Read on Twitter
Thank you @BrennanSpiegel for trying to provide insight into limitations and how to interpret the results.
Unfortunately, this will create more panic among media, patients, and clinicians than actually change management of patients with #COVID19. In this thread, I explain why: https://twitter.com/BrennanSpiegel/status/1280527527325265920
Unfortunately, this will create more panic among media, patients, and clinicians than actually change management of patients with #COVID19. In this thread, I explain why: https://twitter.com/BrennanSpiegel/status/1280527527325265920
2/ Most reported associations in observation clinical research are FALSE.
-To be credible, you need higher magnitude of effect as shown below.
-Weaker associations are probably due to unmeasured confounders/bias
-To be credible, you need higher magnitude of effect as shown below.
-Weaker associations are probably due to unmeasured confounders/bias
3/ This has been proven again and again with all the retrospective/observational studies with PPIs.
Below from @AGA_Gastro about "associations" that have been shown with PPIs.
-Majority of these did not hold true in prospective studies (outside of enteric infections)
Below from @AGA_Gastro about "associations" that have been shown with PPIs.
-Majority of these did not hold true in prospective studies (outside of enteric infections)
4/ Thought exercise to prove the above point on studies with PPI.
This study used "falsification approach" to estimate whether PPI use is also implausibly associated with other common medical conditions for which no known pathophysiologic link exists.
This study used "falsification approach" to estimate whether PPI use is also implausibly associated with other common medical conditions for which no known pathophysiologic link exists.
5/ Here is what they did:
6/ Not surprisingly, they found that PPI use was "associated" with everything they looked at, which shows again that these types of study designs will nearly always show some association.
7/ One can try to use Hill Criteria to determine if observational studies might hold "true" impact, as mentioned by @BrennanSpiegel about biological gradient and plausibility.
8/ So, this study @AmCollegeGastro and #COVID19 is problematic due to ORs (magnitude of effect) is mostly within zone of bias (OR for Qday PPI was 2.15 and BID PPI was 3.67).
-Problem: Unmeasured confounders (what if pts on PPI are more likely to not social distance/wear mask)
-Problem: Unmeasured confounders (what if pts on PPI are more likely to not social distance/wear mask)
9/ This wasn't asked in the survey and looking at demographics, majority of the cohort that tested positive for COVID 19 (85.9%) were <39 y/o and 69.6% had high school degree or less.
10/ The clinical impact of this study is hard to discern:
-If you are on BID PPI, yes, you should always try to go down to Qday PPI (lowest effective dose). This is true because it is good medical practice (not because of this study showing COVID-19 risk) #GERD
-If you are on BID PPI, yes, you should always try to go down to Qday PPI (lowest effective dose). This is true because it is good medical practice (not because of this study showing COVID-19 risk) #GERD
11/ To simplify for patients/media
-IF you are NOT on PPI, doesn't mean you are "lower risk"
-If you are ON PPI, doesn't mean you are "higher risk"
-Regardless of your PPI status, EVERYONE should wear a mask/practice social distancing.
#COVID19 @AmCollegeGastro @AmerGastroAssn
-IF you are NOT on PPI, doesn't mean you are "lower risk"
-If you are ON PPI, doesn't mean you are "higher risk"
-Regardless of your PPI status, EVERYONE should wear a mask/practice social distancing.
#COVID19 @AmCollegeGastro @AmerGastroAssn
12/ Please talk to your physician before stopping your acid blocking medications.
-Unfortunately, studies like this can create panic/misinterpretation which can have real consequences as shown by this study @AmCollegeGastro
-Unfortunately, studies like this can create panic/misinterpretation which can have real consequences as shown by this study @AmCollegeGastro
13/ The important study to do would be to see if use of PPI in COVID 19 lead to meaningful changes in clinic outcomes:
--Does it lead to higher hospitalization rates or higher mortality with #COVID19 in a prospective cohort?
--Does it lead to higher hospitalization rates or higher mortality with #COVID19 in a prospective cohort?
Thoughts from other esophageal/GI docs? Does this study change your clinical practice?
@AllonKahn @KristleLynchMD @WalterChanMD @RishiNaikMD @MTPapaD @AfrinKamalMD @JPandolfinoMD @ZubairMalik_MD @DJodorkovskyMD @JHorsleySilvaMD @BaldeepPablaMD
#MedEd #GITwitter
@AllonKahn @KristleLynchMD @WalterChanMD @RishiNaikMD @MTPapaD @AfrinKamalMD @JPandolfinoMD @ZubairMalik_MD @DJodorkovskyMD @JHorsleySilvaMD @BaldeepPablaMD
#MedEd #GITwitter
