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.

Primary Care Corner with Geoffrey Modest MD: PPIs and increased mortality

20 Jul, 17 | by gmodest

A large longitudinal study of US veterans found a 25% increased risk of death associated with proton pump inhibitor (PPI) usage (see 10.1136/ bmjopen-2016-015735)​.


— the researchers assessed three cohorts of patients, with 5.7 years of follow-up after the first acid suppression therapy prescription was written, and all patients had at least one outpatient serum creatinine value before acid suppression therapy was chosen:

— primary cohort: new users of PPI or H2 blockers (n=349,312)

— PPI versus no PPI users (n= 3,288,092)

— PPI versus no PPI and no H2 blocker (n= 2,887,030)

— covariates assessed included age, race, gender, renal function, number of outpatient serum creatinine measurements, number of hospitalizations, diabetes, hypertension, cardiovascular disease, peripheral arterial disease, cerebrovascular disease, chronic lung disease, cancer, hepatitis C, HIV, dementia, and diseases associated with acid suppression therapy for such GI conditions as GERD, upper GI bleeding, ulcer disease, H. pylori infection, Barrett’s esophagus, achalasia, stricture, esophageal adenocarcinoma.

— Analyzing the difference between the baseline demographics of these different groups: those on PPIs were older (eight months), and had more diabetes, hypertension, cardiovascular disease, and hyperlipidemia than the overall cohort, however looking at their numbers, these differences were on the order of 1% or less. The differences between those put on H2 blockers and PPIs was somewhat more significant, more typically on the order 5%.



— overall PPI use was associated with a 25% increased risk of death vs H2 blockers, HR 1.25 (1.23-1.28)

— using high dimensional propensity scoring (see below) there was a 16% increased risk, HR 1.16 (1.13-1.18)

— comparing PPI use to no PPI use, a 15% increased mortality risk, HR 1.15 (1.14-1.15)

— comparing PPI used to neither PPI nor H2 blockers, a 23% increased risk, HR 1.23 (1.22-1.24)

— in patients without underlying gastrointestinal conditions (those conditions being the ones noted above), those on PPI had 24% increased risk of death vs H2 blocker, HR 1.24 (1.21 1.27)

— among new PPI users, there was a graded association between the duration of exposure and the risk of death (all statistically significant, as compared to those who have taken PPIs <30 days): 5% increased mortality risk for those on PPIs 31-90 days, 17% for 91-180 days, 31% for 181-360 days, 51% for 361-720 days, all of these controlled for the array of comorbidities and risk factors above.



— the above analysis included both propensity scoring, a means to statistically equalize the PPI versus non-PPI users for all of the covariates above, as well as high dimensional propensity scoring, which controlled for an additional potential 500 variables to further minimize confounding

— a recent blog reviewed the risks and benefits of PPIs from the perspective of the American Gastroenterological Association, assessing the literature on the PPI association with kidney disease, dementia, bone fractures, small intestinal bacterial overgrowth, non-typhoidal Salmonella, Campylobacter, spontaneous bacterial peritonitis, C. difficile, pneumonia, calcium/iron/magnesium deficiency, B12 deficiency, and GI malignancies.

— There is also emerging in vitro evidence that PPIs can result in inhibition of lysosomal acidification and impairment of proteostasis, potentially increasing oxidative stress, endothelial dysfunction, telomere shortening, and accelerated human endothelial senescence. In addition animal studies suggest that PPIs decrease the regenerative capacity of livers following partial hepatectomy.

— This VA study, as well as these others reviewed by the AGA, is an observational one. For observational studies, the increased risk found is on the lower side to be significant. The impressive aspects of the study include the fact that it is huge, that they have controlled for a lot of variables (though they do not have specific data on obesity, smoking, and the use of specific medications including anticoagulants, antiplatelet agents, and nonsteroidal anti-inflammatories); that the risk was independent of prior GI symptoms, and that there was a clear dose-response curve (and the 51% increase in those on PPIs for >1 year is getting more impressive). In addition, looking at survival curves for those on PPIs versus H2 blockers, the difference increases linearly over time (ie the curves are splaying apart)


— As mentioned in previous blogs, PPIs are significantly overprescribed, in part exacerbated by their availability over the counter, with studies suggesting that 53-69% are used for inappropriate indications, that the risks of PPIs are poorly understood by patients as well as perhaps some clinicians, that many patients even with severe GI symptoms respond well to H2 blockers or antacids, and that it is often difficult for an array reasons for clinicians and patients to step down therapy to H2 blockers or antacids (e.g., clinicians having relatively limited time with patients are more likely to discuss other pressing medical or psychosocial problems as opposed to spending time trying to convince a patient to step down their therapy; and patients may well be so satisfied with the effectiveness of the PPI, perhaps prescribed by a specialist which might sway them more towards taking the PPI, that they are reluctant to give H2 blockers or antacids a fair trial).

–In my experience with many patients on PPIs, often they self-titrate to use them intermittently, but in any event are usually quite willing to try H2 blockers or antacids when I express my concern about the potential long-term problems by continuing to take such a potent drug when usually a less potent and possibly less harmful one often works as well.


Primary Care Corner with Geoffrey Modest MD: Decreasing opiate prescriptions

19 Jul, 17 | by gmodest

by Dr Geoffrey Modest


Some (likely) good news about prescription opioids from the CDC (see ).


— Opioid prescribing in the United States between 2006 and 2015 peaked in 2010 at 782 morphine milligrams equivalents (MME) per capita then decreased annually to 640 MME per capita in 2015

— there was pretty striking variation by US county, from 203 MME in the lowest quartile to 1319 MME in the highest; multivariate adjustment for many risk factors for opiate prescribing (more non-Hispanic white population, higher rates of uninsured and Medicaid, lower educational level, higher unemployment, living outside the urban areas, more dentists and physicians per capita, more diabetes/arthritis/disability, higher suicide rates) only accounted for 32% of the variation in prescribing [ie, there was much more to it than the expected risk factors… ?perhaps issues like the specific medical cultures in the different areas? other major differences in accident rates, which might reflect very different occupational exposures in different areas??….]

— Prescription rates also increased from 72.4 to 81.2 prescriptions per 100 persons between 2006 and 2010, then declined to 70.6 per 100 persons from 2012 to 2015, a 13.1% decrease.

— High-dose opioid prescribing (daily dose of >90 MME) has decreased much more dramatically: 11.4 per 100 persons in 2010, then 6.7 per 100 persons in 2015.

— The average days supplied of opiates increased 33.0% from 13.3 in 2006 to 17.7 in 2015 [presumably from decreased prescriptions for shorter term opiate prescriptions].


— the situation is still dire: opioid-related deaths from overdoses continues to increase, with 33,091 deaths in 2015, approximately half involving prescription opioids, though a good part of this problem is from heroin laced with fentanyl or other very high potency derivatives.

— Approximately 2 million individuals in the United States have opioid use disorder associated with prescription opioids, with an estimated economic cost of $78.5 billion

— though the opioid prescription rate has decreased since 2012, it still is three times higher than in 1999 (180 MME per capita) and is four times higher than in Europe in 2015.

— It is likely that the decreased prescriptions for opiates of late is related to increased awareness among clinicians, policies implemented by states to decrease opioid prescriptions, use of prescription monitoring programs, and other local efforts. An initial concern, which has not borne out, was that decreasing prescription opioids would lead to more use of more dangerous illicit drugs. However it does seem that these more restrictive policies have decreased the amount of opioids prescribed, prescription opioid involved overdose deaths, and also all opioid involved deaths (see Dowell, D. Health Affair (Millwood). 2016, 35(10):1876).

— but, unfortunately, such data do not take into account the other half of the picture: what is the effect of lower prescription rates on patient outcomes?? are more patients functionally impaired or incapacitated by lower doses of opiates? are fewer patients able to work or lead productive lives? what are the effects on their families? their communities?



— as most of us in clinical practice know, is usually very difficult to convince patients already on chronic prescription opioids to stop taking them entirely, let alone decrease the dose significantly; patients often state that they need these meds to function. So, in many ways, the biggest long-term solution really is to dramatically decrease the number of new patients who are put on opiates. This involves a concerted effort by surgeons to avoid automatically prescribing opiates for surgery/pain (some patients do fine on non-opiates, with studies suggesting for example equal efficacy of NSAIDs for renal colic), emergency rooms to significantly decrease opiate prescriptions (which often make it very hard for us in primary care to stop prescribing them), as well as us in primary care avoiding using opiates whenever possible. There are many non-opiate approaches, medical and psychological (e.g. cognitive behavioral therapy, yoga, etc.) which may do well (see below). And, at least my experience in Boston is that many clinicians are prescribing opiates much less readily than previously.

— As mentioned in prior blogs (see below), once giving opiates to both young people and elderly seems to lead to more likelihood of continued opiate use/prescription subsequently.

— So, to me it is quite impressive that over the last five years there is been such a significant decrease in overall opiate prescriptions, but especially in those on higher doses (which may be associated with increased mortality), the apparent number of short-term prescriptions, as well as the increased documentation of nonopiate approaches which seem to help.​ But i am still concerned that at least some patients are now being undertreated for their pain….



blog of study which found benefit for cognitive behavioral therapy for low back pain

blog of study finding pain control for knee pain through an internet-based home program

blog of a study showing that in older patients, those prescribed opiates for similar conditions were more likely to continue on opiates later

blog of a study finding that teens at low risk of illicit drug dependence who were given legitimate prescription opiates as 12th graders were 3-fold more likely to have opiate use disorder at age 23

Primary Care Corner with Geoffrey Modest MD: Dyspepsia guidelines

18 Jul, 17 | by gmodest

by Dr Geoffrey Modest

​​The American College Of Gastroenterology and the Canadian Association of Gastroenterology updated their guidelines on the management of dyspepsia (see doi: 10.1038/ajg.2017.154​)


Recommendations (note: the recommendation is much weaker when they use the word “suggest” vs “recommend”):

— they suggest endoscopy for dyspepsia patients greater than 60 years old: conditional recommendation/very low-quality evidence. they raised the age from 55 of prior guidelines since the age-specific incidence of gastric cancer has fallen in the US, and they feel the cost of endoscopy per case of cancer detected is prohibitive. they also do suggest this guideline be individualized, so that for patients coming from areas with high upper GI malignancy rates there should be a lower threshold for endoscopy, especially those coming from Southeast Asia and some countries in South America.

— they suggest not to do endoscopy to investigate alarm features (e.g. weight loss, anemia, dysphagia, persistent vomiting) for dyspepsia patients under the age of 60 . This recommendation is based on seven studies finding that alarm features had limited value for detecting any organic pathology: conditional recommendation/moderate quality evidence

— they recommend that dyspepsia patients under the age 60 should have a noninvasive test for H. pylori, with therapy if positive: Strong recommendation/high quality evidence

— they recommend that dyspepsia patients under the age of 60 should have empiric PPI therapy if they are H. pylori negative or remain symptomatic after H. pylori eradication therapy: Strong recommendation/high quality evidence

— they suggest dyspepsia patients under the age of 60 not responding to PPI or H. pylori eradication therapy should be offered prokinetic therapy: Conditional recommendation/very low- quality of evidence.  Metoclopromide should be given for less than 12 weeks [a problem given the chronicity of dyspepsia, but this drug does have significant adverse effects such as tardive dyskinesia], and domperidone dose should be 30 mg a day or less (this medication is not available in the US).

— the recommendations for patients with functional dyspepsia mirror the above, except that those patients who fail PPI or H. pylori eradication therapy should be offered tricyclic antidepressant therapy (Conditional recommendation, moderate quality evidence), and that those not responding to any of these be offered prokinetic therapy (conditional recommendation, very low-quality evidence). Those with functional dyspepsia not responding to any drug therapy should be offered psychological therapy (conditional recommendation, very low-quality of evidence.) They do not recommend the use of complementary or alternative medicines or the routine use of motility studies, except when gastroparesis is strongly suspected (several studies have found that the relationship between dyspeptic symptoms and gastric emptying is poor). The basis of their recommendation for tricyclics is that there were 3 studies looking at patients with functional dyspepsia finding a significant effect in reducing dyspepsia symptoms. No effect has been seen with SSRIs.



— as we all know, dyspepsia is quite prevalent, approximately 20% of the population globally. More common in women, smokers, and those on NSAIDs. The cost to the US healthcare services is over for $18 billion a year, societal costs are likely double that from time away from work.

— The definitions they are using:

— dyspepsia: predominant epigastric pain lasting at least one month, can be associated with epigastric fullness, nausea, vomiting, or heartburn, provided that epigastric pain is the patient’s primary concern (they are trying to minimize the inclusion of GERD in this category)

— functional dyspepsia: patients with dyspepsia where endoscopy and other relevant tests have ruled out organic pathology that explains the symptoms


So, a few comments:

— there are a few items above which counter conventional approaches, including not doing endoscopy to investigate alarm features in those <60 years old (they do comment that studies do not suggest that the predictive value of alarm features is very good, though the quality of evidence to support this is very low)

— I personally think the flow of the above algorithm is somewhat flawed. Those either with functional dyspepsia or plain old dyspepsia who have significant life stressors or other psychosocial conditions should be appropriately treated for those, usually at the same time acid suppression therapy is used, instead of the old biomedical approach, reiterated above, that one tries to treat all medical problems first and if all else fails treat the psychological problems. There have been studies in the past which have shown, not surprisingly to those of us in clinical practice, that stress can produce dyspepsia, with other old studies also showing an increase in gastric acid production.

— The above recommendations apply to symptomatic patients and strongly recommend treating H Pylori infections. As mentioned in many prior blogs, I tend to be quite aggressive in diagnosing and treating H. pylori infections, even if asymptomatic, because of the association with gastric cancer. This is reinforced to me because I treat a population with a very high prevalence of H. pylori, and what seems to me a high incidence of gastric cancer (which in fact is prevalent in the countries from which they hail). See blog reference below.

— another issue is the potential problems with the long-term use of PPI therapy. As commented previously, many patients have been put on PPIs for inappropriate indications (and they are available OTC, to boot). Many dyspepsia patients do respond to H2 blockers or even calcium antacids, which do not put them at higher risk of the many possible complications of long-term PPIs (a blog will come out soon on the association of PPIs with increased overall mortality). In fact, in seven RCTs involving 2456 dyspepsia patients, there was no statistically significant difference between PPI and H2 blocker use in providing symptom relief. And, patients initially treated with PPIs very often can be stepped down to H2 blockers or antacids, though many of these patients do well with stepped down therapy​.



blog for a summary of risks and benefits of PPIs

blog  for a summary of the H Pylori treatment regimens

blog  for a review of H pylori eradication and decreases in gastric cancer

Primary Care Corner with Geoffrey Modest MD: Benzos may not increase mortality risk

17 Jul, 17 | by gmodest

by Dr Geoffrey Modest

​ The BMJ just had an article assessing mortality from benzodiazepines from a large US commercial healthcare database, showing minimal increased mortality risk (see


— 1,252,988 randomly selected patients, comparing those initiated on a benzodiazepine during a medical visit within the prior 14 days vs 1,252,988 non-initiators, from 2004-2013

— all patients were required to fill at least one prescription for any medication both in the 90 days and 91- 180 days before the index date (ie, they were plugged into medical care and filling prescriptions), and high dimensional propensity scoring was done (see below).

— Mean age 46, 85% men, mean Charlson comorbidities score 0.5 (ie, low), 5% smokers, 4% obesity/overweight, 28% hypertension, 1% heart failure, 5% ischemic heart disease, 25% hyperlipidemia, 10% diabetes, 3% COPD, 5% asthma, 10% neuropathic pain, 20% back pain, 3% kidney disease, 10% cancer, 10% anxiety, 10% sleep disorder, 11% depression, 2% drug or alcohol misuse [reaffirming that this is a pretty healthy and younger population overall]

— Medications included SSRIs in 18%, opioids in 30%, barbiturates in 2%, antipsychotics in 2%, other anxiolytics in 1%

— of note, in comparing benzodiazepine initiators vs non-initiators, prior to propensity scoring, the benzodiazepine group had more smokers, hypertensives, atherosclerotic disease, hyperlipidemia, COPD/asthma, neuropathic pain, cancer, a lot more anxiety and depression, and were much more likely to be on beta blockers, steroids, opioids (35% vs 24%!!), anticonvulsants, SSRIs (22 vs 12%), and other hypnotics. All of these characteristics were well-balanced after propensity score matching

— Short acting benzodiazepines were more frequently prescribed, 75% of the filled prescriptions, and alprazolam was the most commonly prescribed of them (47.2%), and diazepam was the most commonly prescribed long-acting agent (87.8%). On review of their supplementary materials, they did include clonazepam as a short acting benzodiazepine, though it’s half-life is actually quite similar to that of diazepam  (both about 20 hours, sometimes much more: >60 hours)​. Not sure why they did that.

— main outcome: all-cause mortality, as determined by linking to the Social Security Administration Death Master File. The overall mean follow-up was 159 days for the benzodiazepine initiators and 146 days for the non-initiators.

— secondary analysis: comparing mortality in patients initiating benzodiazepines with other active treatments (i.e. SSRI antidepressants), also with high dimensional propensity score matching



— over 6 months of follow-up, there were 12.2 events per 1000 person-years in the benzodiazepine initiators vs 6.9 events per 1000 person-years non-initiators, a 78% increased mortality risk. But, given the different baseline characteristics of these groups, probably  the most relevant finding was that after the high dimensional propensity scoring, there were 5061 deaths in benzodiazepine initiators vs 4,691 in non-initiators, 9.3 vs 9.4 events per 1000 person-years; HR 1.00 (0.96 1.04 ). ie no difference

— a 4% increased mortality risk was observed in those on benzodiazepines when the observation period was extended to 12 and 48 months of follow-up.

— benzodiazepines were associated with a 9% increased risk as compared to those starting SSRIs

— in subgroup analysis, older patients initiating benzodiazepines with a longer half-life had no increased risk of all-cause mortality, however younger patients and patients using the short-acting benzodiazepines did have a 9% increased risk.


— Propensity match scoring was used to mathematically control for potential measured confounders. The high dimensional propensity score algorithm also used above is an automated technique which prioritizes/controls for more than 300 covariates that may serve as proxies for unmeasured confounders in large electronic databases. but it is important to reinforce that even large observational studies as this one do not enable us to draw definitive conclusions about causality: there still could be unmeasured variables which are primarily responsible for any associations. This population overall was pretty healthy, those on benzodiazepines less so, emphasizing that there might well have been other significant differences between these groups (though the lack of association is reassuring, since these sicker patients, controlling for their measured sicknesses but were probably at higher risk for other unidentified sicknesses and more likely to have a less healthy lifestyle, and they did not have higher mortality than the much less sick non-benzo initiators).

— As we know, benzodiazepines are frequently used in the outpatient setting, in 2008 approximately 5.2% of US adults aged 18 to 80 used benzodiazepines, increasing from 4.1% in 1996 to 5.6% in 2013. Similar numbers were found in British Columbia, Canada. Use increases with age, with a higher usage in those older than 50, especially for anxiety and sleep disorders. The concerns about their use in the elderly is related to prior reports of a threefold or more increased risk of all-cause mortality, even for short duration usage. And concerns remain about increased falls and fractures in the elderly. it should also be emphasized that this population above is a younger one, a selection bias related to the fact that this was a commercial healthcare database.

— It seems pretty remarkable that in the overall population, 35% of those who initiated benzodiazepines were on opioids vs 24% who did not start benzodiazepines. Given the apparent higher mortality of benzodiazepines in those on opioids found in a few observational studies (eg, see blog), it would have been useful to know specifically how the opioid subgroup fared. One concern that I have regarding the potentially increased mortality of combination of benzodiazepines and opioids is whether it is really from the combination or from the patient mortality associated with the conditions that the benzodiazepines might be treating (e.g. the significantly increased mortality of panic or other anxiety disorders).


—  the increases in mortality found in the subgroup analyses above were very small, though statistically significant because of the huge number of patients evaluated​. so, clinically they found essentially no difference in those initiating benzos

— from a clinical practice perspective, this study to me is largely reassuring​: I have certainly seen many older patients (again, not well represented above) who are severely functionally affected by anxiety, resistant to non-pharmacologic therapies as well as non-benzodiazepine drugs. I have prescribed benzodiazepines in many if them with excellent results. Preferentially I have used longer acting benzodiazepines, such as clonazepam (though as noted, they consider this a short acting benzodiazepine in the above study, but i think that might be an erratum), even in patients in their 90s. Patients certainly understand the potential increased risks of falls and possible increased mortality, but are desperate for immediate symptomatic relief.

Primary Care Corner: Decreasing sudden-death in heart failure

13 Jul, 17 | by gmodest

by Dr Geoffrey Modest

A recent meta-analysis found that the risk of sudden death in patients with symptomatic heart failure and reduced ejection fraction (HFrEF) has decreased significantly over the past 20 years (see Shen L. N Engl J Med 2017; 377: 41).
— 12 clinical trials from 1999 through 2014, in which patient-level data were available, included 40,195 patients, but excluded those with implantable cardioverter-defibrillators (ICD)
— Mean age 65, 77% men, 95% with NYHA class II or III heart failure, mean ejection fraction 28% (varied from 23% to 32%), 62% with ischemic heart failure, ACE-I/ARB use was >90%.
— sudden death was determined in 3583 patients
–Those with sudden death were more often older (low 60s vs mid 60s​), male (low 80% vs mid 70%), had an ischemic cause of heart failure (70% vs 60%), and had worse cardiac function(LVEF 26% vs 29%). There was also minimally lower systolic blood pressure, minimally higher heart rate, minimally worse heart failure symptoms, minimal difference in renal dysfunction, but there was a somewhat more prominent history of myocardial infarction and diabetes in those with sudden death.​ [These were my rough calculations from the supplementary material, no formal calculations done by them]. Also NT-proBNP was higher in those with sudden death, though this was not checked in many of the studies.
— over time, there was a 44% decline in the rate of sudden death across the trials over the 19 years
— the cumulative incidence of sudden death at 90 days after randomization was 2.4% in the earliest trial, and 1.0% in the most recent trial. At 180 days the cumulative incidence of sudden death was approximately double that at 90 days, with these same trend of decrease in the more recent trials.
–My review of the individual trials showed a relatively consistent pattern: those with the same heart failure medications in both the control and experimental groups (as in a trial looking at the role of statins or another medication added on) did not show much difference between these groups in terms of sudden death outcomes. However those in which an additional heart failure medicine, either a beta-blocker or a mineralocorticoid antagonist, was added, there generally was a more impressive benefit: ie, maximizing the standard heart failure regimens seems to be beneficial. Only one study used sacubitril/valsartan, finding some added benefit on top of otherwise maximal therapy
— sudden cardiac death was not higher among patients with a recent diagnosis of heart failure than those with longer standing heart failure
–This study basically reinforces that maximal therapy for HFrEF patients significantly decreases the risk of sudden death, and to levels where automatically implanting an ICD may not be indicated, particularly in the primary prevention of sudden death. It would be important to determine if there is a subset of those patients who might benefit. One leading contender has been ischemic vs nonischemic cardiomyopathy, where the myocardial scar in those with ischemic cardiomyopathy might be the nidus for ventricular arrhythmias (and, indeed, this study confirmed about a 10% increased risk vs non-ischemic cardiomyopathy).
— Another issue is how long to wait to see the degree of myocardial recovery after the presenting symptoms of heart failure or myocardial infarction. There are studies confirming that multiple drugs and maximal doses lead to more reversal of ventricular remodeling and that the LVEF may still improve over 6 to 12 months after starting treatment (which is significantly longer than the 40 days post myocardial infarction highlighted in the ICD guidelines, suggesting that this time interval may still be too short).
–Current recommendations for ICDs include patients with NYHA class II or III symptoms and LVEF< 35%, independent of cause. However, the recent DANISH trial of patients with nonischemic systolic heart failure showed that adding ICD to maximal medical therapy  led to no difference in total mortality, though there was a reduction in sudden death in those with ICDs (4.3% vs 8.2%). It was notable in the above meta-analysis that only 9% of the patients (3180 in total) had had an ICD implanted and were excluded from this study. Given the low LVEF (median 28%) in these symptomatic patients, so that the majority actually met guideline criteria for an ICD, this 9% number suggests that we clinicians have not been pursuing ICDs very vigorously, even though these patients in the meta-analysis were more likely to have been in cardiovascular centers in big academic medical sites, replete with academic cardiologists….
— This study reinforces that we clinicians are doing quite well in preventing sudden cardiac death by putting patients on maximal heart failure medications.  These meds really work
— As with all interventions, in this case ICD placement, there are real risks for the intervention (in this case: infections, ICD failure, documented decreases in quality of life, and in my experience significant PTSD from patients having been shocked either appropriately are not). Therefore, the more targeted we are at selecting patients for this intervention, the better. Further investigation is warranted to determine those subgroups of patients who really benefit from ICD placement (and those who really do not benefit much and ICDs should be a non-issue).
see blog   for a recent review of  the medications for heart failure (including HFpEF), and another which reviewed Medicare patients showing that only 8.1% who qualified for ICDs got them

Primary Care Corner with Geoffrey Modest MD: New colon cancer screening guidelines

12 Jul, 17 | by gmodest

The Multi-Society Task Force of Colorectal Cancer (MSTF), a combo of the Am College of Gastroenterology, Am Gastroenterological Assn, and the Am Society for Gastrointestinal Endoscopy, just published significantly revised guidelines on colorectal cancer (CRC) screening (see doi: 10.1038/ajg.2017.174 ).



–They differentiate between programmatic screening/screening done in an organized system which involves consistent planning, documentation, monitoring of quality, and follow-up (which exists in many industrialized countries, as well as some healthcare plans/medical organizations in the US) vs opportunistic screening, which is largely up to the provider or designee to identify patients who need screening and arrange it

— given the potential for multiple different screens (9 are available), there are various approaches to the patient: offering the patient multiple options, though studies suggest that offering only 2 or 3 preferred options may improve patient adherence; sequential options, where the patient is offered a preferred option 1st, with subsequent options available if the 1st one were declined; or risk-stratified approach, where colonoscopy would be offered to those with a high likelihood or prevalence of advanced precancerous lesions (or potentially patients with higher than average risk, such as older patients, males, those with obesity, diabetes, or smoking), with other tests (eg FIT) being offered to patients at lower risk.


— offer CRC screening beginning at age 50 in average risk patients (strong recommendation, high quality evidence)

— any of the approaches of multiple screening options, sequential screening options, or risk stratified approaches are reasonable (weak recommendation, low quality evidence)

–cascade of tests:

tier 1 tests: colonoscopy every 10 years, or annual FIT tests. (They prefer the former in sites where there are not good follow-up systems; also may be preferable in men older than 60, and women older than 65 with no prior screening)

–tier 2 tests: CT colonography every 5 years, FIT-fecal DNA every 3 years, flexible sigmoidoscopy every 10 years (or every 5 years)

–tier 3 tests: capsule colonoscopy every 5 years (and do not offer Septin 9)

Other considerations:

— family history:

–CRC in a first-degree relative increases the risk of CRC regardless the age of diagnosis of the affected relative, though the younger the relative, the greater the risk.

–family history of CRC diagnosed <60 years of age or for those with 2 first-degree relatives with CRC or advanced adenoma at any age: colonoscopy began at age 40 or 10 years before the age the relative was diagnosed, whichever comes first, with subsequent colonoscopy every 5 years. FIT testing should be offered to those who decline colonoscopy (a randomized trial showed there was a  trend favoring colonoscopy, but this was not significant).  (weak recommendation, low quality of evidence). prefer colonoscopy

–family history of CRC or advanced adenoma in 1st degree relative diagnosed at age > 60 should have screening beginning at age 40, with normal testing intervals as above  (weak recommendation, very-low quality of evidence). prefer colonoscopy

— age:

–though the incidence of CRC in persons under age 50 is increasing,  the incidence remains low enough that they continue with 50 years old as the starting age overall, for non-African Americans​ (strong recommendation, moderate-quality of evidence).

— in African-American individuals, however, they suggest screening begin at age 45 (weak recommendation, very low quality evidence), since overall they have lower screening rates, higher incidence rates of CRC, early mean age at onset, worse survival and late-stage presentation, and a higher proportion of cancers before age 50. (weak recommendation, very-low quality of evidence): there are a few data suggesting benefit of screening African-Americans before age 50. see yesterday’s blog finding increased interval cancers in black persons soon after a “normal” routine colonoscopy

— age to stop screening: potentially beneficial in persons up to age 86 if they have not previously been screened. Should be put in the perspective of comorbidities and life expectancy. Those with negative screening, especially by colonoscopy, could consider stopping at age 75, though MSTF also suggests the option of  continuing screening until life expectancy is less than 10 years (this is a variation supported by them, as opposed to stopping specifically at age 75, though they did note that this is a weak recommendation with low quality evidence). Consider screening up to age 85 if no prior screening has been done (also weak recommendation with low quality evidence). [But, the risks of colonoscopy increase in those over 75yo (eg, see blog  )]


— adenomas are the precursors of about 70% of CRC’s, and typically take more than 10 years to develop cancer, hence the 10 year suggested interval in those with normal colons and without genetic variants (eg Lynch syndrome). Invasive cancers from adenomas <5 mm is extremely rare, and is <1% in those 6-9 mm in size

— one major concern with adenoma biopsies is the poor-to-moderate interobserver agreement in differentiating high- vs low-grade dysplasia by pathologists, as well as between tubular vs tubulovillous histology. Another concern is that sessile serrated polyps (SSP’s) are particularly common in the proximal colon, are flat or sessile in shape, are difficult to detect a colonoscopy, and, other than hyperplastic polyps, and have a significant cancerous potential. However, there is also poor pathologist interobserver agreement in the differentiation of SSP’s from hyperplastic polyps (!!!)

— Colonoscopy has the clear advantages of high sensitivity for cancer and all classes of precancerous lesions, biopsy and diagnosis can be done right away, and there is a 10 year interval between examinations if all is normal. However, though the data are impressive through case-control studies as well as indirect trials (e.g. studies of fecal occult blood testing where large numbers of patients underwent colonoscopy), it is important to remember that there are no randomized trials of colonoscopy screening at this point. Disadvantages of colonoscopy include: thorough bowel prep (a pretty unpleasant procedure), higher risk of perforation than other modalities (0.5 per 1000), high risk of aspiration pneumonia with deep sedation, small risk of splenic injury, a greater risk of bleeding especially when biopsies are done (2.6 per 1000), and death (2.9 per 100,000). Also colonoscopy is operator-dependent, and they suggest that colonoscopists should have an adenoma detection rate greater than 25% (greater than 30% for males, greater than 20% for females), and cecal intubation rates should exceed 95%. and that patients ask/be aware of these numbers/ask the prospective colonoscopist

–FIT testing: noninvasive, one-time sensitivity for cancer 79%, about 30% sensitivity for advanced adenomas, low cost. Major disadvantage is need for repeated testing, poor or no sensitivity for serrated class precursor lesions (though there is no evidence that cancers arising from this class are less likely bleeds than those arising from adenomas). Given the requirement for annual testing, they suggest this is a more viable option in systems where there is programmatic screening available, and less dependent on clinicians remembering to offer annual screening. Although the sensitivity for FIT-DNA screening is higher (92% as well as 40% for SSP’s > 1 cm), they dismiss it because of substantial decrease in specificity and high cost

— flexible sigmoidoscopy: clearly reduce cancer incidence and or mortality in randomized trials. Lower cost and risk vs colonoscopy, more limited bowel preparation, no need for sedation. They also comment that the absence of sedation leads to low satisfaction experience for patients/less willingness to repeat the examination compared to colonoscopy (I wonder if this is really true). They do comment that there’s no reason to think that sigmoidoscopy needs to be at done at 5 year intervals as previously recommended, but that 10 year intervals are probably reasonable. And there are other blogs highlighting an observational study in UK finding decreased CRC incidence and mortality even after 17 years after a single sigmoidoscopy, as well as commenting on articles showing the benefits of colonoscopy for >15 years after a negative study)

— there are specific recommendations in other documents by MSTF regarding the technical performance of both FIT testing and sigmoidoscopy/colonoscopy, in order to achieve high quality results

— also, as they comment, “the best test is the one that gets done”. I think this reinforces the need to use a patient-centered approach to ordering these tests instead of just prescribing one


— Some significant changes over prior recommendations, including placing the recommendations in tier groups instead of just listing them, elevating FIT testing to the top tier, and even suggesting that sigmoidoscopy could be done every 10 years

— I have sent out several blogs in the past suggesting sigmoidoscopy as a viable option to colonoscopy, even though it clearly misses more proximal lesions. Sigmoidoscopy has fallen out of favor for a variety of reasons in the United States, including, as noted in this recommendation, the poor reimbursement and therefore less availability of this modality. The data are mixed on the importance of finding right-sided lesions, with some studies showing minimal mortality benefit in finding and excising them (ie, the major argument against sigmoidoscopy that it does not see the right side may not matter that much).​ It would be useful to have an RCT with clear clinical outcomes assessing the relative values of sigmoidoscopy and colonoscopy.

— I must admit that I am quite impressed with the ease and acceptability for annual FIT testing, and the relatively low likelihood of requiring follow-up colonoscopy. However, I would certainly reiterate the MSTF concern about making sure there are systems to track this testing along with arranging for repeat testing. [As an aside, I should note that in Canada they recommend FIT testing every 2 years, and they actually discourage using colonoscopy for screening, given the higher cost, conscious sedation (and need a driver to take patients home), an awful bowel clean-out,  more risk,  and lack of data showing benefit over FIT testing].

Primary Care Corner: Racial disparities in interval colorectal cancer

11 Jul, 17 | by gmodest

A recent study looked at the racial/ethnic disparities in the development of interval colorectal cancer, defined as cancers that developed in a screened population but either were missed of the time of screening or developed de novo within the recommended screening/surveillance intervals (see DOI: 10.7326/M16-1154).



— a population-based cohort study of  Medicare enrollees aged 66-75 who had colonoscopy between 2002 and 2011, followed through 2013, with  linkage to the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program

— total study population 61,433: 51,313 white/4196 black/2696 Asian/1164 Hispanic

— median age at index colonoscopy 70, 60% female (though higher in black and Hispanic persons), poverty level as defined by the ZIP Code of the patient’s residence was high in 32% white/69% black/36% Asian/67% Hispanic, urban 83% white/89% black/ 98% Asian/98% Hispanic, Charlson Comorbidity Index>= 2 in 2.5% white/6.5% black/3% Asian/6% Hispanic (reflects higher rate of comorbidities)

— similar screening colonoscopy rates between black and white persons (79.5% versus 80.7%), as well as similar polypectomy rates at the index colonoscopy (23.4% versus 24.7%)

— 2735 cases of interval colorectal cancer (CRC) were identified over 235,146 person-years of follow-up

— proximal 66% white/56% black/46% Asian/62% Hispanic

— rectum 18% white/25% black/35% Asian

— other data not provided because numbers too low/concerns about protecting patient confidentiality

— interval CRC was defined as a diagnosis 6-59 months after colonoscopy

— interval CRC, when found, was mostly within three years of prior screening: 66% white/71% black/72% Asian/65% Hispanic

— the study also evaluated whether any variation in incidence was related to the quality of the colonoscopy, as measured by the physicians’ polyp detection rate (PDR), with higher rate suggesting higher quality colonoscopy performed (a validated instrument)



— the probability of an interval CRC by the end of the follow-up was 7.1% in black persons and 5.8% in white persons, a 32% increase, HR 1.32 (1.15-1.51), after adjusting for age, sex, and year of colonoscopy (this ratio did not change with further adjustment by ZIP Code, comorbidity, or whether polypectomy was done at the index colonoscopy). The probability was lower in Hispanic (4.4%) and in Asian persons (3.8%)

— 52.8% of black persons versus 46.2% of white persons received colonoscopy from physicians who had a lower PDR. This PDR rate was significantly associated with interval CRC risk. Overall, the risk for an interval CRC was higher in patients whose colonoscopy was performed by physicians in the lowest quartile of PDR, with a dose-response curve

— However, adjustment for PDR did not alter the above HRs, decreasing the HR minimally from 1.32 to 1.31. Black persons still had a 35% increased interval cancer risk in those seeing physicians in the third-highest PDR quartile and 74% increase in those in the highest quartile. Also adjustment for colonoscopy by indication did not affect the results (see below).

— the disparity was more pronounced for cancer of the rectum, with a 70% increased risk [HR 1.70 (1.25-2.31)] and a 25% increase in the distal  colon [HR 1.45 (1.00-2.11)], but a nonsignificant difference in proximal cancer [HR 1.17 (0.96- 1.42)]

Black persons also had a significant 60% higher risk of distant disease associated with an interval CRC, but not for regional or local disease



— background: CRC is the third most common type of cancer in men and women, and the second leading cause of cancer-related deaths in the US. Interval CRC accounts for 3-8% of the CRC cases.

–There are important ethnic/racial disparities: black persons have the highest incidence of and mortality from CRC, 22 to 27% higher than white persons, as well as earlier mean age of CRC onset/higher % of cancers under age 50

— It is difficult to pinpoint the causes for these racial differences. Other studies have suggested that approximately 40% of the disparities in CRC incidence are from lower utilization of screening among black persons (though not found in this study). The above study looked at quality of screening, using PDR as a surrogate measure, and did find that black persons had more colonoscopies by physicians with lower PDR suggesting lower quality screening (though adjusting for this did not find any difference). This study did not look at other potential quality measures such as cecal intubation rates, colonoscopy withdrawal times, adequacy of bowel preparation, or completeness of polyp resection. In addition, there are no data about the characteristics/aggressiveness of cancers found, such as microsatellite instability. Also, as a large data-mining study, they do not have the actual specific indications for the colonoscopy itself though they did use patient characteristics and gastrointestinal conditions/symptoms within the prior 12 months to stratify screening versus non-screening. There was also no specific data about adenoma detection rate in the data, though other studies have suggested that PDR is well-correlated in other studies

— there may well be genetic differences in colon cancer predisposition among the groups, see blog  for a discussion of the interplay between nature and nurture

— it was notable that black persons had more interval cancers in the distal colon, though still more than half of the interval cancer lesions were still in the proximal colon in general. These distal lesions are easier to reach endoscopically and have fewer difficult to-detect sessile polyps, so are more effectively assessed/treated by colonoscopy. Estimates of 10 to 16% of proximal colon tumors are missed; and about 37% of interval rectal cancers are missed (? If there are racial/ethnic differences in these numbers). Though, the efficacy of detecting proximal lesions in preventing CRC deaths is much lower than distal lesions. However, despite the fact that distal lesions are easier to detect on colonoscopy, the important shift in pickup for black persons is the 60% higher incidence of distant disease in the interval cancers.

— another caveat is that black persons in this study were more likely to be poorer, urban, female and have more medical comorbidities, so there might well have been uncontrolled biases in the above observational study

— Asian persons have similar polyp detection rates as compared to white persons in prior studies. The lower incidence of interval cancers found in the current study is consistent with slower progression to cancer in this group.


So, the study does find a pretty dramatic increase in interval colon cancers in black persons. And, at the time of these diagnoses, the cancers were more aggressive and had more distant spread. It therefore does raise the question of whether there should be different criteria for colonoscopy screening based on race.

— for example, some guidelines suggest that black persons have colonoscopy beginning at an earlier age, with a recent suggestion of age 45 (this will be commented on further in an upcoming blog on new colorectal screening guidelines).

— the study also raises the question of whether there should be shorter surveillance intervals in black persons (and perhaps longer intervals in Asian patients). I would caution here that we do not have the important clinical outcome data to support this conclusion. And, the fact that these tumors often manifested themselves within three years, many already with distant spread in black persons, may suggest that increased screening would not be useful.

–so, we really need a large prospective RCT to assess the appropriate interval for CRC screening for different racial/ethnic groups, with granular data about the quality of the colonoscopy preparation, the adequacy of polyp resection, more specifics about the aggressiveness of polyps resected, etc., prior to submitting a large number of people do a pretty difficult, invasive and not-entirely-benign screening test, as well as potentially very aggressive treatment regimens depending on the outcome. There might also be a role of FIT testing versus colonoscopy in this group, or perhaps colonoscopy at the usual intervals but more annual FIT testing in between.

Primary Care Corner with Geoffrey Modest: antibiotics for skin abscesses?

10 Jul, 17 | by gmodest

by Dr Geoffrey Modest

A recent article found that either clindamycin or trimethoprim/sulfamethoxazole (TMP/SMX) are superior to placebo for uncomplicated skin abscesses (see Daum RS. N Engl J Med 2017;376:2545-55 ).


— multicenter, prospective, double-blind trial of 786 outpatient adults and children who had a skin abscess <5 cm in diameter (or <3 cm if 6-11 months old, <4 cm if 1-8 yo), and two of: erythema, swelling/induration, local warmth, purulent drainage, tenderness.

— patients randomized to clindamycin 150 mg TID vs TMP/SMX 80/400 mg BID vs placebo, for 10 days

— 57% male, mean age 25.5 (64% > 18yo/13% 9-17yo/21% 1-18yo/2% <1yo),  62% African-American/31% white, temperature 37°C, area of wound 4 cm², surrounding erythema 27 cm²

— all had incision and drainage of their abscess

— Exclusion criteria: infection in a body site requiring specialized management (perirectal, genital, hand infection), human or animal bites, oral temperature higher than 38.5°C (38.0°C in children 6-11 months of age), systemic inflammatory response syndrome, immunosuppressive therapy, immunocompromising conditions (e.g. diabetes, renal failure), BMI >40


— S. aureus was isolated from 527 (67%); MRSA was isolated from 388 (49%)

— intention-to-treat analysis 10 days after therapy: cure rate in clindamycin group was 83%, TMP/SMX group 82% (no statistical difference), but in the placebo group was 69%, significantly lower (p<0.001)

— For those who could be evaluated (those that completed the required study visits): 93% of clindamycin group, 93% of TMP/SMX group, and 81% of placebo were cured.

— children did slightly better than adults overall, especially in the clindamycin group

— For those without S. aureus, 90.5% on clindamycin, 90.8% on TMP/SMX , and 90.8% on placebo were cured (i.e. no benefit in non-S. aureus infections by antibiotics)

— for those with S. aureus (similar numbers for MRSA and MSSA), around 95% on clindamycin, 92% on TMP/SMX , and 70% on placebo were cured

— treatment failure was mostly because of a lesion at a different body site, or use of a rescue medication (done largely in the placebo group), but rarely due to worsening of the original lesion.

— at the one month follow-up visit:

–79% of the clindamycin group, 73% on TMP/SMX , and 63% on placebo remained cured.

–new infections at one month were less common in the clindamycin group (7%), than in the TMP/SMX group (14%), p=0.03, or the placebo group (13%), p=0.06

— adverse events were more common with clindamycin (22%) than with TMP/SMX (11%) or placebo (13%). All adverse events were without sequelae. One participant on TMP/SMX had a hypersensitivity reaction (fever, rash, thrombocytopenia, and hepatitis). Most common adverse events were diarrhea (16% clindamycin, 5% TMP/SMX, 7% placebo) and nausea (2%, 4%, 2%). No C difficile infections


— Skin abscesses are quite common, affecting 4% of people in the United States annually. Often these are treated as outpatients with clindamycin or TMP/SMX,  given the large percent of community MRSA

— In this study TMP/SMX was effective at a lower dose than often prescribed, though a 10-day course was given (vs 7 days). It is possible that the cure rates might have been higher with higher doses of TMP/SMX

— 13  patients had resistance to clindamycin and did not fare as well (54% cure vs 85% in clindamycin-susceptible infections). There were no cases of TMP/SMX resistance

— this study suggests that antibiotics help, though held-wisdom previously was that there was not much additional benefit after incision and drainage alone.

— there were minimal severe adverse reactions. Prior studies have found about 5% on TMP/SMX have severe reactions (and 1 person did above), and these can be fatal.

— for clindamycin, there was a meta-analysis (see Brown KA. Antimicrobial Agents and Chemotherapy 2013; 57: 2326) that found that, as compared to no antibiotic exposure, the Odds Ratio of C difficile infection for clindamycin was 16.80, fluoroquinolones 5.50, cephalosporins 5.68, macrolides 2.65, TMP/SMX 1.81

— it was mentioned in the supplementary materials of the current study that 33 in the placebo group (13%) used “rescue meds”, whereas 12 in clindamycin (5%) and 15 in TMP/SMX (6%)​ did. There is no comment on what meds were used or what the outcomes were for those who continued with their assigned meds/placebo and did not require rescue meds


–it seems quite likely from this study that antibiotics help, as also noted in a prior blog , a study comparing TMP/SMX at 4-fold higher doses (320/1600 mg bid) finding benefit from the antibiotic but noting that >80% responded just to I&D.  And, 80+% of people seem to get better without antibiotics (in several other studies), and even in this study, 81% of those completing the study were cured. (70% of those with documented S aureus infections).

— there are really bad/occasionally lethal short-term potential adverse events from these antibiotics, especially severe systemic reactions to TMP/SMX and severe C difficile infections with clindamycin

–and there are potential long-term bad effects on the microbiome. As an example, this blog reviews a large prospective observational study finding antibiotic usage was associated with later development of colonic adenomas, including high-risk ones

–it was also notable in this study that there were not many cases of the abscess area getting worse in the placebo group with “treatment failure”

–so, in lower risk patients (nondiabetics, immunocompromise, etc), it might be reasonable just to watch the patient, holding the antibiotics but prescribing them if they were not getting better in a few days. I would be inclined to use antibiotics even in this lower-risk group if there were uncertainty about being able to have close followup (returning to clinic or phone followup), to assess the progress.

Primary Care Corner with Geoffrey Modest MD: Teens birthrate and sexual activity/contraception use

28 Jun, 17 | by gmodest

by Geoffrey Modest

2 related articles were recently published by the CDC.

1. In the United State from 1991 to 2014 the birth rate among teens age 15 to 19 hasdeclined by a dramatic 61%from 61.8 per 1000 to 24.2 per 1000, with larger % decreases in ethnic/racial minorities (see ). However the birth rate remained approximately twice as high for Hispanic and non-Hispanic black teens compared to non-Hispanic white teens. There are also significant geographic and socioeconomic disparities. See prior blog  for the full assessment . In brief:

— from 2006 to 2014, a 41% decline in birthrate overall, to 25.4 per 1000 female teens

— Hispanic: decreasing to 39.8 per 1000, a 51% decrease from 77.4 per 1000

— Black: decreasing to 37.0 per 1000, a 44% decrease, from 61.9 per 1000

— white: decreasing to 18.0 per 1000, a 35% decrease from 26.7 per 1000

— there is large geographic variability, for example the Hispanic birthrate in 2014 varied from 17.0 per 1000 in Maine to 58.0 per 1000 in Oklahoma; Black birthrate varied from 14.0 per thousand in New Hampshire to 54.6 per thousand in Arkansas; white birthrate varied from 4.8 per 1000 in New Jersey to 39.2 per 1000 in West Virginia. And, within states, sometimes the racial disparities remain very high: eg, in Nebraska the birth rate for whites was 16.2 (approx the national average) whereas the rates for black and Hispanic (42.6 and 53.9) were far above the national average for these groups.

–and, there was a higher birth rate in those who are unemployed, have lower levels of education and lower incomes.


–as a perspective, the overall US birth rate in 2011 for 15-19 years old females was 34.0 per 1000, vs 13 per 1000 the same year in Canada. And the rate in France was 7 per 1000 and in Germany 5 per 1000.


–very dramatic changes in teen birth rate, with a narrowing of the gap for racial/ethnic minorities as compared to whites. However, as noted, the gap remains significantly discordant.

–And, I suspect, a large part of the geographic gap reflects access to and quality of care. And there are major concerns about the future: the trump administration etc are pushing for decreased Planned Parenthood (potentially leading to even less access to care/contraception for poor and minority patients, with anticipated increases in pregnancy rates, and likely maternal and fetal death rates) and even cutbacks in maternity care. For example, another blog showed that a restrictive abortion law in Texas led to an 18.2% decline in abortions; and there have been a plethora of studies linking lack of prenatal care to poorer outcomes.

–and, the overall social environment, getting worse in the trump era, no doubt will add to the problem: lower incomes, cutbacks in social programs (including perhaps health insurance), fewer safety net programs overall, and social upheaval in general (including targeting immigrants) will predictably lead to even less access to care, less sense of hope for the future, lower self-esteem, and poorer health outcomes, including pregnancy rates.


2. Another CDC article evaluated sexual activity and contraceptive use among teens aged 15-19 in the US from 2011-2015 (see,finding:


–42.4% of never-married female teens (4.0 million) and 44.2% of never-married male teens (4.4 million) had sex at least once by the time of the interview; these numbers were similar to those from 2002 and 2006-2010, though looking back to 1988 there was a decline (downward trend, with p<0.05). By ethnicity/race, from 2002 to  2011-15:

–Hispanic female: 37.4% in 2002, 41.2% in 2011-15; Hispanic male: 54.8% to 45.7%

–non-Hispanic white female: 45.1% in 2002, 44.3% in 2011-15; non-Hispanic white male: 40.8% to 42.8%

–non-Hispanic black female: 56.9% in 2002, 45.7% in 2011-15; non-Hispanic black male: 63.3% to 58.6%

–of these, the differences in females in 2011-15 were not statistically significant; though the difference/decrease in non-Hispanic black males was significantly higher than the others

–assessing sexual activity by family structure at age 14: for females, a significantly lower % were sexually experienced if they lived with both parents (36.8% vs 50.8%) and for males (39.4% vs 51.9%); for males, if their mothers gave birth to first child by age >=20 (39.3% vs 56.7% if mothers younger). Also for males, less sexually experienced if mother had at least some college vs high school diploma or GED (41.0%​ vs 46.7%)

–by age: males more likely than females to have sex younger (age 15-16). Though probabilities were the same by age 17

–partners for first sexual experience: 74.1% of females but 51.1% of males were “going steady” with their partner; 13% of females and 27.3% of males with “just friends” (the latter especially true with younger teens)

–reasons for not having sex: most common: religion/morals (35.3% of females, 27.9% of males), then “not found right person yet” 21.7% females, 28.5% males; then “don’t want to get pregnant” 19.3% females, 21.2% males

–no change from 2002 in terms of % of teens who have had sex in past 3 months, though older teens (18-19 yo) were twice as likely as those 15-17.


–female teenager use of contraception at first sex:

–increased from 74.5% in 2002 to 81.0% in 2011-15, though lowest in non-Hispanic black teens at 62%, and higher in Hispanic (79%) and non-Hispanic white (87%) teens; overall, mostly using condoms (66.4% in 2002, increasing to 74.6% in 2011-15)

–dual protection (condom plus pill) also increased significantly from 13.1% in 2002 to 18.5% in 2011-15

–overall 5.8% of females had a long-acting reversible method (IUD in 2.8% or implant in 3.0%) in 2011-2015

–male teenager use of contraception at first sex:

–condom use increased from 70.9% in 2002 to 79.6% in 2006-10 and remained stable at 76.8% in 2011-2015

–emergency contraception use has increased significantly from 2002 (8% of female teens who ever used it) to 2011-15 (23%)

–no change in ever using condoms (97%), withdrawal (60%), pills (56%), depo-medroxyprogesterone (17%)

–feelings about hypothetical pregnancy (which does correlate with risk of teen birth, pregnancy risk behaviors): in 2011-15, more females would be very upset (60.5%) vs males (46.1%)

–teens who have sex at an earlier age are not only less likely to use contraceptives at time of first sex, but also at the last sex as well.



–although there are lots of statistics above, the report has even lots more….

–it is notable that 81% of females used contraception the first time they have sex, condoms were used by 77% of males the first time they have sex

–overall trends seem to confirm some decreases in sexual activity overall and increase in contraceptive usage since 1988, aligning with the observed decrease in teen pregnancy in the early 1990s (as above). though the contraceptive use largely plateaued or only decreased slightly since 2002, the types and effectiveness of contraceptives used has changed

–there is a general trend to promote long-acting reversible contraceptives (IUDs implants) in adolescents (eg, see ), and there is dramatically increasing use of emergency contraception

–one important potential utility of all of these statistics is to help us clinicians and public health people focus on teens where the statistical predictors point to those at highest risk of not using contraceptives or becoming pregnant (eg. family and social situation, education, age of first sex, ethnicity/race…)


–but, as noted in commentary after the first article, there are real concerns about the future, especially with access to high quality, affordable care (see above)

Primary Care Corner with Geoffrey Modest MD: Yoga for chronic low back pain

26 Jun, 17 | by gmodest

​​by Dr Geoffrey Modest

A recent community-based study of chronic low back pain found that yoga was non-inferior to physical therapy for function and pain (See doi:10.7326/M16-2579).


— 320 predominantly low income adults with nonspecific chronic low back pain were enrolled in a 12 week, single-blind, 3 group randomized trial, with a 40 week maintenance phase.

— Mean age 45, 60% female, 18% non-Hispanic white/58% non-Hispanic black/14% Hispanic, 32% earned college degree or higher, 45% currently employed, 60% with annual income less than $30,000, BMI 31, mean back pain intensity 7 (on scale of 11), RMDQ score 15 (score range 0 to 23, higher score means worse function), 70% on pain medications (52% NSAIDs/35% acetaminophen/20% opioids), comorbidities of hypertension 35%/neck pain 30%/pulmonary disorders 25%/diabetes 18%/depression 20%. [the Roland Morris Disability Questionnaire (RMDQ) of 15 and pain intensity score of 7 reflect moderate to severe pain]

— the study was done in an academic safety net hospital and 7 affiliated federally-qualified community health centers in ethnically diverse neighborhoods in Boston

— the 3 groups were:

— 12 weekly 75-minute yoga classes, each class beginning with relaxation and meditation exercises, yoga breathing, and yoga philosophy; then yoga poses; and then relaxation. 30 minutes of daily home practice were encouraged, with patients getting a DVD, manual, and take-home yoga supplies

— fifteen 60-minute PT visits, including treatment-based classification, graded exercise, and screening for fear-avoidance beliefs

— an educational book (the Back Pain Helpbook) and newsletters on low back pain self-management, stretching, strengthening, and the role of emotions and fear-avoidance

— the maintenance phase compared yoga drop-in classes vs home practice; or PT booster sessions vs home practice

— primary outcomes: back-related function as measured by the RMDQ, and pain measured by an 11 point scale, both measured at 12 weeks. Prespecified noninferiority margins were 1.5 on the RMDQ scale and 1.0 for the pain scale.

— Secondary outcomes included pain medication use, global improvement, satisfaction with intervention, and health-related quality of life


— median yoga attendance was 7 classes, median PT attendance was 7 appointments.

— home practice was reported by 75% of yoga participants (for median of 27 minutes) and 64% of PT participants (median of 4 exercises, 4 days per week)

— fewer than half of the participants met the predefined adherence goal: at least 9 yoga or 11 PT sessions

— of 59 participants randomly assigned to yoga drop-in classes during the maintenance phase, 53% attended at least one class (median 13); of 54 participants randomly assigned to PT booster sessions, 56% attended at least one (median 2)

— primary outcomes:

— improvement in RMDQ for yoga -3.8, which was noninferior to PT, -3.5. Education was -2.5

— decrease pain for yoga was -1.7, which was noninferior to PT -2.3. Education was -1.4

— yoga and PT were not superior to education at 12 weeks, however both yoga and PT were more likely than education to have a clinically meaningful response in RMDQ (which they define as 3.0 points), but yoga did not reach quite reach their meaningful cutpoint of 2.0.

— secondary outcomes:

— greater than 30% reduction in RMDQ: 48% yoga vs 37% physical therapy vs 23% education

— greater than 30% reduction in back pain 35% yoga vs 43% physical therapy vs 25% education

​– at 12 weeks, use of any pain medication was 55% yoga (33% acetaminophen, 33% NSAIDs, 23% opioids) vs 54% using any pain medication with PT (22% acetaminophen, 43% NSAIDs, 14% opioids) vs 75% using any pain medication in the education group (39% acetaminophen, 48% NSAIDs, 18% opioids)

— self-rated global improvement: 34% improved with yoga vs 42% PT vs 21% education

— satisfaction with intervention: 43% very satisfied with yoga, 50% with physical therapy, 21% with education

–no difference between groups for either SF-36 physical health score or SF-36 mental health score [Short Form-36’s are validated questionnaires]

— in patients who were adherent to the protocol, mean RMDQ changes at 12 weeks were: -4.64 yoga, -5.7 PT, -2.74 education; clinically meaningful improvements in RMDQ in those adherent to the protocol occurred in 57% for yoga, 56% PT, 21% for education. For pain intensity 12 weeks, -2.1 for yoga, -2.6 or PT, -1.34 education, with clinically meaningful pain improvement were present in 50% for yoga, 52% for PT, and 14% with education

— in the maintenance phase, RMDQ or pain changes did not differ significantly between yoga drop-in and yoga home practice, or between PT booster sessions and PT home practice.

— Adverse effects were mild, self-limited joint and back pain.


— as we all know, chronic low back pain is both extraordinarily common (estimates of 10% of the US adults), costly (total costs, including work loss, exceed $200 billion per year), but has low patient satisfaction with treatment.

— Physical therapy is the most common evidence-based, reimbursable, nonpharmacologic therapy prescribed by clinicians. There are several studies suggesting that yoga also helps. So, this study is helpful for showing that yoga is on a par with physical therapy in decreasing pain and improving function.

–although they found that yoga/PT were not superior to education for both function and pain, there were meaningful improvements for RMDQ with yoga and both for PT, the improvements were maintained after one year (which unexpectedly was irrespective of whether the patients had ongoing yoga or PT booster classes or not), and the likely explanation of these results was the low adherence rates to the yoga/PT programs. Looking at the adherent patients, the results were dramatically more impressive for RMDQ and pain, as highlighted above.

— I should also note that several important findings  (eg, use of medications, patient satisfaction, >30% improvements) were also really impressively better with yoga/PT

–this study was particularly useful to many of us because it focused on community-based care in poor and non-white areas, in which there has been documented increased incidence of pain and disability from low back pain, fewer referrals for specialists, and less-intensive rehabilitation for occupational back injuries. Yoga classes, however, are less available and less used in poorer and nonwhite areas

–unfortunately, the use of yoga was not able to decrease the use of opioids much, though other pain meds were decreased



–yoga may well be a useful ​aid for patients with chronic low back pain

–yoga may also be cheaper for patients who have high copays for PT

–and, yoga instruction is readily available to patients with internet connection or DVD players. My overwhelming experience is that patients who do PT for back pain (or other symptoms) rarely continue with significant home-based exercises to maintain their PT benefits, frequently requiring repeat PT programs in the future and/or persistent pain/functional loss. Perhaps one of the difficulties for patients is transitioning from the array of PT therapies (eg using expensive exercise machinery or treatments unavailable at home) to the home-based, low-tech PT exercises.  Yoga might provide a more fluid transition from office-based yoga training and performance, since patients continue to do the same thing at home.

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