You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.

Archive for November, 2013

Primary Care Corner with Dr. Geoffrey Modest: Morning After Pill INEFFECTIVE in Overweight!?

27 Nov, 13 | by EBM

a disturbing story today that morning-after pills may not be as effective in overweight women (see for story). basically, European health regulators found that a French morning after pill marketed as Norlevo had decreasing effectiveness in women over 165 pounds and was not effective at all in those over 176 pounds, prompting the Europeans to order a label change.  turns out that Plan B in the US is basically the same composition. there was a comment in Physician’s First Watch that “According to the CDC, the average American woman weighs 166.2 pounds, which may raise concern about the efficacy of this type of emergency contraception among many U.S. women. The battle to make emergency contraception available has been long and arduous. Hopefully, this recent information will not discourage users, as this may still be the best option available.”

a couple of comments:

1. it is pretty surprising to me that if the average US woman weighs in the “decreasing effectiveness” group, and that there are lots in the “not effective at all” group, that failure rates in the US would have been pretty apparent before this.

2. seems that BMI would be more useful than weight per se, since (i would assume) the issue is the amount of fat present, with its effects on hormone metabolism (ie and the amount of adipose tissue in a 6 foot tall woman weighing 176# is pretty different from a 5 footer of same weight)

2. there was a recent article in ObGyn journal (see  DOI: http://10.1097/AOG.0b013e31828317cc) which assessed efficacy of oral contraceptive pills, patches, and vaginal rings in 7500 women of varying BMIs followed 2-3 years.  no difference in failure rates in those with BMI <25 (8.5%), BMI 25-30 (11%), or BMI >30 (9%). this was an observational study with self-reported outcomes. dose of ethinyl estradiol in the pills ranged from 20-35 micrograms (though no information on whether the higher dose pills were taken more by the more overweight women). other observational studies have had mixed results.

so, quite concerning finding in Europe. i would assume that the Food and Drug Admin will assess this aggressively to see, for example, what the US experience is and if the morning after pill should have differing hormonal content based on BMI (or weight).


Primary Care Corner with Dr. Geoffrey Modest: And more on pharma shenanigans

26 Nov, 13 | by EBM

so, just after i posted the blog yesterday, this came through:  “FDA Opens Door to Wider Avandia Prescribing” – see, or your local newspaper. has the slight whiff of the hand of big pharma…..

i can’t figure out why this is an issue (expect from drug company pressure). the data are pretty clear that rosiglitazone (avandia) in several studies is associated with more clinical heart disease.  it has adverse lipid effects. and, if one were to choose a glitazone, there are data that pioglitazone is beneficial to lipids and leads to favorable cardiac outcomes  (see dm pioglit PROACTIVE study. lancet 2005, or Lancet 2005; 366: 1279–89).

and the beat goes on.



Primary Care Corner with Dr. Geoffrey Modest: More drug co. shenanigans

26 Nov, 13 | by EBM

as part of the safeguard for clinical trials, the FDA modernization act mandated establishment of the Registry of Clinical Trials, with public access to information.  in 2007 the USA FDA Amendments Act expanded it to include prior studies, posting almost all trials at (by law: posting all prospective clinical study of health outcomes, studies involving drugs or other products regulated by the FDA, and controlled intervention studies in humans). Failure to comply with this mandate could result in penalties of up to $10,000 per day and withholding of funds from investigators sponsored by the NIH.

In a recent studies reported in BMJ (see doi: 10.1136/bmj.f6104) investigators looked at all trials with at least 500 participants in the clinical trials registry and completed prior to January 2009 and published by November 1, 2012. (These researchers focused on large studies because these would most likely be published if presented to a medical journal and therefore if not published, more likely represented a conscious decision not to pursue publication).  they found 585 registered trials, of which 29% (171 studies) remained unpublished.  300,000 people were enrolled in these 171 studies. Of these 171 trials that remained unpublished, 38 did have results available on clinical But 133 had no results available either in published form or in (one could argue that even if these results were available at that website, there was an obligation to publish the data from these large trials in a peer-reviewed journal). They also found that non-publication was more common among trials that received industry funding (32%) than those that did not (18%).

The indications of the above are quite profound.  As I mentioned in prior posts, there has been a dramatic shift in this country (the US)  in funding of research.  Ronald Reagan as president effectively privatized much of healthcare research.  Prior to Reagan, articles in medical journals were predominantly sponsored by the government (NIH, NSF).  However,over the past several decades the vast majority of research has been sponsored by the private sector (almost all studies in the best peer-reviewed medical journals are drug or medical device company sponsored), with the inherent conflict of interest in presenting negative results for drugs/medical devices they want to promote. There has been public exposure of some of the flagrant examples, such as publishing only a smaller positive study on gabapentin for neuropathic pain and suppressing a much larger study which found a negative results. This article reveals that the safeguards that the FDA had placed to protect the public have been frequently ignored.

I will also cut and paste a relevant recent article in NY Times  (“Doctors Say Heart Drug Raised Risk of an Attack” by Andrew Pollack, 11/19/2013):


Cardiologists have accused a small drug company of withholding data from a clinical trial showing that the company’s drug, meant to reduce the risk of heart attacks, increased the risk instead.

The cardiologists said that the company, Anthera Pharmaceuticals, did not turn over data to academic investigators, as it was required to do, for more than a year.

“Despite a contract that required transfer to the academic authors, the company stonewalled every attempt to acquire the data,” Dr. Steven Nissen, a cardiologist at the Cleveland Clinic, said in an email on Tuesday.

Studies and lawsuits have shown that many clinical trial results, particularly negative ones, are not published. Critics say that hampers medical practice and violates an obligation to patients, who try experimental treatments in part to advance knowledge.

“We think that when you enter into a clinical trial, and we enter into contracts with our patients, there’s an obligation that we are going to do the right thing,” said Dr. Nicholls, a cardiologist at the South Australian Health and Medical Research Institute in Adelaide.

The Phase 3 trial involved 5,145 patients in various countries with acute coronary syndrome, a sudden blockage of blood flow to heart muscles. It tested whether the drug varespladib could reduce the risk of heart attacks, stroke, death and chest pains requiring hospitalization.

The trial was terminated early by the safety monitoring committee in March 2012. The company announced that the committee had determined that there was no chance the drug would succeed in the trial.

The results show that patients who got the drug actually had a higher risk of cardiovascular problems, mainly heart attacks, than those who received a placebo.

Dr. Nicholls said in an interview that Anthera continually promised to provide the complete data but did not. Finally, the company told him this spring that it had given the drug back to Eli Lilly, the original developer of the drug. The investigators were then able to get the data almost immediately from Lilly, he said.

Dr. Nicholls said that Anthera was also supposed to ensure that all patients were surveyed six months after the trial ended to see if they were still alive. But that data was collected for only 31 percent of the patients, making it impossible to clearly determine if the drug had increased the risk of death, he said.

so, this all strongly suggests that we as clinicians be very careful about being early-adopters to new medications or medical devices, since the studies are largely supported by drug/medical device companies, and we cannot be certain about the accuracy or completeness of the data we see. and to make matters worse, the FDA (under pressure from drug companies) has been releasing drugs earlier than before with the expectation of post-marketing surveillance….


Primary care corner with Dr. Geoffrey Modest: Nocebo

25 Nov, 13 | by EBM

in the new york times there was a good analysis of the “nocebo effect”–people on placebo in various studies who have significant side-effects. can be up to the 40% range.  an example of a patient who attempted suicide by taking an overdose of meds, which, unbeknownst to him, was placebo — and his blood pressure dropped!! other studies suggest that when provider describes potential side-effects of a new med, there is a greater likelihood of the patient developing that side-effect.  see:

in my practice, i mention common side-effects but more in a dismissive way (like, “although some people may develop some inflammation of the liver when on a statin, it is really unlikely, i have seen it rarely in the many patients i have on statins, and i will monitor you closely just in case”), especially if i think the med in question is really important for the patient. however, i have had several patients who refuse to take the med after reading the attached pharmacy paper detailing the side-effects. so, my sense is that we do need to mention important side-effects to new med, we need to make sure that we put this in perspective and that the patient understands our rationale for prescribing the med and then reassure the pt that we will be accessible and will follow them closely to make sure all is ok (also does reinforce utility of seeing the patient sooner than later after starting a new and important med, to make sure there is no problem).


Primary care corner with Dr. Geoffrey Modest: Dual renin angiotensin system blockade?

25 Nov, 13 | by EBM

The question arises occasionally as to whether there is an advantage of dual blockade of the renin angiotensin system, especially in cases of patients with heart failure or severe proteinuria.  The rationale of dual blockade is that ACE inhibitors by themselves lead to incomplete and often transient RAS blockade in most patients because of a physiologic escape mechanism/alternative pathway, and the objectives of RAS blockade are to decrease cardiac remodeling, reduce endothelial dysfunction, and decrease renal dysfunction by decreasing the putative nephrotoxic effects of severe proteinuria.  Dual blockade has been mostly achieved through the combination of an ACE inhibitor and an angiotensin receptor blocker (ARB), or the combination of an ARB/ACE-I with the direct renin inhibitor aliskiren.  There was meta-analysis in the BMJ recently of 33 randomized controlled trials with 68K patients with a mean age of 61 and followed a mean of 52 weeks  (see BMJ 2013;346:f360 doi: 10.1136/bmj.f360), with the major indications for the dual blockade being congestive heart failure and hypertension.  Results:

–No overall benefit for all-cause mortality, although there was a trend to a decrease in the subgroup with heart failure but a significant 7% increase  in the subgroup without heart failure (i.e., mostly in those treated for hypertension)

–No benefit in cardiovascular mortality

–In the subgroup with heart failure, there was an 18% decrease in hospital admissions

–However, dual therapy was associated with the 55% increase in the risk of hyperkalemia (defined as a potassium greater than 5.5), a 41% increase in the risk of renal failure (defined as a creatinine above 2 or a doubling of the creatinine), and a 66% increased risk of hypotension.

so, the bottom line is that dual RAS blockade seems only to help in decreasing hospitalizations in patients who have heart failure.  In terms of severe proteinuria, the COOPERATE trial did find that dual therapy led to a decrease in renal failure, spurring many of us on to advocate aggressive dual therapy in this circumstance.  However, this study was ultimately retracted and the ONTARGET study found that the combination of the ARB telmsartan and ACE-I ramipril decreased albuminuria, but led to more renal dysfunction.  So, yet again, what seemed mechanistically to be an attractive therapy (dual RAS blockade), ultimately proved not to be clinically beneficial, except perhaps in decreasing hospitalizations in patients with severe heart failure.  Although dual blockade does decreased blood pressure, I am unaware of any study showing positive clinical outcomes with this therapy and the increase in all-cause mortality in patients in this meta-analysis with dual therapy not for a heart failure indication (which is mostly in patients with hypertension) was associated with an increase in all-cause mortality.



Primary care corner with Dr Geoffrey Modest: Non-celiac gluten sensitivity?

25 Nov, 13 | by EBM

It is not uncommon to see patients who have typical symptoms of celiac disease who seem to respond to a gluten-free diet but have negative workup for celiac disease. These patients are said to have nonceliac gluten sensitivity (NCGS), which is characterized by irritable bowel-type symptoms after the ingestion of gluten, improvement after gluten withdrawal from the diet, and negative celiac serologies/biopsies. they can have both intestinal sx (diarrhea, abd discomfort/pain, bloating, flatulence) and extra-abdominal ones (headache, lethargy, poor concentration, ataxia, oral ulceration). Additionally (and previously unknown to me), there are some food items which can also cause these symptoms, called FODMAPs (fermentable, oligo-, di-, monosaccharides and polyols) which are poorly absorbed short-chain carbohydrtates. Presumably, FODMAPs lead to luminal gut distension via osmotic effects and gas production from bacterial fermentation in the gut. A small Australian study was done with a double-blind cross-over methodology of 37 patients with NCGS meeting the criteria for IBS (subjects aged 21-64, 6 men, fulfilling Rome III criteria), continued on their gluten-free diet but also given a 2-week diet of reduced FODMAPs, then randomly placed on diets with high-gluten (16g/d), low-gluten (2 g/d) or control (16 g whey protein isolate/d), with a washout period of 2 weeks, then crossed over to the other diets (see  results:


–several of the patients who stated that they were gluten-responsive did have mild to moderate symptoms on a gluten-free diet, with at least 50% of the patients showing significant improvement on the low FODMAP diet during the run-in period of the study, with decrease in both GI symptoms and fatigue (see their figure 1)

–subsequent addition of the high gluten, low-gluten, or control diet was associated equally with worsening of their symptoms. only 8% had gluten-specific effects. 30% had a significant symptom response to the placebo diet (nocebo effect: patients who have symptoms on placebo meds — i will send around again an interesting article on nocebo effect)

–22 subjects returned  8-17 months later for a 3-day rechallenge trial, already on gluten-free diet, then given FODMAPs reduced diet and additionally elimination of other putative symptom triggers in some patients (salicylates, amines, MSG, and preservatives benzoates, priopionate, sultfites, nitrites, and sorbic acid, as well as added antioxidants and food colors). finding — no gluten-specific differences in response to the addition of gluten vs placebo to the meals. of note, in both of these studies there was a consistent difference by the order of the trial: the first intervention induced more symptoms independent of what the intervention was…


so…. what are these things they call FODMAPs???  there is a good website with a one-page handout for patients (and for us) at basically,

–fructose (honey, fruit, high fructose cornsyrup (HFCS)


–fructans (wheat, onion, garlic)

–galactans (beans, lentils, legumes such as soy)

–polyols (sweeteners containing sorbitol, mannitol, xylitol; fruits such as avocado, apricots, cherries, nectarines, peaches, plums)


therefore, you may ask, what can one eat???

–meat, poultry, eggs, fish

–dairy — lactose-free, or small amounts of cream cheese, 1/2 and 1/2, hard cheeses, sherbet

–grains — gluten-free grains, (ie, no wheat)

–fruits — bananas, berries, cantaloupe, grapes, grapefruit, melons, kiwi, lemon, lime

–veges — bamboo shoots, bell peppers, cukes, carrots, celery, corn, eggplant, lettuce, potatoes, squash, tomatoes, zukes

–beverages — none with HFCS, but other fruit/vege juices (1/2 cup at a time), coffee, tea

–seasonings — most spices and herbs. not garlic, honey, onions, molasses, jams, jellies, coconut


bottom line: from my experience, i think it is pretty common for patients to have abdominal bloating, pain, etc. my first approach is to give fiber, make sure bowel movements are regular and soft. but symptoms may well persist. sometimes low gluten diet works in spite of negative workup for celiac dz. and, if so, that’s an easier solution than the above FODMAPs diet. but i will now try using this diet if the patient continues to have symptoms in spite of regular bowel movements and nonresponse to low gluten diet.




Primary care corner with Dr Geoffrey Modest: Prostate screening conundrum continues

25 Nov, 13 | by EBM

There have been a couple of articles of note on prostate cancer and screening.


One was in this week’s New England Journal of Medicine (see DOI: 10.1056/NEJMsa1201141). The impetus for the study was to determine if urologists purchasing their own office-based, expensive intensity-modulated radiation therapy (IMRT) equipment were more likely to use this form of radiation therapy vs urologists who do not own the equipment.  The researcher looked at Medicare claims from 2005 through 2010 in 2 samples: 1 was in private practice groups with 35 urologists self-referring for IMRT who started using this equipment during this time period versus 35 groups not owning this therapy; the other study was of a group of 11 non-self-referring urologists at National Comprehensive Cancer Network centers compared to 11 self-referring urology groups in private practice.  The intention of the second study was to compare urologists who were likely to practice on the basis of clinical evidence and unlikely to derive personal financial benefits from the services, versus 11 matched self-referring private urology practices in close proximity to the centers.  Findings:

–for the 35 private practice urology groups, the rate of IMRT use by self-referring urologists increased from 13.1 to 32.3%.

–for 35 non-self-referring urology groups during the same time period, the rate of a IMRT use increased only from 14.3 at 15.6%.

–For those urologists working at National Comprehensive Cancer Network centers, the use of IMRT remained stable at 8% but increased to 33% in the 11 matched nearby self-referring urology groups.

–The mean cost estimate of her IMRTwas $31,574, approximately twice that of a radical prostatectomy or brachytherap

Not surprising.  Similar findings have been found for using advanced imaging techniques, clinical laboratory testing, and pathology services by self-referring physicians.  There is also an increased use of surgery in physician-owners of specialty hospitals. Although the feds prohibit much self-referral, there are notable exceptions, including for radiation therapy.


The second article which came out several months ago was an update of the American Urologic Association guidelines for early detection of prostate cancer (see: There were several changes in these guidelines over prior ones.  They still focus on PSA screening, given that there is little evidence supporting the clinical efficacy of DRE, PSA derivatives (eg free PSA, PSA velocity, etc), or novel biomarkers such as PCA3.  They do acknowledge several harms from screening, including that more than 75% of people with a PSA greater than 3 have no cancer on biopsy, the cumulative risk of having at least one false positive test after 4 rounds of annual testing was around 13%, and over-diagnosis of prostate cancer was on the order of 66% (approximately 1/3 of the screen-detected cases).  Their general recommendations are as follows:

–Not screen men under age 40

–Not do routine screening in men between 40 and 54 years of age at average risk. They recommend individualized decisions on prostate cancer screening in these younger men at higher risk (e.g. positive family history, or African-American men)

–For men aged 55-69, “the panel recognizes that the decision to undergo PSA screening involves weighing the benefits of preventing prostate cancer mortality in one man for every one thousand men screened over a decade against the known potential harms associated with screening and treatment.  For this reason, the panel strongly recommends shared decision-making for men aged 55-69 years that are considering PSA screening, and proceeding based on a man’s values and preferences.”

–The screening interval of 2 years or more may be preferred over annual screening to reduce the harms of screening, preserving the majority of benefits and reducing over-diagnosis and false positives.

–Not screening men over 70 or men with a less than a 10-15 year life expectancy.

so, these new guidelines definitely reflect a softening of their position on screening. still clearly supports screening with the caveat that there should be shared decision-making.  pretty hard to do that when we, as trained medical providers able to decipher the intricacies of the different studies and their methodologies, are pretty clueless as to what to do.  I will also follow this post with a prior discussion–see below.


From March 2013:

in one of my precepting sessions, the question came up about the utility of finasteride as well as the predictive accuracy of PSA screening.  finasteride does do well in decreasing prostate volume, though it takes months to work. but i wanted to circulate 2 important articles (both from the prostate cancer prevention trial), which i think sheds light on these issues.

1. prostate cancer prevention trial — 7 year study of 19K men >55yo, with PSA <3 and nl DRE, randomized to finasteride 5mg/d vs placebo. rationale of study is high prevalence of prostate cancer in men (17% lifetime risk). found that prostate cancer was reduced by the prophylactic administration of finasteride by 25% (24% in the placebo group, and 18% with finasteride). BUT, higher incidence of high-grade prostate cancer (Gleason scores 7-10) in the finasteride group (6.4%) vs the placebo group (5.1%).  also sexual side-effects of finasteride. but, bottom line, concern that there may be more high-grade, potentially lethal prostate cancers with finasteride.

2. prostate cancer in those with low PSAs — this looked at the placebo group at the end of the above 7 year study, where they did a prostate bx in 3K men in the placebo group who never had PSA>4 or abnl DRE. rather striking findings as follows:

-15.2% had prostate cancer, breakdown as follows:

-psa <0.5, 6.6% with prostate cancer (!!!)

-psa 0.6-1, 10%

-psa 1.1-2, 17%

-psa 2.1-3, 24%

-psa 3.1-4, 27%

-of those who had prostate cancer, 15% of them had high grade cancers (Gleason 7-10), breakdown as follows:

-psa <0.5, 12.5% of those in that psa group with cancer had high grade lesions

-psa 3.1-4, 25% with high grade lesions.

bottom line, increasing psa is associated with more cancers and high-grade cancers, BUT psa is a pretty miserable test, does very poorly in terms of its sensitivity (and, if you lower the sensitivity to a cutpoint of 3 or 2, the number of biopsies done in the population increases dramatically and you still seem to miss a fair number of even high grade cancers).

hope this is helpful.


Primary care corner with Dr Geoffrey Modest: Adverse statin effects

25 Nov, 13 | by EBM

There have been a slew of articles in the literature in the past year on the harmful effects of statins. A recent study-level network meta-analysis was reported, incorporating 135 randomized controlled trials with 250K participants to assess the reported adverse effects (note that some of these effects could have been under-reported, such as myalgias, given the relative frequency of such in many of our clinical practices but their low numbers) — see  DOI: 10.1161/CIRCOUTCOMES.111.000071.

The findings:

–no difference between individual statins and controls for: myalgias (subjective muscle pain), creatine kinase elevations (varied in studies from 3x to 10x more than baseline), cancer, or discontinuations because of adverse events. Likelihood of rhabdomyolysis small and no diff from controls.

–overall 9% higher likelihood of developing diabetes

–51% higher likelihood of developing transaminitis (ALT or AST >3x baseline, though note that this is not a clinical marker only a biochemical one)

–simvastatin and pravastatin best tolerated, higher doses of atorvastatin and rosuvastatin less tolerated

–higher dose simvastatin (>40mg/d) assoc with more marked increase in CK elevations (OR 4.14)

Some specifics:

–myalgias: in the larger analysis of all studies, found in 2% of participants, no diff between statins

–transaminitis: 1% of participants. Atorva the worst in dose-response fashion, with OR 5.25 if >40mg/d.  fluvastatin pretty bad, but no one uses this anymore….

–CK elevations: 0.6% of individuals. Worst with high-dose simva (though interesting that myalgias in this group were not significantly different from controls and actually had a trend to be 28% less than controls…)

–diabetes: overall 9% increased risk, though worse for high-dose rosuvastatin (16%) in the placebo-controlled direct trials, but no individual statin difference in the overall network analysis.

–simva and prava were best in the combined outcomes (includes discontinuations from combo of adverse effects, myalgia, transaminitis and CK elevation)

–pitavastatin recently approved by FDA (have never used it myself), but in general seems to be worse than the others overall.

The background here is that statins are dramatically effective in reducing clinical cardiac/cerebrovascular events, in the elderly, in women, in patients with diabetes, with relative risk reductions in the 25-30% range (even in patients with diabetes felt by their endocrinologists not to need statins  because their cholesterol levels were “okay”). Of course the absolute risk reduction depends on the risk status of the cohort involved,but recent meta-analyses have shown pretty conclusively that statins are beneficial even in primary prevention and even if some cases of diabetes develop) — see Lancet 2012; 380: 565–71 for Jupiter study with rosuvastatin, the worst one in the above meta-analysis, and argument that cardiovasc benefits exceed diabetes risk. Overall, there do seem to be some differences between the statins, with pravastatin and simvastatin having the fewest adverse effects (I was a bit surprised here, since some earlier individual studies found worse problems with simvastatin vs atorvastatin,  which reinforces the utility of these larger analyses).


Primary Care Corner with Dr. Geoffrey Modest: Ca++ based phos binders increase mortality

25 Nov, 13 | by EBM

for those of us either treating patients with chronic kidney disease or following them for other reasons, there was an important article in lancet which found that calcium-based phosphate binders are associated with increased mortality (see  this was a meta-anal of 5 new RCTs and 9 older ones  from a previous meta-anal, with 11 reporting mortality as an outcome. 4622 pts involved. compared those on calcium-based phosphate binders (calcium carbonate or calcium acetate) with non-calcium based binders (sevelamer hydrocholride or carbonate — brand name eg of renagel — or lanthanum carbonate — brand name fosrenol). results:

–overall 22% reduction in all-cause mortality with non-calcium based phosphate binders.  most studies in dialysis patients, but the few with predialysis (only 134 patients) with almost significant 46% decrease in all-cause mortality (0.28-1.03). decreased mortality independent of degree of phosphate reduction

–for the 7 RCTs which assessed coronary artery calcification, there was a very highly significant decrease in those on non-calcium based formulations vs calcium-based (looked at the Agatston score, for what that’s worth)

turns out, not to shock you, that the non-calcium based phosphate binders are a lot more expensive (4-70x), but the patents are due to expire next year.  the background here is that renal failure by itself is highly associated with coronary artery disease (studies suggest that even slightly elevated creatinine levels or even high normal microalbumin levels are associated with higher clinical CAD, and the combo of both has even more association), and the vast majority of renal failure patients die of CAD. so, makes sense to me to switch patients to the non-calcium based binders….


Primary Care Corner with Dr. Geoffrey Modest: Sequential or standard Rx for H pylori?

25 Nov, 13 | by EBM

previously sent out an article finding that sequential H pylori therapy (5 days of PPI plus amoxicillin 1 g twice a day, followed by 5 days of PPI plus clarithromycin 500 twice a day plus metronidazole 500 twice a day) was superior to standard triple therapy.  This sequential therapy worked even with clarithromycin resistance, with 89% of clarithromycin-resistant strains responding vs 29% with standard therapy. they posited the mechanism for resistance being that h pylori developed efflux channels for clarithro that rapidly remove the drug from the bacteria and that amox changes the structure of the cell membrane to prevent effective clarithromycin efflux and therefore resistance. (see Ann Intern Med 2007; 146:555-563)


there was a recent systematic review and meta-analysis in the BMJ assessing the efficacy of sequential versus triple therapy (see  BMJ 2013;347:f4587 doi: 10.1136/bmj.f4587). for this review, 46 randomized control trials were assessed with 5666 patients on sequential therapy versus 7866 on standard therapies.  Results:

–22 studies compared sequential therapy with a seven-day course of standard triple therapy (mostly PPI plus amox plus clarithro, all bid; or PPI plus amox plus metronidazole, all bid).  While in the 5000 patients randomized to each of these 2 groups, sequential therapy resulted in H. pylori eradication in 86.5% versus 71.5% for triple therapy, highly statistically significant

–14 studies compared sequential therapy with triple therapy for 10 days.  2700 patients were divided into the 2 groups with eradication rates of 84.3% for sequential therapy and 75.3% for triple therapy of 10 days duration, a statistically significant result.  No significant differences in adverse effects.

–7 studies compared sequential therapy will triple therapy lasting 14 days. 2500 patients were divided into the 2 groups with 81% eradication rates in each group.  No different in adverse effects

–There was also no difference between sequential therapy and therapy that contained bismuth or with quadruple therapy.

–in pts with clarithro resistance, efficacy of sequential therapy was 73% vs 50% in those with 7 or 10 days of standard therapy,  but no diff with 14 days of standard therapy , or with quadruple therapy or sequential therapy with levoflox instead of clarithro

–in pts with metronidazole resistance, eradication rates for sequential therapy were 86%, vs 62% with 7 day and 59% for 10 day standard therapy, both significant, though the only one was a small study (50/group) of 14day therapy found eradication rate of an unusually low 73% with sequential therapy and 89% with 14-day standard therapy.

–in pts with both metronid and clarithro resistance, studies too small to have generalizable conclusions

so, h pylori infection is really common, at least in boston, and global studies suggest that gastric cancer assoc with h pylori (non-cardia gastric carcinoma and MALT) in 2008 occurred in 470K people in less developed regions and 190K in more developed regions.  data of the standard therapy (per Maastriccht conference in 1997) with PPI, clarithro, and either amox or metronidazole suggest waning efficacy over time.  In Europe, resistance to clarithromycin is 17.5%, to levofloxacin is 14%, and metronidazole is 35%.  it would be very useful to use local resistance patterns as a way of choosing empiric therapy. in 2007 when the first sequential therapy article came out (above), i called several endoscopists at boston medical center and no one had data on local resistance patterns.  i called again today and this is still true. Their guess is that resistance is very common here (perhaps in the 30-50% range!!!), though they still mostly prescribe the 10 day course of standard therapy, without great results. one big issue here is that NONE of the studies have been done in the US, again clouding the picture of what we should do. I personally have been prescribing the sequential therapy since that annals article in 2007, since I suspected that antibiotic resistance was high given the commonplace use of metronidazole and clarithromycin locally for other infections. the only problem is that it is a more complex regimen and takes a while to explain the therapy, for which I made a written handout, which elineates day-to-day therapy. from the new bmj article, perhaps standard therapy for 14 days is equivalent (again, this may not be true if our resistance patterns are really different…. who knows??)



EBM blog homepage

Evidence-Based Medicine blog

Analysis and discussion of developments in Evidence-Based Medicine Visit site

Creative Comms logo

Latest from Evidence-Based Medicine

Latest from EBM