Studies on factors predicting GORD response to proton-pump inhibitors: NERD subpopulations need to be analysed separately.

We read with interest the paper by Zerbib et al on the factors predicting the response of patients with gastro-oesophageal reflux disease (GORD) to proton pump inhibitors (PPIs) therapy and believe that the results they obtained must be interpreted with caution. The French authors concluded that multivariate analysis, including both clinical and physiological parameters, did not identify any reflux pattern associated with PPI failure. This is at variance with many previous findings, which have clearly shown that an increased oesophageal acid exposure is a robust predictor of reflux response to PPIs. We believe that the surprising result achieved by Zerbib et al is probably due to the fact that only 35% of the patients enrolled in the study had an abnormal oesophageal acid exposure, while the majority of their patients (56%) showed symptoms pertaining to non-erosive reflux disease (NERD) (although it is not clear how 14 patients not identified as NERD or erosive oesophagitis or functional heartburn (FH) have been classified for the analysis and these patients were considered as a unique population instead of separate subgroups as identifiable by pH-impedance. It is well known that the subgroups with hypersensitive oesophagus (HO) and FH have a normal acid exposure. Even though they have performed a subanalysis in patients with only documented GORD (abnormal oesophageal acid exposure and/or positive symptom association analysis) in order to exclude those with FH, this subgroup still contains a greater part of patients with HO (about 50%), who do not have an excess of acid in their oesophagus, which can be responsible for the unexpected finding that increased oesophageal acid exposure is not able to predict a good response to PPIs. Therefore, subgroups of NERD should be analysed as distinct populations instead of pooling them together in order to avoid relevant confounding factors. Also, the presence of functional digestive disorders as predictive factors of PPI failure is further confirmation that many NERD patients studied by them pertained to the subgroups with HO and FH, because previous studies have clearly documented that functional dyspepsia prevails in patients with FH and HO2 and also irritable bowel syndrome is much more represented in patients with NERD than in those with erosive oesophagitis. Moreover, the authors concluded that a BMI #25 kg/m2 is a determinant factor in predicting lack of response to PPIs in all the groups of GORD patients analysed in their study. On the contrary, previous data have clearly shown that an increase in BMI is significantly associated with poor response to PPI therapy. Once again, this result reported by Zerbib et al may be due to the enrolment of many patients with HO and FH in their study, who have been shown to be significantly thinner than those with erosive oesophagitis and NERD with abnormal oesophageal acid exposure. Finally, among the clinical factors having the potential to affect the response to PPI therapy, the authors have not considered the presence of hiatal hernia, which has been shown to play a key role in determining the right dosage of PPIs to obtain therapeutic success. Therefore, we believe that the factors influencing the response to PPI therapy can no longer be analysed by pooling together all the subpopulations included in the GORD realm. In particular, as pH-impedance has allowed us to subdivide the heterogeneous population of NERD, which represents about 80% of reflux patients, into more definite subgroups (NERD with acid excess in the oesophagus and HO) that are clearly separated from FH, we believe that all variables that can potentially guide our clinical management of GORD patients should be evaluated by taking into account the above categorisation and thus the old concept of analysing them as a unique population should be given up