We thank Gray et al for constructive comments to our article regarding vagotomy and Parkinson's disease (PD). 1 These authors have previously hypothesized that the appendix may be a primary source for pathological aggregation of a-synuclein with secondary spread to the brain 2 and contend that our results instead provide support for this hypothesis. In our study, we detected no protective effect following superselective vagotomy, in which only part of the stomach is surgically denervated. In the opinion of Gray et al, this largely excludes a gastric origin for PD initiation. However, this claim is based on the hypothesis that only a single gastrointestinal region serves as the point of initiation, which is not supported by the many reports of widespread a-synucleinopathy throughout the entire gastrointestinal tract. 3,4 Also, Holmqvist et al recently demonstrated that injections of PD brain lysate or recombinant fibrillar a-synuclein into the duodenum of rats is readily transported through the vagal nerve and reaches the dorsal motor nucleus of the vagus. 5 Thus, at least in animal models, it has now been demonstrated that a-synuclein originating outside the stomach and appendix propagates to the brainstem. To us, it seems more probable that many gastrointestinal regions (including the appendix) more or less simultaneously serve as initiation points in PD pathogenesis. In this explanatory framework, only a full truncal vagotomy would efficiently prevent ascending a-synuclein propagation to the dorsal motor nucleus of the vagus. Nevertheless, we have performed an initial reanalysis of our data set, in which we adjusted the analysis for the presence/absence of previous appendicitis. This did not change the risk estimates.Gray et al also wondered whether we were able to discriminate the PD risk reduction after selective vagotomy vis-avis superselective vagotomy. However, owing to issues regarding procedure code validation, we chose to group patients with truncal and selective vagotomy together. These procedures both require pyloroplasty, which has robust entries in our database. For this reason, we cannot comment on any difference in risk reduction between these two vagotomy subtypes. Furthermore, given that truncal and selective vagotomy are grouped together, the effect of truncal vagotomy might actually have been underestimated in our original report.In summary, we find the "appendix hypothesis" of Gray et al highly interesting and an inspiration for important future studies. We do, however, not believe that the observed lack of protection after superselective vagotomy in itself points to any particular gastrointestinal region as being a more important point of origin for further spread.