H Pylori: test and treat

A systematic review and network meta-analysis found that a ”test and treat” approach to H Pylori infections in patients who have dyspepsia is the optimal strategy (see dyspepsis test and treat H Pylori bmj2019 in dropbox, or doi: https://doi.org/10.1136/bmj.l6483)

Details:
-- 15 randomized controlled trials with 6162 adult participants with symptoms of dyspepsia were included
-- the strategies tested were prompt endoscopy; testing for H Pylori and performing endoscopy in those testing positive (“test and scope”); testing for H Pylori and giving eradication treatment in those who test positive (“test and treat”); empirical acid suppression; or symptom-based management
-- trials reported symptom status at follow-up, at least 12 months later
-- the strategies were ranked by the P score, the probability of each treatment being ranked is the best in a network analysis

Results:
-- as compared to symptom-based management:
    -- test and treat: P score 0.79 [the best]
    -- prompt endoscopy: P score 0.71
    -- test and scope: P score 0.57
    -- empirical acid suppression: P score 0.30
--test and treat also ranked first on analysis by per-protocol (ie, what patients they did, instead of which group they were randomized to)
--none of these strategies were significantly less effective than test and treat, or more effective than each other, by either direct comparisons of studies or by indirect comparisons through the network meta-analysis
--endoscopies done, as compard to the prompt endoscopy:
   -- test and treat had the lowest, with 77% fewer endoscopies than the prompt endoscopy group, RR 0.23 (0.17-0.31), P score 0.98
    --symptom-based management had 61% fewer, RR 0.39 (0.20-0.74), P score 0.61
    --empirical acid suppression also had 61% fewer, RR 0.39 (0.27-0.56), P score 0.59
    --test and scope had 45% fewer, RR 0.55 (0.30-0.99), P score 0.32
--relative risk of remaining symptomatic after therapy: all were about the same, at about 0.90, but performed similarly (test and treat eeked out a slight edge over prompt endoscopy, 0.89 vs 0.90)
--participant dissatisfaction (there were only 6 RCTs and and less good data on this):
    -- the lowest risk of dissatisfaction was with prompt endoscopy, with the relative risk of being dissatisfied at 0.58 (0.37-0.91) P Score  0.95 [ie, this was the approach that  patients were happiest with]
    -- next lowest was test and scope, RR 0.83 (0.43-1.59), P Score 0.46
   -- then test and treat, RR 0.85 (0.54-1.34), P Score 0.41
-- upper GI cancer, of 5028 participants 20 cancers were detected (0.40%)
    -- 11 (0.67%) in those undergoing prompt endoscopy, 4 (0.24%) in test and treat, 2 (0.41%) in test and scope, and 2 (0.24%) on empirical acid suppression, and one (0.26%) on symptom-based management

Commentary:
--“dyspepsia” is a diverse set of symptoms for patients who have upper abdominal pain or discomfort, including epigastric pain, early satiety, postprandial distress, heartburn, regurgitation, nausea. And it affects about 20% of people at some point in their lives, with substantial effect on quality of life/work productivity in a large many.
-- the approach in the above study is only for those without alarm features (eg weight loss, anemia, dysphagia), who should all go to prompt endoscopy
-- though many old studies concluded that test and treat was an appropriate strategy, international guidelines have differed: the Am College of Gastroenterol and Canadian Assn Gastro consider it first line in those <60yo, strength of recommendation is strong with high level of evidence; NICE in the UK also considers it first line, but holds back a bit suggesting to offer it to most patients (they actually consider either test and treat, or empirical acid suppression as acceptable); the Asia-Pacific working Party in 1998 said to “consider it” if empirical acid suppression fails (though these are really old recs, and this would be a counterintuitive approach now since there is much higher incidence of gastric cancer in that region)
-- the network meta-analysis, as used above, attempts to evaluate different strategies in different RCTs, even if there was no direct comparison of 2 or more modalities in a single RCT. For example, it might compare 2 strategies vs placebo, then compare the effect size in each one to ascertain which strategy was likely to be better. Though this might be the only way to compare 2 strategies, it is fraught with significant concerns, since the 2 studies likely had very different patient groups with different demographic and clinical features, different inclusion and exclusion criteria, and different overall methodologies. And, in this case where endoscopy was an option, there was no blinding of recipients, so none of them were at “low risk of bias”. In the above analysis, there was actually some direct comparison for 9 of the 10 possible combinations, making their conclusions a bit more robust. But  there was no access to individual patient data, eg to see if there were differential effects within the studies regarding using different dyspepsia-related resources or how the intervention was applied or specific endpoints used
--also the RCTs in this study were largely from Western populations (one study was from Malaysia), so not sure the conclusion is generalizable to many areas in the world (esp  Asia) where the rates of gastric cancer are higher (which is largely due to H pylori infection)
--upper GI cancer rates: not surprisingly low in this study of patients of mostly younger age (the studies were all very different, though most had mean age of 40-45) and followed only 12-18 months
--it is also notable that the actual symptomatic response to treating H pylori was so low; several issues here:
    --most people with dyspepsia have functional dyspepsia and not peptic ulcer disease, and the effect of treating H pylori in the former group is low per many studies (number needed to treat in the 8-14 range, per H Pylori gastritis kyoto report gut2015 in dropbox, or doi:10.1136/gutjnl-2015-309252). But, treating H pylori functional dyspepsia, when it does help, would alleviate the need for long term acid suppressants; and the long-term efficacy and safety of PPIs are not totally clear (see http://gmodestmedblogs.blogspot.com/2017/04/risks-and-benefits-of-longterm-ppis_26.html , or http://gmodestmedblogs.blogspot.com/2017/07/ppis-and-increased-mortality.html , though more recent studies do suggest that they seem to be safer than some of my concerns in these blogs. But still, best to use PPIs only if necessary and for shortest timeframe appropriate)
    --the studies in this network meta-analysis were pretty old, from 1994-2009, and all used H pylori treatments that are not acceptable now because of limited efficacy (lower doses of meds, and only 1 week of therapy)
    --and, there might be other benefits of treating: reduction in future dyspepsia in some patients or later peptic ulcer disease; reductions in gastric cancer (eg see http://gmodestmedblogs.blogspot.com/2019/10/h-pylori-rx-vitamins-and-dec-gastric.html)
-- why was the satisfaction rate so high for prompt endoscopy??
    -- patients may feel better knowing definitively what is going on through direct assessment. Or perhaps they just have expectations that a more invasive test is better???
    --but, of course, endoscopy is quite expensive (one study found that endoscopy was cost effective if the willingness to pay per patient cured of their dyspepsia was $180,000), and it does have some small chance of adverse events
--one concern is the actual long-term role of untreated H pylori infection in countries (eg the US) where the actual incidence of gastric cancer is quite low, and even decreasing. Should we really be concerned here?? the projected incidence of stomach cancer this year is still 27,510 cases and 11,140 deaths (see https://www.cancer.net/cancer-types/stomach-cancer/statistics ). And H pylori is one of the most profound risk factors. But, even with this “low” number of stomach cancers, this is often a pretty miserable cancer and, I think, the bar to detect H Pylori and treat should be pretty low.

So, this study does support the pretty standard approach used nowadays of testing for H pylori in those with symptomatic dyspepsia (and no alarm features) and then just treating it. And, I expect that symptomatic relief is higher now with the more effective treatment regimens we prescribe. A few points:
    --several of the older studies just did antibody testing for H pylori. I think this is still okay, since the constellation of symptomatic dyspepsia in the setting of old H pylori infection (as per IgG positivity) does warrant treatment (and the stool antigen does have some false negatives, esp if the patient has been taking PPIs, which are available OTC)
    --though, as per the blog yesterday (see  http://gmodestmedblogs.blogspot.com/2019/12/h-pylori-in-us-veterans.html ), it does make sense to have a final documentation of H Pylori eradication through the stool antigen of breath test.
    --and, this last blog does reinforce the pretty common occurrence of H Pylori in the US (mostly asymptomatic) and the likely benefit of detection and treatment

geoff​

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@uphams.org

to get access to all of the blogs:

1. go to http://gmodestmedblogs.blogspot.com/ to see them in reverse chronological order

2. click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

3. or you can just click on the magnifying glass on top right, then  type in a name in the search box and get all the blogs with that name in them

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list