Yeah, lots of risk for confounders and selection bias in a study like this. If you take the huge number of patients with a history of GI bleeding (for which PPI therapy is well-proven), those are patients who already have an elevated all-cause mortality risk. They also selected patients who had at least one creatinine measure in their EHR; so probably people who are more likely to have background renal disease. And using US veterans makes it hard to generalize to the greater population. However, they also saw elevated risk in patients who didn't have a de-facto indication for PPIs (GERD, GI bleeding, etc), so maybe there is something to this.

Honestly, it's kind of an amateur study design. I think the conclusion to draw here is we need a large randomized control study to look at this issue.

Regardless, as a physician, I agree with the study authors that if a patient doesn't have a specific indication for PPI, they shouldn't be on one. Then again, I don't need a study to tell me that; it's my feeling about ALL medications.

There’s an obvious issue that you stomach acid is there to digest food correctly and also provide the environment for growing the correct microbiome; both of which are pretty vital functions I’d say.

They didn't find a similar effect with H2 blockers, which pretty strongly suggests a specific problem caused by the mechanism of PPIs.

Only if the choice between PPI and H2 blockers is random. If there's some self-selection going on, there isn't much to say about PPI's impact on patient health given what we know about H2.

This study is bogus because they didn't analyze cause of death or account for the reason for using PPI.

H2 blockers and PPIs are often not prescribed for the same things. For many serious diseases PPIs are commonly prescribed as a secondary medication (eg with Plavix) where H2 blockers would make no sense. That's why the group sizes differ greatly.