, p = 0.027) were independently associated with the cure of chronic cough (Table 4). we also performed the same analysis procedure on cure+excellent vs fair+poor and cure vs fair+poor of chronic cough, and found that only present of typical symptoms was the independent predictor (OR = 4.114, 95%CI = 1.091–15.522, p = 0.037; OR = 5.857, 95%CI = 1.290–26.590, p = 0.022).

Among the 79 patients, 19 patients didn’t have pathologic AET and positive reflux-cough association. Fourteen of which had their cough cured or significantly improved (Table 5).

Seventy-nine patients with chronic cough under LF were included in our study, At a median follow-up time of 35 months after surgery, chronic cough, heartburn and regurgitation were cured or significantly improved in 57(72.2%), 52(88.1%) and 53(85.5%) patients. By using multivariate model, we found that present of typical symptoms (OR = 6.435) and number of Reflux episodes (impedance) ≥73 (OR = 0.306) but not AET and reflux-cough association were independently associated with the cure of chronic cough. GERD, Asthma, and Postnasal drip syndrome, alone or in combination, were responsible for 93.6% of the cases of chronic cough [21, 22]. The 2016 ACCP guidelines recommended a clinical profile to excluding other potential chronic cough causes, and the clinical profile was estimated to be 91% predicive that a patient’s cough would respond to anti-reflux treatment [23]. In our study, asthma and non-asthmatic eosinophilic bronchitis were not excluded, so the respond rate was lower (72.2%).

Francis [24] found that preoperative heartburn with or without regurgitation and esophageal acid exposure of greater than 12% at baseline were significant predictors of response of the primary extraesophageal reflux symptom to anti-reflux surgery. In our study, present of typical symptoms was consistently significantly associated with good outcome in three different multivariate analysis. Number of Reflux episodes (impedance) ≥ 73 was only significantly associated with the cure of chronic cough. Thus we believed that typical symptoms was the only predictor in our study. We utilized a favorable response (cure+excellent) to LF as the “gold standard” test for comparison with present of typical symptoms (consistency test) and found that the consistency was poor (κ value:0.219, p = 0.033), thus although typical symptoms can predict better outcome, its ability to diagnose GERD-related cough is poor.

The recent Lyon consensus on ambulatory reflux monitoring illustrated that SI and SAP have a predictive value for the effect of medical and surgical treatment of reflux disease, and this is independent of AET [2]. While data about predictive role of SI and SAP on outcome of LF for chronic cough is still limited and controversial. Marco [9] found that even single time of correlation between cough and reflux could improve cure rate of chronic cough, but the sample size was too small (8 patients in one group and 10 in the other). Michael [25] found that positive SAP independently predicted good outcome of chronic cough, but most patients were under PPI therapy and only pH mornitoring was used. In contrary, Francis [24] found that no difference in response to surgery based on SI/SAP parameters, but the sample size was small too (12 cough patients). Thus our study is basically the first one to investigate the role of reflux-cough correlation on outcome of chronic cough after anti-reflux surgery. We found that positive SI/SAP was not significantly associated with the cure rate of chronic cough. The following two studies may provide a perspective on the reason. Paterson and Murat [26] showed that using a diary or event marker for determination of cough is inadequate because patients underestimate the frequency of cough events or misreport their timing. Sifrim and Dupont [27] found that the number of cough bursts detected by manometry was significantly higher than by patients using the event marker, and the delay was 28 (7–80) seconds, most of all, they found that the delay could change cough-reflux pattern to reflux-cough pattern. Thus correctly calculation of SI/SAP was impossible, not to mention using them as predictive factor for better outcome of LF.

DeMeester scores and AET were also not significantly associated with good outcome of chronic cough. The reason may be that pathologic AET are not the only implication for considering LF. In our center, other evidence of GERD such as hiatal hernia, typical symptoms, No. of reflux episodes, DeMeester scores and PPI response are under consideration. Thus symptomatic outcome of patients without pathological AET may similar to those with pathological AET. As we can see that 14 of 19 patients without pathologic AET and positive reflux-cough association had their cough cured or significantly improved after LF.

There were several limitations in our study. First, as the dose of PPI, duration of PPI, and response of chronic cough to PPI was not standardized in our medical record, thus PPI response was not included. Since a positive response to initial empiric PPI therapy is the best indicator for eventual resolution of gastroesophageal reflux-related chronic cough [28, 29], thus the lack of this part may have an impact on the results. Further, this is a retrospective study and the symptomatic outcome were relied on telephone interview, the patients may misunderstood what we asked, as the cure of chronic cough was easy to judge, thus we adopted cure of chronic cough as the outcome metrics to minimize the impact.

The clinical profile to excluding other potential chronic cough causes is useful in selecting patients most like to benefit from LF, and present of typical symptoms but not postitive reflux-cough correlation predicts cure of chronic cough. Futher prospective study using manometry to detect cough is needed to illustrate the association between reflux-cough correlation and symptomatic outcome of chronic cough after LF.