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 January, 2014

Primary Care Corner with Geoffrey Modest MD: more vitamin d studies, including multiple sclerosis and dosing in infants

22 Jan, 14 | by EBM

several articles have come out over the past few months on vitamin d. there are a slew of them from before, including review articles (eg  N Engl J Med 2007;357:266-81) highlighting the vast array of possible important actions of vitamin D in immune function, cancer prevention, heart disease prevention, as well as the better documented effects in bone health and preventing falls in elderly. last year there was an article finding that obese adolescents with low vitamin D had improved insulin sensitivity on repleting vitamin D levels (see  doi: 10.3945/ajcn.112.050013). here are 3 more articles with more mixed results.


1. large meta-analysis of vitamin d supplementation and bone mineral density/BMD (see included 23 studies (mean duration 2 yrs, 4082 people, 92% women, average age 59). mean baseline 25-OH vit D was <50 nmol/L (<20 ng/ml) in 8 studies. in general they found not much benefit from vit d supplementation, the only really significant one being in femoral neck BMD, specifically in those with the lowest baseline 25-OH vit D levels and in those supplemented with <800 as opposed to >=800 IU/d.

a couple of issues. there seem to be large individual variations in the physiologic effect of vit D deficiency, with only approx 50% having secondary hyperparathyroidism (none of the studies geared therapy to the hyperPTH group). and, somewhat surprisingly, in this meta-analysis the improvement in BMD was not in the highly cortical bone of the forearm, where hyper PTH has the most osteopenic effect and is the purported mechanism for osteopenia. the authors argue that vitamin D is mostly important for the maintenance of circulating calcium in a specific range, and in fact in high doses can stimulate osteoclastogenesis. it is important to keep in mind that BMD is a surrogate marker, which does not necessarily predict risk of fracture (BMD is a quantitative assessment of bone density, not a qualitative evaluation of its microstructure/strength, and there can be discordance — eg fluoride increases bone density but actually increases risk of fracture).

it is not easy to reconcile this meta-anal with others. i posted last year on a meta-anal in NEJM (see  DOI: 10.1056/NEJMoa1109617) which included 31K people (mean age 76 and 91% women) which found a nonsignificant 10% reduction in risk of hip fracture and a significant 7% reduction in risk of any nonvertebral fracture, with those in the highest vitamin D intake (median 800 IU) having a significant 30% reduction in risk of hip fracture and 14% reduction in risk of any non-vertebral fracture. this result was consistent across age groups, community-dwelling vs in an institution, baseline 25-OH vitD levels, and additional calcium intake, concluding that higher dose vitamin D (>=800IU/d) was “somewhat favorable in prevention of hip fracture and any nonvertebral fracture in persons 65 years of age or older”.

2. article in JAMA neurology looked at vitamin d status and progression of multiple sclerosis (see doi:10.1001/jamaneurol.2013.5993). this study was an intervention study looking at the efficacy of early vs delayed use of b-interferon in pts with clinically isolated syndrome (CIS) to assess progression to clinically definite MS (CDMS) or MS defined by clinical and MRI criteria (MDMS).  in this 5 yr study, they measured 25-OH vit d in the beginning and several points in the study. 400 pts. finding:


–average 25(OH)D following a CIS strongly associated with MS activity and progression (MRI and clinical) of MS over 4-5 years.

–those with serum 25(OH)D >50nmol/L (20 ng/ml) had 4x lower change in T2 lesion volume on MRI, 2-fold lower rate of brain atrophy, and lower disability on the Expanded Disability Status Scale (EDSS)

–and, the above findings were in a population on meds to decrease MS progression/relapse


there have been observational data that low vitamin D levels are associated with development of MS, thought to be related to the vitamin D effects on immune function (there are vit d receptors pretty much everywhere in the body, not just the bone and muscle).  i sent out an article 1-2 years ago which showed improved response to TB meds in a cohort also given vit d supplementation (see vit d and tb treatment pnas 2012 in dropbox, or doi:10.1073/pnas.1200072109/-/DCSupplemental).  but, the current study was an observational component of an intervention trial for b-interferon, and it is limited by inherent biases of observational studies (and those with higher vit D levels in the study tended to be younger, have lower BMI, lower number of T2 lesions on MRI, and higher brain volume when they presented with CIS, though the above results did control for age, sex, treatment time and T2 lesion score)). also almost all white. but at least they looked at patients with very early MS symptoms (independent of vitamin D levels) and looked at clinical/MRI progression. also, no ceiling effect was observed with vit D levels.


3. study done in montreal to assess different doses of vitamin d in breast-fed one-month olds and assess 25(OH)D levels at 3,6,9 and 12 months of age (seeJAMA. 2013;309(17):1785-1792). vitamin D3 doses were 400, 800, 1200, or 1600 units.  current recommendations are 400IU/d supplement in US and Canada, though some countries have higher recs (france and finland recommend ?1000IU/d).  results, assessing 2 different 25(OH)D goals:


–for the goal of 25(OH)D level >75mmol/L (30ng/ml): at 3 months, 55% in 400IU group, 81% with 800IU, 92% with 1200IU, 100% with 1600 IU.  But not sustained at 12 months in any group, with falloff proportional to dose: about 35% of those on 400IU/d up to 80% in those on 1600 IU/d

–for the goal of 25(OH)D level >50mmol/L (20ng/ml): essentially all achieved this goal at 3 months and at 12 months.

–bone mineral content and plasma PTH levels increased over time, while ionized calcium, and urinary calcium/creat ratios declined over time,  but no diff between groups in any of these measures

–of note, the 1600IU dosage was discontinued during the study because of 25(OH)D levels that were felt to be dangerously high.


so, all in all, seems reasonable to continue with the 400IU recommendation at this point, since by all of the markers noted including bone mineral content, this was as good as higher doses. the real clinical outcome, both for bone and non-bone benefits of vitamin D, will require longer term followup into later childhood. other issues include the applicability of the results of this study to kids of different ethnicities/skin pigmentation, or living at different latitudes.


so, overall, the data on vitamin D remains mixed. my sense, pending large intervention studies, is that there may well be important benefits to correcting vitamin D deficiency (decreased falls in elderly have been documented, but also effects on bone, immunologic function, potentially preventing cancer) outweigh the pretty much insignificant toxicity (except for patients on very high doses)



Primary Care Corner with Geoffrey Modest MD: type 2 diabetes breakthrough?

21 Jan, 14 | by EBM

a new article appeared in Nature which may be an important breakthrough in diabetes care/prevention in the future, targeting probably the primary pathophysiologic disturbance in the development of type 2 diabetes (see  doi:10.1038/nature12656). adiponectin has been known for years as a naturally occurring hormone produced in adipocytes, with lower levels found in obese people and those with diabetes. there is a strong relationship between lower adiponectin levels and visceral adiposity as well as the development of insulin resistance, and by increasing adiponectin levels there is improved insulin sensitivity (presumably through increasing fatty acid oxidation and inhibition of hepatic glucose production) and lowered blood glucose levels. adiponectin levels are increased by the glitizones and that may be their mechanism to improve insulin sensitivity. in addition adiponectin has anti-inflammatory and anti-atherogenic effects in several studies.

with this background, there was this japanese study of a molecule (AdipoRon), an oral adiponectin analog, which stimulates both of the known adiponectin receptors. this med was found to improve insulin resistance and glucose tolerance when normal mice were fed a high-fat diet, to improve diabetes in genetically obese mice, and to prolong the otherwise shortened life span in these genetically obese mice fed a high-fat diet.

so, fyi. the main reason i am posting this is that, to me, having low adiponectin levels is associated with the development of many of the characteristics of metabolic syndrome and diabetes, and at least theoretically increasing the adiponectin effect medically may be an important future intervention in preventing the development of type 2 diabetes and its complications. (of course, it would be even better if we could develop a comprehensive public health program to decrease obesity and increase exercise in the population as a whole, which is even more physiologic…..).


Primary Care Corner with Geoffrey Modest MD: varenicline, combo Rx and use in mental illness

21 Jan, 14 | by EBM

there were a few important articles in the jan 8th JAMA on smoking — i will highlight 2 of them on varenicline, a partial agonist which binds with high affinity as well as specificity to neuronal acetylcholine receptors. smoking cessation in smokers is the most significant single intervention to decrease mortality, and 62% of deaths in female smokers/60% of deaths in male smokers are attributable to smoking.


1. i had postedt a general smoking cessation article about one year ago from JAMA which referenced some small study finding benefit of combining varenicline with bupropion. the current study (see  doi:10.1001/jama.2013.283185) was a 12-month multi-center RCT comparing varenicline with bupropion vs varenicline with placebo, with both medications dosed as is generally recommended.  results:


–500 patients (mean age 42, 94% white, 45% women, 78% completed at least some college — study done at Mayo clinic and univ minnesota), though only 62% completed the study (dropout rates similar in both groups)

–of those completing study:

–at 12 weeks: 53% in combo group achieved smoking abstinence and 56% 7-day point-prevalence smoking abstinence (ie, self-reported tobacco cessation for previous 7 days, confirmed by carbon monoxide level), vs 43% and 49% with varenicline alone

–at 26 weeks: 37% with prolonged and 38% with 7-day point prevalence abstinence in combo and 28% and 32% with varenicline alone

–at 52 weeks: 31% with prolonged and 37% with 7-day point prevalence abstinence in combo and 25% and 29% with varenicline alone

–adverse effects: combo with more anxiety (7% vs 3%) and depressive sx (4% vs 1%) — a bit surprising since the combo group had bupropion, a pretty potent anti-depressant….

–of the above, the only findings reaching statistical significance were: smoking abstinence at 12 and 26 weeks. no significance at 52 weeks, or any of the measures of 7-day point prevalence.

–subgroup analysis: for heavy smokers (>20 cigs/day), combo therapy assoc with statistically significant prolonged smoking abstinence at 12 weeks (50% vs 36%), 26 weeks (35% vs 19%) and 52 weeks (32% vs 17%), with similar numbers for those with high levels of nicotine dependence.

–less weight gain in combo group than in varenicline alone (1.1 vs 2.5kg at 12 weeks, 3.4 vs 3.8kg at 26 weeks, and 4.9 vs 6.1kg at 52 weeks)


this trial confirms the concept of combination therapy and extends it to varenicline plus bupropion. other studies have found improved abstinence rates with combo of bupropion and nicotine replacement therapy (NRT), and with using combo NRTs (eg patch for maintenance nicotine levels, with inhaler or lozenge for times of increased urge to smoke).  this study, though small numbers of patients and pretty high dropout rate, does suggest that the combo of bupropion and varenicline is effective, esp in those who are heavy smokers.


there are minimal studies of electronic cigarettes, as per a prior blog. mostly short-term (eg 12 weeks: 14% abstinence with e-cigarettes vs 4% with nicotine-free device; another study for 6 months found no statistical difference between e-cigarettes, nicotine patch or placebo). of note, the FDA has found important carcinogenic contaminants in 2 brands of e-cigarettes: polycyclic aromatic hydrocarbons and nitrosamines. (though my experience in a relatively small number of smokers is that the e-cigs did help them quit, and cigarettes have many more dangerous additives. however, i and others are concerned that e-cigarettes may change the rather pervasive anti-smoking ethic developed in this country over many years and make tobacco more acceptable and even chic for young people).


2. the above study excluded patients with major medical or psych illnesses, and there is concern that varenicline could exacerbate psych problems. in addition, the baseline smoking rates are significantly higher in people with mental illness (25% vs 19%, in 2011). not surprisingly, those with mental illness who received mental health treatment quit smoking at a higher rate than those who did not receive treatment (37% vs 33%). another JAMA article looked at patients with schizophrenia or bipolar disease who had quit smoking on varenicline and then were put on maintenance of varenicline plus cognitive behavioral therapy (CBT) vs CBT alone (see  doi:10.1001/jama.2013.285113), as follows:


–of 203 smokers (60% smoking more than 20 cigs/d) with medically-treated and stable schizophrenia or bipolar dz, then treated with varenicline in a 12-week open trial, 87 had at least 14 days of continuous abstinence by the end of this period (88% had schizophrenia, 12% bipolar)

–these 87 abstinent patients were randomized to continued varenicline to complete 52 weeks, along with tapering CBT sessions (median of 26 sessions), vs  placebo and CBT (24 sessions)

–at the end of the varenicline intervention (52 weeks), point prevalence abstinence rate of 60% with varenicline vs 19% placebo

–at the end of the study (76 weeks — ie, 24 weeks after stopping intervention), abstinence rates of 30% vs 11% (pretty impressive!!)

–and, perhaps most noteworthy, no difference in adverse events with varenicline in the open-label part of the study, and there were actually fewer reports of agitation or excitement with the maintenance varenicline than placebo (though more headaches)


so, small study, but does suggest that we can use varenicline in patients with schizophrenia or bipolar disease, which i think many of us were wary about.



Primary Care Corner with Geoffrey Modest MD: COPD patients at risk post MI–don’t withhold the beta blocker, but be careful

21 Jan, 14 | by EBM

There have been several articles over the past 1-2 decades suggesting that the vast majority of patients denied b-blockers post-MI (most of whom had COPD) clinically should have been prescribed b-blockers, albeit carefully. a recent analysis in BMJ retrospectively accessed a large UK electronic database from 2003-8, assessing whether COPD patients with an MI were prescribed b-blockers, and what their clinical outcome were (see doi: 10.1136/bmj.f6650).  results:

–there were 1063 patients with COPD who were hospitalized with an MI in the cohort. not surprisingly, some differences between those prescribed b-blockers and those not (those put on b-blockers were younger and less likely to have history of hypertension, PAD, heart failure, dyslipidemia, angina prior to MI). GOLD staging was similar between the groups. only 40% of patients with MI and COPD were prescribed b-blocker while in the hospital

–med adherence was 80% at one year, decreasing to 60% at 2 years (suggesting those prescribed b-blockers tolerated them pretty well)

–those started on b-blockers during the admission for MI had significantly increased survival (adjusted hazard ratio of 0.50), with median followup of 2.9 years. one-year survival in those prescribed a b-blocker during the hosp admission for MI also had adjusted hazard ratio of 0.48.

–those already on b-blockers when admitted to hospital with their MI had survival benefit at well (hazard ratio of 0.59)

–cause of death evenly divided between cardiac and non-cardiac causes, and both were decreased by about 50% in patients put on b-blockers (a bit surprising, but they cite a paper — , or BMJ 2011;342:d2549 — finding that b-blockers were associated with decreased COPD mortality as well as cardiovasc, in line with their findings)

so, one issue here is that patients with COPD are at higher risk for cardiovasc dz (common risk factor/smoking, perhaps inflammation). recent studies have challenged the “prevailing wisdom” that b-blockers are dangerous to use in patients with COPD by increasing bronchospasm (that being said, it makes sense to start low, titrate up more slowly, and preferentially use cardioselective b-blockers — eg atenolol, metoprolol, though in this UK group bisolprolol was most commonly prescribed). based on a few studies done in the US showing tolerability of b-blockers and suggestive evidence that survival is improved post-MI, the use of b-blockers post-MI has increased into the 90+% range in patients with COPD. the patients in this study put on b-blockers were younger and healthier. they tried to mathematically control for these disparities, but this observational study might well have had unmeasured confounding biases.

bottom line: this is an observational study and may well be flawed. but the convergence of evidence suggests strongly that COPD patients, who are at high risk for CAD, should be given a trial of b-blockers post-MI, although probably beginning at low dose and titrated up as tolerated, and preferentially using a cardioselective one. (i think the same argument holds for patients with heart failure and significantly reduced ejection fraction).  i would, however, be especially careful in those patients with asthma or those with COPD but a strong reversible component.


Primary Care Corner with Geoffrey Modest MD: Mediterranean diet and CAD primary prevention

21 Jan, 14 | by EBM

The recent blog on the Mediterranean diet and diabetes prevention is a substudy of the larger PREDIMED study, and i forgot to mention that i had posted another article on this study finding decreased development of cardiovascular disease with either of the Mediterranean diets (with extra–virgin olive oil or nuts) as compared to the low-fat diet.  this blog post is  below.  geoff

Nejm with article on mediterranean diet in primary prevention (see r DOI: 10.1056/NEJMoa1200303), as follows:


–7500 spanish patients aged 55-80, 57% women, with high cardiovascular risk (type 2 dm, or at least 3 risk factors of smoking, htn, inc LDL, low HDL, overwt/obese, or fam hx premature cad) but no evident cardiac disease. Other baseline characteristics: 40% on statins, 20% on antiplatelet rx, 50% on ACE-i. pts assessed for primary outcome of major cardiovasc event rate (MI, stroke, of death from cardiovasc causes). put on one of 3 diets:

-mediterranean with extra olive oil (given 1 qt of extra-virgin olive oil/week)

-mediterranean with extra mixed nuts (given 30 g mixed nuts/day)

-control diet (advise to reduce dietary fat).

–study stopped early, after 4.8 yrs, with (compared to control diet):

-30% decrease in primary outcome in those on medit diet and olive oil

-28% decrease in primary outcome in those on medit diet and mixed nuts

-these translate to absolute risk reduction of 3 major cardiovasc events per 1000 person-yrs


These studies fit in with others i’ve sent out, which show that mediterranean diet decreases metabolic syndrome, has anti-inflammatory effects, improves endothelial function.


And, as with the few other dietary intervention studies in high risk pts (eg Lyon diet), the magnitude of the effect (30% reduction) is of the same order of magnitude as putting patients on statins.  Reinforces the importance of diet/exercise in preventing disease as a very powerful tool



Primary Care Corner with Geoffrey Modest MD: The many functions of albumin in cirrhosis

21 Jan, 14 | by EBM

recent review of the role of albumin in cirrhosis (see HEPATOLOGY 2013;58:1836-1846), which strays a bit from the primary care focus of the vast majority of these blog posts, but i always find it interesting when new data significantly changes the way we look at how the body works.  in this case, albumin therapy was always presented as a plasma volume expander, but it turns out that there are many other roles it plays (and who knows which is the most important physiologically???). basically, its biological roles (in patients with cirrhosis) include:


–plasma volume expansion (provides 75% of the plasma oncotic pressure) and is >50% of total plasma proteins

–binds to diverse array of molecules (bile acids, hormones, metals, long-chain fatty acids, nitric oxide, endotoxins….): solubilizes , transports them

–the thiol group on albumen is the most important extracellular antioxidant (80% of the thiol groups are on albumin)

–by binding to endotoxins, in vitro studies show that albumin decreases endotoxin activity

–albumin inhibits TNF-a upregulation of vascular adhesion molecules.

–albumin binds to nitric oxide and can lead to vasodilation and inhibit platelet aggregation

–endothelial stabilization, perhaps related to albumin’s ability to modulate inflammation, reduces oxidative damage and interferes in neutrophil adhesion

–in cirrhosis, double effect of decreased albumin quantity as well as dysfunction as above

–clinically, albumin is useful in pts with cirrhosis and spontaneous bacterial peritonitis, hepatorenal syndrome, paracentesis-induced circulatory dysfunction.

–in those without cirrhosis, some data suggests efficacy in women with ovarian hyperstimulation syndrome and in pts with malaria, burns,  and possibly has protective role in those with ischemic stroke and Alzheimer’s.


so, seems to be more than just a volume expander…. (and again shows that our models of physiologic function are often rather primitive. and need to be continually challenged)



Primary Care Corner with Geoffrey Modest MD: Mediterranean diet for diabetes prevention

21 Jan, 14 | by EBM

recent spanish study, sponsored mostly by the spanish governnment but with material support from the california walnut commission and morella nuts (though without their input on design, data analysis or reporting of results) assessed the effect of a Mediterranean diet in the prevention of diabetes (see  Ann Intern Med. 2014;160:1-10). weight loss has been repeatedly shown to decrease diabetes incidence, and there are several observational trials suggesting that rice and high glycemic index diets lead to more diabetes and that green vege diets decrease incidence. this spanish trial was a subgroup analysis of a 4-year randomized controlled trial done in several primary care centers, involving 3500 men and women aged 55-80 at high cardiovasc risk (either diabetes or at least 3 of: smoking, hypertension, high chol, low HDL, overwt/obesity, fam hx of premature cad; study randomization was not stratified by diabetes status, but diabetics were excluded from the analysis) randomly assigned to one of three diets:

–Mediterranean diet supplemented with extra-virgin olive oil (EVOO) — pts given 50 ml to use per day

–Mediterranean diet supplemented with nuts — patients given 30g/d: 15g of walnuts, 7.5g of almonds, an 7.5g of hazel nuts

–control diet, with advice on a low-fat diet


baseline mean BMI of 30, 16% smokers, 90% with htn, 54% with hyperlipidemia (mean LDL 140, HDL 56). none of the dietary interventions focused on weight loss or increasing physical activity.  results:


–reasonable adherence to diet by food questionnaires as well as by objective biomarkers in small random sample (urinary hydroxytyrosol level in EVOO group and a-linolenic acid in nut group). only minor changes in body weight, waist circumference and physical activity, and no diff between the groups

–EVOO diet, nuts diet, and control diet had, respectively 80, 92, 101 cases of new-onset diabetes, with rates of 16.0, 18.7 and 23.6 cases per 1000 person-years and multivariate-adjusted hazard ratios of 0.60 (0.43-0.85) and 0.82 (0.61-1.10) comparing EVOO and nuts diets, respectively, to controls (ie 40% and 18% reductions, with the 18% reduction not quite being statistically significant). diabetes incidence curves separate after 1.5 years and maintain this separation for up to 5.5 years.


Mediterranean diet does have significant fats (34-40% of energy, mostly derived from vege sources such as olive oil and nuts) and low dairy. includes moderate consumption of alcohol (mostly wine) and frequent use of sauces with tomato, onions, garlic, other spices. in this study dietitians gave personalized input to patients about using EVOO for cooking and dressing, weekly intake of nuts, increased fruit/vege/legume/fish, eating white meat instead of red or processed meat, and avoiding butter, fast foods, sweets, and sugar-sweetened beverages. also reduction in alcohol other than wine (moderate amount, with meals). diet assessed q3months in individual and group visits, focusing on shopping lists, meal plans, recipes. pts in control group also with personalized advice and group sessions of same frequency and intensity as those in each of the Mediterranean diet groups. new-onset diabetes defined as fasting bs>126 or 2-hr >200 after 75g glucose oral load (ie, not A1C).


so, impressive intervention trial showing that there seems to be significant benefit from a diet based on lots of fruits/veges/low animal fats with additional EVOO and with wine thrown in, as compared to a low-fat diet. we cannot conclude from this study, however, that adding the olive oil to the Mediterranean diet is better than the Mediterranean diet alone, just that it is better than a low-fat diet. it is likely that the EVOO, which adds more monounsaturates (olive oil) and anti-oxidants (the extra-virgin preparation) are additionally beneficial.



Primary Care Corner with Geoffrey Modest MD: New 1st line antiretroviral recs

21 Jan, 14 | by EBM

DHHS panel on antiretrovirals came out with an upate on initial ART recommendations for adults and adolescents (see   these new recommendations are based on 144-week data showing that the combo pill of tenofovir/FTC/elvitegravir/cobicistat was non-inferior to atripla, and 3 studies finding that dolutegravir-based regimens (dolutegravir vs raltegravir, both with abacavir/3TC; dolutegravir plus abacavir/3TC vs atripla; dolutegravir vs ritonivir-boosted darunavir, with either abacavir/3TC or truvada) showed non-inferiority overall, and the dolutegravir regimens actually did better than darunavir and efavirenz based regimens because of more discontinuations of the latter because of adverse events. in brief, the following 4 INSTI (integrase strand transfer inhibitor)-based regimens are now recommended as first-line:


–raltegravir 400mg bid plus tenofivir 300/emtricitabine 200 (truvada) daily – this one was approved in last recommendations

–elvitegravir 550/cobicistat 150/ tenofovir 300/emtricitabine 200 daily (one pill)

–dolutegravir 50, abacavir 600/ lamivudine 300 daily (only if HLA B*5701 negative) –(2pills)

–dolutegravir 50 plus truvada once daily –(2pills)


these regimens are in addition to prior approved regimens of tenofivir/ FTC (truvada) in combination with either efavirenz, atazanavir/ritonivir, or darunavir/ritonivir


dolutegravir can be given once daily with or without food. really well-tolerated, with only adverse reactions of moderate to severe intensity being insomnia (3%) and headache (2%).  but, it does decrease tubular secretion of creatinine without affecting glomerular function, so can see increase of creat by 0.11 on average, but no drug-related renal adverse events (ie, the increase in creat is not clinically significant.  my only caution is that tenofovir is pretty nephrotoxic, so important to follow renal function to make sure creatinine plateaus out and doesn’t keep increasing.  also, several of the meds (eg, tenofovir, FTC, 3TC) need renal dosing if CrCl<50.  dolutegravir (as opposed to raltegravir) has low incidence of drug resistance. seems like a really good drug, but (not to shock you too much) ain’t cheap: $1400/month.



Primary Care Corner with Dr. Geoffrey Modest: Mammography, US Task Force on BRCA testing

2 Jan, 14 | by EBM

1. The US preventive service taskforce (USPSTF) just released their guidelines for BRCA testing (see in brief, they suggest:


–estimated prevalence of BRCA1 or 2  is 0.2% to 0.3% (ie, 1 in 300-500 women) in the general population of women, 6.0% in women with cancer onset before age 40 years, and 2.1% in the general population of Ashkenazi Jewish women. In a meta-analysis of studies in which recruitment was based on family history of breast or ovarian cancer, BRCA1 mutation prevalence was 13.6%, BRCA2 mutation prevalence was 7.9%, and prevalence of either mutation was 19.8%.

–A woman’s risk for breast cancer increases to 45% to 65% by age 70 years if there are clinically significant mutations in either BRCA gene. Mutations in the BRCA1 gene increase ovarian cancer risk to 39% by age 70 years, and BRCA2 mutations increase ovarian cancer risk to 10% to 17% by age 70 years.

–USPSTF recommends that primary care providers screen women who have family members with breast, ovarian, tubal, or peritoneal cancer with 1 of several screening tools designed to identify a family history that may be associated with an increased risk for potentially harmful mutations in breast cancer susceptibility genes (BRCA1 or BRCA2). Women with positive screening results should receive genetic counseling and, if indicated after counseling, BRCA testing. (B recommendation).  one of the simplest tools is below (they list several in their guidelines, though note that the Gail model is not designed for BRCA testing/risk assessment):


Did any of your first-degree relatives have breast or ovarian cancer?
Did any of your relatives have bilateral breast cancer?
Did any man in your family have breast cancer?
Did any woman in your family have breast and ovarian cancer?
Did any woman in your family have breast cancer before age 50 y?
Do you have 2 or more relatives with breast and/orovarian cancer?
Do you have 2 or more relatives with breast and/or bowel cancer?

One positive response initiates referral.

–The USPSTF recommends against routine genetic counseling or BRCA testing for women whose family history is not associated with an increased risk for potentially harmful mutations in the BRCA1or BRCA2 genes. (D recommendation)


2. USPSTF also recently came out with recommendations about medication use to reduce risk of development of breast cancer in high risk women (see here are there conclusions:


— adequate evidence exists that treatment with tamoxifen or raloxifene can significantly reduce the relative risk (RR) for invasive ER-positive breast cancer in postmenopausal women who are at increased risk for breast cancer

–tamoxifen and raloxifene, in a systematic review, reduced the incidence of invasive breast cancer by 7 to 9 events per 1000 women over 5 years; tamoxifen reduced breast cancer incidence more than raloxifene. Tamoxifen also reduces the incidence of invasive breast cancer in premenopausal women who are at increased risk.

–women with an estimated 5-year risk of 3% or greater (basically using the Gail model) should be considered for treatment.  see their figures 1-4, which display benefit-risk of using tamoxifen vs raloxifene  in 4 subgroups at different risk levels — nonhispanic white women with and without uterus, and similarly for black women.

–the benefits of tamoxifen and raloxifene for breast cancer risk reduction are no greater than small in women who are not at increased risk for the disease. (a bit convoluted: basically, not use meds if woman not at increased risk)

–In addition to breast cancer risk reduction, the USPSTF found adequate evidence that tamoxifen and raloxifene reduce the risk for nonvertebral or vertebral fractures, respectively, in postmenopausal women.

–BUT, tamoxifen and raloxifene increase risk for venous thromboembolic events (VTEs) by 4 to 7 events per 1000 women over 5 years and that tamoxifen increases risk more than raloxifene. the potential harms from thromboembolic events are small to moderate, with increased potential for harms in older women.

— tamoxifen but not raloxifene increases risk for endometrial cancer (4 more cases per 1000 women — and i might add that these are typically more aggressive cancers). Potential harms from tamoxifen-related endometrial cancer are small to moderate and depend on hysterectomy status and age. The potential risks for tamoxifen-related harms are higher in women older than 50 years and in women with a uterus. Tamoxifen may also increase the incidence of cataracts.

–Vasomotor symptoms (hot flashes) are a common adverse effect of both medications that is not typically classified as serious, but these symptoms may affect a patient’s quality of life and willingness to use or adhere to these medications.

–so, their conclusions are that there is moderate certainty that there is a moderate net benefit from use of tamoxifen and raloxifene to reduce the incidence of invasive breast cancer in women who are at increased risk for the disease, and with moderate certainty that the potential harms of tamoxifen and raloxifene outweigh the potential benefits for breast cancer risk reduction in women who are not at increased risk for the disease.



3. an editorial in the NY times this week by H. Gilbert Welch ( highlighted several issues about mammography, with the following statistics.  for one thousand 50yo women screened yearly for 10 years (with range from low to high estimates, based on studies):

–490-670 will have a false positive test

–3-14 will be overdiagnosed (ie, diagnosed with cancer, getting therapy, but having a “cancer” which would never have caused a problem: ie, leading to overtreatment)

–but only 0.3-3.2 will avoid a breast cancer death



–mammography is a pretty rotten screening test, with high numbers of false positives and overtreatment, and small numbers of “lives saved”. note that the estimates by Welch are for annual mammograms for ten years in women over age 50, and not annual from age 50-75 as typically recommended, or even 40-75 as is often done.

–one very intriguing comment by Welch is that perhaps much of the apparent benefit of screening may be subsumed by the improvement in treatments (ie, does screening continue to be beneficial if treatment is much more effective and better tolerated than before? perhaps screening really adds nothing now since the new and earlier treatments are more effective? remember that the above benefits for treatment in Gilbert’s article are based on old studies prior to current therapies).  i would add here that more aggressive screening of very high risk patients (eg BRCA’s, as above) and other high risk as determined by Gail model (using prophylactic tamoxifen/raloxifene as above) would also likely undercut the benefit of mammography further.

–BUT, mammography is what we have now, and above considerations (which i think are important) are not adequately studied to result in a shift in current screening recommendations. the only change i have made in my practice as a result of recent studies is to offer women at age 50 the option of screening every 2 years, which (per the annals of intl medicine article i sent out several months ago) is associated with fewer false positives and little impact on true positives. i also do screening mammograms on women from 40-50 if they so desire, though i indicate that screening is quite controversial (similar to my approach on PSA screening — shared decision-making).

–and, unfortunately, none of these guidelines focus on the most important issue: breast cancer results from an accumulation of pathologic mutations over decades, and that the likelihood is quite high (in my opinion) that industrial and environmental toxins play an important role here. given the huge numbers of untested toxins currently used, and given the large number introduced into industry yearly that are never tested for potential carcinogenesis, probably the single most important preventative strategy is to focus on widespread testing of chemicals used, which often make it into the environment or the food chain — a proposal which is consistently rebuffed by industry and their governmental lobbies.



Primary Care Corner with Dr. Geoffrey Modest: Time to screen for lung cancer with CT scanning?

2 Jan, 14 | by EBM

News on New Years Eve Day:

Directly from USPSTF – here is the URL for their formal position:


this is from the Boston Globe:

“Current cigarette smokers ages 55 to 80 who have smoked the equivalent of a pack a day for 30 years, or people who had those same smoking habits within the past 15 years, should be screened, advised the US Preventive Services Task Force, a group created by Congress. Under the federal health law, insurance companies will have to begin covering the $300 to $400 cost of the screening by the end of 2014.”

Here for easy reference is my blog post from 6 months ago on the subject, with review of the most important trial (National Lung Screening Trial):

Recent recommendations by the Am College of Chest Physicians to perform low-dose CT screening of smokers (see DOI: 10.1378/chest.12-2377). these recommendations parallel the interim recommendations of the American Lung Association from 2012

Baseline: lung cancer is common and has generally poor prognosis (esp with lesions greater than stage 1), causing as many deaths as the combo of the next four cancers. one thing to keep in mind is that there are newer therapies that work better than the old ones — targeted to the specific tumor-associated genetic mutations engendered by the cancer (ie, possible that these treatments could change the risk/benefit analysis of screening in the future).  of note, the arena smoking-related morbidity is changing: tobacco companies have seen shrinking local markets (showing that sometimes after decades of obvious connection with lung cancer/persistent denial by the corporations, public health initiatives may work….); as a result,  there has been huge-scale exporting (“dumping”) of cigarettes to developing nations, with likely huge increases in tobacco-related morbidity and mortality in the near future.

Cancer prevention: attempts to prevent cancer in smokers mostly with different antioxidants or anti-inflammatories (eg b-carotene, aspirin, selenium, inhaled steroids, vitamine E, retinoids) have not panned out and are not recommended. preventing smoking initiation is the clearest prevention (though 15% of lung cancers are not smoking related. we do know, however,  from many epidemiologic studies over the decades that cancer risk geometrically increases with multiple insults, including air pollution/environmental exposures and occupational exposures in addition to smoking). for those who smoke,  smoking cessation clearly helps!, with about a 15 year lag to reducing the lung cancer risk to near non-smoker levels (unlike the heart disease risk, which decreases dramatically within 6 months of smoking cessation).

screening methods: old studies have not shown clinical benefit with either CXR of sputum cytology screening.

–low-dose CT screening (LDCT). lots of nodules identified (in 10-50% of smokers — for example, the National Lung Screening Trial Research Team (NSLT)—(see (doi 10.1056/NEJMoa1102873) –screened 27K high risk patients with LDCT and 27K with CXR yearly for 3 years and followed another 3.5 yrs, and found 25% with positive screen on LDCT and 7% with CXR, finding 645 cases/100K person-yrs with LDCT and 572/100K person-yrs with CXR –13% more. most notably, there were 247 lung cancer deaths/100K person-yrs with LDCT and 309 lung cancer deaths/100K person-yrs with CXR, a significant 20% decrease (though not very large absolute numbers – difference of only 62 deaths/100K person-yrs…), and all-cause mortality decreased 7%. the LDCT pickup of cancer was similar each of the 3 years (suggesting that it would be useful to continue screening annually). but, very large number of false positives (>95% of positives were false ones). the vast majority of those with abnormal screens had follow-up radiologic procedures, a small  minority with invasive testing (1.2% of pts not found to have cancer had a biopsy or bronchoscopy).  BUT, given the high number of abnormal screens, the “low-dose” radiation did not remain so low. the CT delivered 1.5 mSv of radiation (vs 8 mSv for regular chest CT). because of the large number of positive LDCT who then received follow up chest CT or PET CT,  the average dose overall for the LDCT cohort was actually 8mSv. the rough calculation is that this degree of radiation exposure (mostly based on atomic bomb and some medical imaging studies) would create one cancer death per 2500 people screened.

–the recommendation:  for smokers and former smokers aged 55-74 who have smoked >30 pack-yrs and either continue smoking or have stopped within the past 15 years should be offered annual LDCT, if comprehensive care can be provided as in the NLST trial.

so, this recommendation, at this point, is by pulmonary specialist organizations, which may have some self-interest (organizationally, or by the individuals involved in crafting the recommendations) to be aggressive (eg, as with the american urology assn and PSA screening).  we may want to wait for a more neutral group (eg USPSTF, though i suspect they will follow suit, given that the NLST is a well-done study). my fundamental concern is that at the same time we are getting recommendations about expensive, intensive, high-tech screening for a largely preventable cancer (and with a significant but low difference in absolute death rates), we in the trenches are getting less and less support for programs to prevent or stop smoking (cutbacks in health educators, varying and variable insurance-based support for smoking cessation devices).  In addition, i am very concerned about the additional radiation exposure (will also resend some of my previous emails about risks of radiation exposure).

–i spoke with radiology at BMC (in fact, one of the participants involved in designing and analyzing NLST) who said:

–not ready to do screening yet. need to develop protocols. but will be a low dose CT, probably no more than the 2mSv (as point of reference: CXR is 5-10x less than that).   more later….


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