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 August, 2015

Primary Care Corner with Geoffrey Modest MD: New Guidelines on CVD Prevention in Diabetics

31 Aug, 15 | by EBM

By Dr. Geoffrey Modest

The Am Heart Assn and Am Diabetes Assn just published joint guidelines regarding the prevention of cardiovascular disease, by far the greatest killer in diabetics (see dm prev CVD ada2015 in dropbox, or DOI: 10.2337/dci15-0012). Given the sometimes discordant recommendations of the different organizations in the past, this is a positive approach. Details:

Lifestyle management of diabetes

  • Physical activity: studies confirm that exercise is associated with weight loss, increased HDL levels, and decreased A1c. One large study (Look AHEAD) did not find fewer cardiovasc events with exercise, but the exercising group did take fewer cardioprotective drugs (i.e., exercise seemed to decrease the use of meds)
  • Nutrition: the Mediterranean diet in a Spanish study found a 30% decrease in CVD events (see for more detailed review: ). Some but not all data support a low glycemic index diet for improving A1C levels and lowering triglycerides, which is commonly elevated in diabetics esp if not in good control [other approaches to lowering TG levels include decreasing alcohol intake, exercise, losing weight and decreasing saturated and trans fats (which lower HDL and raise TG) and substituting monounsats and polyunsats].
  • Weight management: diet and exercise, as above, but much more successful in structured programs with monitoring the lifestyle changes, calorie reduction (including through meal replacements as done in studies), counseling, patient self-monitoring with frequent weight assessments. The studies finding the most weight loss tended to be those with the most intensive interventions (including motivational interviewing, intensive counseling, goal setting/contracts, refresher courses, incentives to those who lose the most weight…). One issue mentioned is whether the weight gain associated with smoking cessation might lead to increased CVD mortality: an observational study found that even the mean weight increase of 3.6 kg had better outcomes than continued smoking.
  • Pharmacological therapy for weight loss (I don’t use these meds, but will provide their review): the Obesity Society suggests that meds are indicated in those with BMI>30 or if 25-30 with medical co-morbidities. One should always promote the lifestyle changes, though these are often hard for patients to maintain long-term, and also to avoid drugs that increase weight, such as several antiepileptics, diabetic drugs (e.g. thiazolidinediones, insulin, glinides, and sulfonylureas cause weight gain; metformin and GLP-1 agonists produce weight loss on the order of 5%), many psych meds. Only 3 drugs are approved for long-term (orlistat, lorcaserin, and extended-release topiramate/phentermine), the rest (including regular phentermine alone) are mostly approved for <12 weeks only.
  • Surgery (mostly now called “metabolic” instead of “bariatric” surgery since there are pretty profound metabolic improvements even before large weight changes): the guidelines vary some (International Diabetes Federation in 2011 recommended surgery for BMI >30 if diabetic patients not able to reach diabetic goals with meds, esp if comorbidities present), but the new AHA/ACC/Obesity Society guidelines stick to the BMI 35 cutoff for those with obesity-related comorbidities such as diabetes. The best long-term data come from the Swedish Obese Subjects (SOS) study at 15 years: weight loss as % of total body weight was 27% for gastric bypass, 18% for vertical-band gastroplasty and 13% of gastric banding (will append my blog for the SOS study at the bottom). The metabolic improvements with “metabolic” surgery are pretty profound: 78% have resolution of diabetes (though data are from meta-analysis of short-term studies, though SOS did find 72% remission at 2 years and 36% at 10 years). Also 78% risk reduction in development of diabetes.  Also 63% improvement in hypertension, 65% in hyperlipidemia. Some retrospective data show that nephropathy may be reversed with surgery. And most studies find decrease in CVD as well as all-cause mortality. Complications of the surgery are uncommon, other than nutrient deficiencies.
  • Aspirin: the data are pretty mixed. basically, there “may be” a modest effect of aspirin in diabetic patients, on the order of a 9% risk reduction which in a meta-analysis was statistically non-significant, but with a >2-fold increase in GI bleeding, leading to the tepid recommendation: low-dose aspirin (75-162 mg) is “reasonable” in those with 10-year CVD risk >10% and no increased risk of bleeding, and (per “expert opinion”) in those with risk of 5-10% risk [note: several years ago this was a strong recommendation, which was down-graded by ADA when the POPADAD trial came out (diabetic patients with documented PAD), finding a trend to less cardioprotection with aspirin (see org/10.1136/bmj.a1840)]
  • Target A1C: <=7% for most patients (to reduce microvascular complications), <6.5% in younger people without CVD and shorter duration diabetes, 8% or even slightly higher in those with history of severe hypoglycemia, limited life expectancy, cognitive impairment repeated counseling.
  • Diabetic meds: metformin is still king.
  • Hypertension: the JNC8 goals are supported: <140/90 for most individuals. Lower targets may be appropriate for some individuals “although the guidelines have not yet been formally updated to incorporate this new information” and seem to include younger patients who can achieve systolic <130 without undue treatment burden. They do note that the 2015 ADA goal is currently <130/80 if this goal can be achieved safely. ACE/ARB should be used.
  • Lipids: the efficacy of statins is similar in diabetics and nondiabetics, with a 39 mg/dl (1 mmol/L) decrease in LDL being associated with a 21% decrease in major vascular events and a 9% decrease in total mortality, and statins work similarly independent on the baseline LDL. Recommendations: check the lipid level at least annually. Continue to emphasize lifestyle changes. These guidelines are similar to the ADA guidelines for the past few years: all diabetics between 40-75 yo and with LDL from 70-189 should be on a moderate-intensity statin. Those at higher risk should be on a high-intensity statin. If <40 or >75  yo, individualize therapy.
  • Screening for renal and cardiovasc complications: note that proteinuria and/or decreased GFR are associated with increased cardiovasc and total mortality. So, check annual microalbuminuria and creatinine/eGFR. In terms of cardiovascular screening, they feel that EKGs are reasonable (given that up to 20% of diabetics have abnormalities, those with abnormalities have a higher all-cause mortality, so the EKG might help in risk stratification — which, to me, means that those with abnormal EKGs we should be more aggressive in risk factor reduction, including more aggressive lipid lowering. It is unclear that there is benefit to pursuing EKG changes with more invasive testing/treatment, since there are no studies showing improved outcomes in asymptomatic diabetics with EKG changes. The guidelines do consider it “reasonable” to measure ABIs in asymptomatic diabetics (a marker of generalized atherosclerosis, which would lead to more aggressive risk factor modification), coronary artery calcium scoring by CT scanning (which is more often positive in diabetics, though a “0” score effectively precludes cardiovascular events over 5 years). They also feel it is reasonable to do stress myocardial imaging in those with the highest CAC scores (>400), given the high rate of cardiovascular events (though no RCTs on this)


  • The A1C goals they suggest are mostly reasonable, as I argue in the blog However, I do have several patients where it is dangerous (I think) to lower the A1C even to 10% — they are cognitively limited, unable to stick to diet despite many interventions including home care and senior day programs, are on very high dose insulin, metformin and exanitide, and have blood sugars in the 45 to >500 range with A1c in the 12% range. For them, I cannot just ramp up the insulin (real risk of hypoglycemia, which by itself confers a risk of ischemic and other cardiac events, including increases in blood pressure, heart rate, myocardial contractility, arrhythmias, ischemic EKG changes, endothelial dysfunction, platelet reactivity, inflammatory mediators…) and, bottom line, I need to treat the patient and not the A1C
  • The new meds: I am really hesitant to use new meds just based on lowering A1C levels, a surrogate marker. We have seen too many drugs (e.g. rosiglitazone) which by my reading of the literature is a harmful drug. And the new DPP-4 inhibitors seem to be clinically lackluster (see ). The FDA does allow diabetes meds to be marketed just for lowering A1C levels, though it is good that they now request evidence that there is no increase in cardiovascular risk by the meds​. But, bottom line: I really need to see pretty solid evidence that a new drug does more than just lower A1c levels (and doesn’t create more problems) before I will use them. Of note, the PROactive trial of pioglitazone, which I do use infrequently, did find a 16% decrease in their secondary outcome of MI, stroke and cardiovasc mortality.
  • Hypertension: I still use a lower target for blood pressure control in diabetics (e.g. 130/80) but am very concerned in the elderly or those with longstanding diabetes in terms of autonomic dysfunction and orthostatic hypotension (which I check for regularly in the office). And I am much less aggressive in blood pressure lowering if I really need to flog the patients with lots of meds to do so. For a fuller review including a large meta-analysis of blood pressures achieved and outcomes, as well as a critique of JNC 8, see and​  . Though I do preferentially use ACE/ARB in diabetics (they may, in some but not all studies, prevent the development of diabetes), I am not aware of any good data showing that this is beneficial in those without evidence of any renal disease.
  • Screening and lipids: I am pretty aggressive in lipid management of my diabetics (80% of diabetics die from cardiovasc disease). I still adhere to the treat-to-goal philosophy, so in my diabetics who have any other significant risk factor, I use only high potency statins (except in rare cases where the lipids are already really good, then I still use lower dose statins, since trials have found that lowering LDL from 96 to 70 is as useful as lowering from 130 to 100. But I do have a few patients with baseline LDLs under 100 and given atorvastatin 10-20 mg brings it down to 50). But my target overall is to try to get below 70 if possible, with escalating the statins to rosuvastatin 40 if needed. See a recent blog for more info on CAC scoring and my concerns about the 2013 AHA/ACC guidelines:


Primary Care Corner with Geoffrey Modest MD: Regulation of Medical Devices

31 Aug, 15 | by EBM

By Dr. Geoffrey Modest

If you thought that the FDA was approving too many drugs without adequate testing, just look at the approval process for medical devices!!! (see JAMA. 2015;314(6):604-612). The process is a tad complex. Studies are designated as “pivotal” if they were used by the FDA for premarket evaluation of safety and effectiveness as the primary basis for approval; others done premarketing were considered “non-pivotal”. Premarket approval is the primary way that novel devices are approved — those devices which support/sustain life, prevent illness, or prevent unreasonable risk to the patient. The overall process also allows for post-market approval studies, including for changes made to the devices through supplemental applications. The current study assessed this overall process, given the increasing concerns that high-risk devices make it into the market without adequate testing: the FDA has indeed been accepting more “flexible” premarket evidence for new devices in order to expedite patient access. The FDA reserved the potential to require postmarketing data collection to help assure that the devices are safe and effective, looking at the “total product life cycle” to try to fill this gap.


  • They looked at 28 high-risk therapeutic devices approved by the FDA through the premarket approval pathway between 2010-2011. Then in October 2014, they looked at all the studies in (studies on devices, per FDA regulations in 2007, are required to be logged into, as well as any others they could find (including the manufacturer websites), assessing the study status of pre-approval studies, FDA-required post-approval ones, and manufacturer/investigator-initiated studies.


  • 286 clinical studies were identified for high-risk therapeutic medical devices:
    • 82 premarket (28.7%):  30 were pivotal and 52 were nonpivotal.
    • 204 postmarket: 33 were required post-approval studies by the FDA (11.5% of the studies done), 171 were manufacturer/investigator-initiated studies (59.8%).
  • Study completion rates:
    • Premarket: 44 (84.6%) nonpivotal and 30 (100%) of pivotal were reported as completed. For these approved devices, the median total number of completed studies for pivotal and nonpivotal studies was 1, the number of patients was 241 for pivotal and 65 for nonpivotal, and the median duration of the longest follow-up for pivotal studies was 3.0 months. (1.5 months for nonimplantable devices and 12.0 months for implantable). Of note, 63.3% of the studies were open-label, only 20% were double-blind, and 20.5% relied on surrogate markers of disease.
    • Postmarket: 6 (18.2%) of the 33 FDA-required studies and 20 (11.7%) of manufacturer/investigator-initiated studies were completed, with 23 (69.7%) and 130 (76.0%) respectively listed as “on-going” 3+ years later.  Of the completed postmarketing studies, the median longest follow-up was 9 months.
    • In terms of funding, all the pivotal studies premarketing were done by industry, as were 50% of the postmarketing studies.

So, bottom line:

  • Medical devices are pretty marginally regulated and based on pretty unimpressive studies: for the FDA-required pivotal studies (which directly measure efficacy and safety), on average only one study was done (and 80% of the studies were not double-blind), with only 241 people, and a paltry 3 months of follow-up, and often used only surrogate markers.  Of those that the FDA required post-market evaluation (often required given the laxity of the premarket evaluation process), the studies were similarly small, unblinded, limited to 1 year, largely focused on surrogate endpoints and without active comparators, and fewer than 1 in 5 actually had data completed within 3 years (though the FDA has not imposed any penalties on the manufacturers). And, per FDA rules, any modification of devices requires very minimalist data for approval (e.g., allowing just surrogate markers). Of concern, several other studies find that clinicians overall do rapidly adopt the new technologies soon after they are introduced into the market.
  • The blog published 8/27/15 on implantable defibrillators brings out another aspect of the inadequacy of regulation: ICD approval was based on studies in younger people who had stable angina in an outpatient setting, but usage of ICDs has dramatically morphed to include older patients in an acute inpatient setting. And this large observational study in fact found that there is NO benefit for this group, despite noting that one-third of ICDs are currently implanted in Medicare recipients in the acute setting, with the attendant very large costs and attributable adverse events.
  • And, I suspect that much of the continuing disempowerment of the FDA and federal regulation in general is related to the increasing trend of popular governmental distrust and simultaneous embracing of privatization. Of current concern,  there is a 2016 spending bill before the US Congress to defund the Agency for Healthcare Research and Quality, AHRQ  (see ), where Richard Kronick, the head of AHRQ notes “the mission of the Agency for Healthcare Research and Quality is to produce evidence to make health care safer, higher quality, more affordable, accessible, and equitable.”  This mission is different from and compliments that of the NIH to promote basic biomedical research. AHRQ looks at the gap between current medical practice and recommended evidence-based care: how medical discoveries are incorporated into medical care delivery. This attempt at defunding AHRQ, needless to say/write, is very disturbing to those of us currently practicing medicine and the patients we serve, given the increasing evidence of exaggerated and sometimes falsified drug/device company promotions: unfortunate complications of medical devices (e.g., mesh used in bladder prolapse surgeries subsequently moving and penetrating nearby organs) as well as several medication issues (e.g., Merck withholding publication of negative studies of gabapentin..), also see several recent blogs: , a blog on the diabetic med sitagliptin, highlighting the lack of clear clinical benefit , one of several I have sent out on dabagratan (some prior to the BMJ on-line blogs), though this one summarizes several issues, including the fact that the drug company hid their own internal evidence showing that drug levels should be monitored (they hid this because their promotion was largely that dabagratan was superior to warfarin since there was no need to monitor INRs/drug levels…) presents a study showing a large number of drug company studies listed on​ did not have final results on the website, and they were are not published in medical journals.​

Primary Care Corner with Geoffrey Modest MD: Implantable Cardioverter Defibrillators in the Hospitalized Elderly

27 Aug, 15 | by EBM

By Dr. Geoffrey Modest

There was a pretty striking analysis of large numbers of elderly patients inappropriately receiving implantable cardioverter defibrillators (ICDs) during acute hospitalizations (see doi: 10.1136/bmj.h3529​)


  • The US implants more ICDs than any other country: 133,262 implants in 2009: 434 new implants/1M people, 1.5x higher than the second largest implanter
  • The age of implantation is slowly increasing, with average now of 74 yo
  • The major trials finding ICD efficacy were in outpatients with stable heart failure, with a mean age of 60 in the SCD-HeFT and 64 in MADIT II trials (in this latter study, there was no difference if symptomatic NYHA class 2-3 symptomatic or asymptomatic). The 23-31% survival benefit in these studies became apparent after 1-1.5 years.
  • But 1/3 of older Medicare beneficiaries have ICDs implanted during hospital admissions for heart failure or other acute co-morbidities
  • The current indications for AICD use for primary prevention include those at high risk of life-threatening ventricular tach or fib despite optimal med therapy (b-blocker, ACE-I), such as those with symptomatic cardiomyopathy (NYHA class 2-3) and LVEF (left ventricular ejection fraction) <35%, or those 40 days post-MI who are asymptomatic with LVEF <30%
  • Not much data in elderly: a substudy of MADIT-II found that those 121 people randomized to an ICD who were >75yo had only a non-significant mortality benefit.


  • This was a retrospective cohort study of 23,111 Medicare recipients who had a history of heart failure, were hospitalized with an acute condition, and were considered eligible for ICD therapy for primary prevention, e.g. with EF <35%. 5258 received ICD and 17,853 did not. Over 90% had the index admissions for heart failure. Follow-up 2.8 years. 53% died.
  • Mean age 80 in those not getting ICD and 75.5 in those getting one. Other differences: higher likelihood of ICD if male, lower LVEF (25% vs 29%), less psych comorbidites, more diabetes, ischemic heart failure, stroke.


  • Crude survival curves showed improvement in survival in the first few months only, no difference thereafter (see figure at the bottom). Matching patients with an ICD vs not by high dimension propensity scoring revealed no statistically significant benefit from ICDs.
  • Subgroup analysis: no statistical difference in cardiac mortality by history of non-recent MI (though there was a 37% lower total mortality). No diff in group with LBBB or by BNP levels. Those >81 yo did do better (RR 0.78, 0.65-0.93), though only 12% received an ICD

So, a few points:

  • A large number of ICDs are put in during acute hospitalizations, which had not been studied. Hospitalized patients are pretty different from stable outpatients: they may be more likely to die from heart failure itself (and ICDs don’t do much for that) instead of arrhythmia. This finding may be similar to the studies of patients admitted with MIs and getting early ICD placement: there was no benefit as compared to waiting 30-40 days and there were more non-sudden cardiac deaths in those getting early ICDs, those least likely to benefit from an ICD. Hence the recommendation to wait 40 days.
  • The benefit for this Medicare population was entirely within the first few months after ICD implantation (as per graph below). This result differs strikingly from the VA and MADIT-2 studies showing benefit only after 1-1.5 years, suggesting that in the Medicare study above there was a significant selection bias to ICDs in healthier patients (i.e., They did better in the next 4 months because they were healthier and therefore more likely to get an ICD). This also could explain the improved response in those >80 yo, where only 12% got the ICDs.
  • As an aside, other data show that women get ICDs less frequently than men, but there are other studies finding that they have lower risk of sudden cardiac death and are more likely to have complications for ICD implantation.
  • This study brings up a few issues: the most striking to me is the use of ICDs in acutely hospitalized patients, mostly hospitalized for heart failure, with no data to support that approach, and the observational data from this quite large Medicare study now going against it. ICDs clearly subject patients to adverse events and the system to very high costs (as another aside, I did have an 80+ year old patient who had a very large MI and the got an ICD, who for many years thereafter lived in fear of the ICD shocking him and therefore limited his activities significantly). The second big issue is the migration of indications from well-documented (those in their mid-60s) to much older individuals, based only on observational data, and not a lot of that.

implantable cardioverter defbrillators

Primary Care Corner with Geoffrey Modest MD: Oseltamivir for Influenza

27 Aug, 15 | by EBM

By Dr. Geoffrey Modest

As we approach flu season again, there was another article looking at the effectiveness of oseltamivir (see DOI: 10.1093/infdis/jiv058​). This was an observational community-based study in Hong Kong from 2008-2013.



  • This was an observational study of 582 patients with laboratory-confirmed influenza (121 children<5yo, 250 were >18yo– I’m not sure of the real numbers since what they stated in the text differed from the table). Baseline characteristics were similar, though more of the treated ones had myalgias and fewer got antipyretics. 223 were treated with oseltamivir (185 with influenza A and 38 with influenza B), and 359 did not get antivirals (258 with influenza A and 101 with B)


  • For those who took oseltamivir within 24 hours of onset of symptoms, the median duration of “all self-reported symptoms” until complete resolution of symptoms was reduced by 56%, including resolution of fever and respiratory symptoms (all with p<0.01). There was no significant oseltamivireffect if the onset of symptoms was >24 hours previously. On reviewing the graphs of symptom reduction, there was a 10% difference after 1 day, 20% after 2 days and 35% after 3 days (i.e., the decrease in symptoms happened pretty quickly). Symptoms were only assessed once a day.
  • Those on oseltamivir had no significant decreases in either viral shedding or in transmission of influenza to household contacts, independent of duration of symptoms in the index patient.

So, these results differed a bit from the BMJ meta-analysis from 4 months ago which really found minimal efficacy (e.g., decrease in time to first alleviation of symptoms from 7 days to 6.3 days). Of note, in the current study, there was no change in viral shedding or household transmission, and its effects were limited to those initiating meds within 24 hours of the onset of symptoms. This was an observational study and is therefore not conclusive​ (we know that those getting meds were a little different than those who did not, and, most importantly, was there a placebo effect??).  But, the CDC still recommends oseltamivir (see ), especially in higher risk individuals (<2yo, >64yo, those with chronic medical conditions, pregnant women, those with BMI>40, those in chronic care facilities), given as soon as possible after onset of symptoms, but especially if symptoms  <48 hours (can be longer with severe symptoms or hospitalization). Bottom line for me: I continue to be underwhelmed by these meds, given the minimal efficacy in the large meta-analysis and high rate of adverse effects. And, of course, the effectiveness of these drugs in the future will be affected by neuraminidase inhibitor resistance, which has played a role in the past.

For very recent flu vaccine recommendations by the CDC, see .​


Primary Care Corner with Geoffrey Modest MD: Another Downside of Lung Cancer CT Screening

26 Aug, 15 | by EBM

By Dr. Geoffrey Modest

There was a small but I think important study on the perceptions of patients who smoke and had low-dose lung cancer screening –LDCT (see lung cancer CT and continued smoking jamaintmed 2015 in Dropbox, or doi:10.1001/jamainternmed.2015.3558​). This was a VA study of patients who had in-depth semi-structure qualitative interviews about their health beliefs relating to smoking and lung cancer, done in 2014.  Details:

  • 37 patients (89% male, mean age 62, 62% white/27% black or pacific islander, mean 49 pack-years smoking, mean Fagerstrom score of 4.75 — suggesting moderate nicotine dependence) who participated in the VA pilot study on lung cancer screening in one of 151 VA medical centers were interviewed, most after they got their results. There was a detailed pretest description of the risks and benefits of screening.
  • Results:
    • 18 of the participants had abnormal screening results (9 with nodules <1 cm, 12 with nonpulmonary incidental findings)
    • Despite education on the limited benefit of screening, many participants thought that everyone undergoing screening would benefit
    • Many patients thought that screening would give a more precise estimate of their individual risk of cancer, to see “how much damage was done do their lungs” by smoking
    • Many acknowledged the need to quit smoking, but focused on “catching” cancer early
    • Many patients were reassured that by finding a nodule meant that the screening worked for them (and protected them from getting cancer), that finding a cancer meant that it could be cured and was harmless [in fact, >1/3 of the cancers detected by screening in National Lung Screening Trial were > stage 1]. Several thought they were going to get “bad news”, and the lack of an abnormality on CT led to “a lack of urgency for quitting” smoking. “It is more of a relaxing thing that there is a part of my body that I know is working and looks like it is continuing working fine for the rest of the year at least”.
    • Some participants felt that screening would itself cure cancer, that they would not need chemotherapy (and perhaps avoid these treatments that their relatives had gone through)
    • Some people thought they were the “lucky ones”, citing that they knew of people who lived to be 100 and smoked over 50 years without a problem. Some acknowledged that smoking caused cancer but that “like most smokers it’s not going to be me”

For my general critique of the lung cancer screening guidelines, see , which mostly focuses on the limitations of the National Lung Screening Trial and the related development of pretty aggressive national guidelines, and the potential of radiation-induced cancers. This current small qualitative study, done in veterans, brings up another major issue of screening: patients’ understanding and interpretation of both the screening process and the results. My major concerns from this study are:

  • It is often very difficult for patients (including some of us when we are patients…) to think objectively about cancer. It is scary, both in what it is/can do and in its treatment (it is the “crab” that extends itself relentlessly throughout your body, the emperor of all maladies). Many people are willing to take large risks of screening or treatment to decrease their risks from even not-so-aggressive cancer: “just take everything out”. This inability to sort out risks and benefits objectively and not just focus on the word “cancer” more emotionally, I think, leads to misperceptions in this study that screening protected them from cancer, small nodules are not significant but only need followup, that it was okay to continue smoking since they were one of the lucky ones who would not get cancer.
  • ​Although lung cancer is certainly a risk of smoking, many more patients die of cardiovascular disease related to smoking (the single most important correctable cause of heart disease), and almost as many die from COPD (158K lung cancer deaths/year vs 135K with COPD). The focus on CT scans to pick up lung cancer effectively de-emphasizes and diminishes even more important morbidity/mortality issues and provides a false sense of reassurance.
  • Ultimately, there are also ethical and social issues regarding the approach to serious preventable illnesses. Huge amounts of funding and resources are going into early detection of lung cancer, which had very small absolute benefit (decreased mortality by 62 deaths per 100,000 person-years of screening in NLST), instead of inexpensive (and severely underfunded), community-based prevention programs — see prior blog describing a very impressive, low-cost, community-initiated and community-based program in a poor rural Maine community, which achieved impressive cardiovascular risk factor reduction, including for smoking (​ ). Devoting 1/10 the $$ from lung cancer screening to such initiatives would likely have much more dramatic population benefits.

So, this adds to my concerns about the appropriateness of the LDCT screening. In some circumstances, medicalizing (getting CT scans) a social issue (smoking) can lead to undercutting the real social message (smoking is really bad for you, for a lot of reasons beyond lung cancer). As another example where medicalizing may undercut the important public health message, there was a recent study showing that patients put on statins tend to stop doing the lifestyle changes (diet, exercise, weight loss) since they were taking a medication which was so effective in lowering the cholesterol.​..

Primary Care Corner with Geoffrey Modest MD: Flu Vaccine Recommendations 2015-6

17 Aug, 15 | by EBM

By Dr. Geoffrey Modest

MMWR just published the influenza vaccine recommendations for 2015-16 (See​ ). In brief:

  • As per usual, vaccine is recommended for all >6 months old who do not have contraindications, and should be offered starting in October, before flu cases start. They also recommend vaccinating early even if the ideal vaccine preparation is not available, in order to avoid missed opportunities [eg, for the elderly, I would wait for the high-dose vaccine only if it will be available prior to the onset of flu season and it is clear that the patient will get the vaccine].
  • Children 6 months til 8 years still need 2 doses (second dose >4 weeks later) if they have not had prior flu vaccines (at least 2 doses before 7/1/15). Unlike last year, there is no special consideration of influenza A (H1N1).4092914530_b23270543a_o
  • The vaccine this year will contain A/California/7/2009 (H1N1), A/Switzerland/9715293/2013 (H3N2), and B/Phuket/3073/2013 (Yamagata lineage). The influenza A H3N2 and influenza B components are new this year. There will also be a quadrivalent vaccine, which includes these three plus B/Brisbane/60/2008 (Victoria lineage), as was included in the quadrivalent vaccine the last 2 years [CDC does not prefer either the trivalent or quadrivalent vaccine at this time, though this could change after the flu season begins.]
  • The live-attenuated vaccine (LAIV), previously recommended as the preferred vaccine in kids aged 2-8, is no longer recommended over the inactivated vaccine injection, since there have been inconsistent findings more recently about the LAIV efficacy (in fact the LAIV4 vaccine in 2013-14 had poor efficacy against the predominant H1N1 flu virus that year). As before, LAIV can be used from age 2-49 yo unless there are contraindications, such as: children 2-4 yo who have asthma or wheezing episode in the last 12 months; LAIV is listed as a precaution in older kids and adults with asthma who “may be at increased risk for wheezing after administration of LAIV” and in those with chronic lung, heart, renal, liver, neuro, heme or metabolic disorders (ie, safety of LAIV is not clearly established in those with these underlying medical conditions) — I would add here that I do have a couple of older kids/adolescents who are quite large, very unwilling to have an injection, and pose a risk to themselves and the injectors, and I have given LAIV, with parental consent, in these cases even if there is mild asthma (ie, protection from flu, I felt, trumped the potential of some increased risk of wheezing). LAIV should still not be given to those with documented egg allergy (not enough data),  those taking aspirin, or those immunocompromised or taking care of severely immunosuppressed persons (or should avoid contact with them for at least 7 days after LAIV, given the theoretical risk of transmission).
  • For those with egg allergy, best to use an egg-free preparation (only indicated in those at least 18 yo), though they note that “severe allergic reactions to egg-based influenza vaccines are unlikely” based on a study of 4172 egg-allergic patients who had no episodes of anaphylaxis after the regular vaccine, though the VAERS (Vaccine Adverse Event Reporting System) does have a few reported cases. Small studies (up to 282 kids aged 2-17, of which 115 had anaphylaxis to eggs in the past) reported no cases of severe allergic reaction to LAIV. Given that these are small studies, they recommend using inactivated vaccine if the recombinant non-egg produced vaccine is not available, and that it should be given by a “physician with experience in the recognition and management of severe allergic conditions”

Primary Care Corner with Geoffrey Modest MD: Spicy Foods and Mortality

13 Aug, 15 | by EBM

By Dr. Geoffrey Modest

A large epidemiologic study from 10 diverse areas in China looked at total and cause specific mortality in those eating spicy foods (see BMJ 2015;351:h3942).


  • Spices (esp capsaicin) have antibacterial activity and modify the intestinal microbiome (no clear data on exactly how the microbiome is altered, or to what effect)
  • Capsaicin has been studied in small populations or experimental conditions, finding it has anti-obesity, antioxidant, anti-inflammatory, anti-cancer, antihypertensive effects, and improves glucose homeostasis.​
  • Ingestion of hot red pepper decreases the appetite of both Asian and white people (?decreasing obesity-related morbidity/mortalitychili-61898_640)
  • A large ecological study has found higher spice consumption is associated with lower cancer incidence (these are rough-and-dirty studies which just show that in areas of high spice consumption there is less cancer, without the specifics showing individuals who eat spice have less cancer)

Details of this study:

  • 199,293 men and 288,082 women aged 30-79, without known cancer, heart disease or stroke at baseline, had an initial food questionnaire including consumption frequency of spicy foods, then followed 7.2 years (3.5M person-years of followup) between 2004-2013.
  • mean age 50, BMI 23, 5% diabetic, 35% hypertensive
  • 11,820 men and 8404 women died during this time period
  • those who consumed more spicy foods were from more rural areas (48% of low consumers, 82% high), more likely to smoke (57% vs 70% for men and 1.8 vs 3.0% for women) and drink alcohol (27% vs 47% men and 1.2 vs 3.8% women), and more frequently consumed red meat, vegetables and fruits (though not huge differences in these numbers).  Most common spice was fresh or dried chili peppers
  • Absolute mortality rates (in deaths/1000 person years):
    • ​6.1 for those eating spicy foods <1x/week
    • 4 for those eating spicy foods 1-2x/week, adjusted hazard ratio compared to those <1x/week=0.90 (0.84-.96)
    • 3 for those eating spicy foods 3-5x/week, adjusted hazard ratio compared to those <1x/week=0.86 (0.80-.92)
    • 8 for those eating spicy foods 6-7x/week, adjusted hazard ratio compared to those <1x/week=0.86 (0.82-.90) — ie, a 14% relative risk reduction
    • The above numbers adjusted for underlying hepatitis, smoking, prevalent diabetes or hypertension, red meat or vege consumption, alcohol, SES, BMI, physical activity, family history
    • Overall, women did better than men, with a 20% risk reduction, vs 10% for men
  • The relationship between eating spicy food was stronger in those not consuming alcohol (p=0.033 for interaction). somewhat stronger benefit if eating fresh vs dried chili peppers
  • Inverse relations also noted for deaths due to cancer, ischemic heart disease, and respiratory diseases

So, a couple of comments:

  • This is a large epidemiologic study, subject to the usual caveats: concerns about adequacy of consumption data (only one initial dietary assessment, and self-reported; also those with high chili consumption likely used other unmeasured spices and ingredients) and adequacy of outcome data (quality of death reports). Also, hard to separate the non-measured differences between those living in more rural than more urban settings (and the difference in spicy food consumption was pretty striking in the different settings). Though it is notable that the high spicy food consumers did have some not-so-great lifestyle parameters (smoking alcohol, red meat). And, perhaps most importantly, epidemiologic studies preclude determining that a relationship is causal
  • In this study, there was a threshold: when people consumed spicy foods 1-2x/week, they achieved almost all of the benefit
  • But my real bottom line is that I should now add fresh chilies to my dark chocolate snacks.

[Here are a couple of the blogs on chocolate: and ]​

Primary Care Corner with Geoffrey Modest MD: Post-op Delirium in Elderly

12 Aug, 15 | by EBM

By: Dr. Geoffrey Modest

A recent study found a significant relationship between inflammatory markers and the development of delirium post-operatively (see doi:10.1093/gerona/glv083). This study assessed 566 patients >70yo undergoing major noncardiac surgery from the Successful Aging after Elective Surgery Study (SAGES), and compared 12 different cytokine levels in those patients who developed delirium vs not. Funded by the National Institute on Aging.


  • Inflammatory markers have been linked to a variety of outcomes in the elderly, including frailty, dementia and cognitive decline, as well as several medical diseases (diabetes, heart disease…)
  • Delirium is associated with systemic inflammatory states (eg, infections), perhaps related to inflammation-related breakdown of blood-brain barrier, microglial activation, and neuroinflammation.

Details of this study:

  • Mean age 77, 56% female, vascular comorbidity approx 45%, General Cognitive Performance mean of 55 (this is a composite measure reflecting vulnerability to delirium, which to my reading is a scale normalized to 50 for the elderly, with a score of 50 roughly correlating to an MMSE of 26, and a score of 55 correlating with 28)
  • Cytokines were measured at 4 times in all patients: pre-op (PREOP), post-anesthesia (PACU), post-op day 2 (POD2), and post-op day 30 (POD1M)
  • This was a case-control study matched for age, sex, surgery type, baseline cognition, vascular comorbidity, apo E genotype
  • Delirium was assessed by the Confusion Assessment Method algorithm
  • There was a 3-stage approach: a discovery cohort (n=272, comparing 39 matched pairs of delirium cases vs no-delirium controls), a replication cohort (36 matched pairs from the remaining SAGES sample), and a combo of these 2 (with 75 matched pairs).
  • Results:
    • The most consistently elevated statistically significant cytokines in the different cohorts:
      • IL-2, especially on POD2
      • IL-6 (the most elevated, and dramatically so), esp POD2
      • IL-12 was decreased, mostly POD1M
      • TNF-a (tumor necrosis factor alpha) and VEGF (vascular endothelial growth factor): both peaking on POD2
      • The association was found in all the cohorts, though less impressively in the replication cohort

This study confirms some previous studies finding an association between cytokines and delirium. In other studies IL-6 has been found to be associated with age-related frailty and dementia. IL-2 regulates growth and differentiation of T cells and affects the development of the immune system and also seems to induce blood-brain barrier dysfunction. A strength of the current study is that it had a real baseline cytokine measurement (some of the studies of patients with hip fractures could have had cytokine increases based on the trauma. This study was in patients getting elective surgery). Also, the cytokines measured in the current study did not include acute phase reactants.

So, delirium is a pretty awful disease for patients and family, with increased mortality, increased hospital stays with their attendant morbidity and cost, high rates of institutional placement, and perhaps even high risk of subsequent dementia. This study provides both a potential means to identify patients at high risk of delirium early, and could ultimately lead to new medications to prevent delirium (all to be determined). I did send out a previous blog on the efficacy of melatonin-like meds in preventing delirium (see )

Primary Care Corner with Geoffrey Modest MD: Early Life Stress (in Mice), Changes in Microbiome, and Later Anxiety

11 Aug, 15 | by EBM

By: Dr. Geoffrey Modest

I heard about this on NPR and realized that I had not blogged for awhile on the microbiome…  This study basically showed that early life stress in mice leads to changes in their intestinal microbiome, which seems to be largely responsible for altered behavior later in life (see doi:10.1038/ncomms8735).  


  • There are a few human studies finding an association between traumatic childhood events and the later development of psychiatric diseases and functional bowel disorders

  • Animal studies find that maternal separation leads to long-lasting behavioral changes, gut dysfunction, hyper-responsiveness of the hypothalamic-pituitary-adrenal (HPA) axis, depression/anxiety, visceral hypersensitivity, and increased intestinal permeability in adulthood

  • Intestinal microbiota can affect host behavior: germ-free animals have altered HPA responses to stress which is reversed by colonization by commensal bacteria, germ-free mice have lower anxiety-like behaviors, and there are an array of altered changes in neural and hormonal function (eg, microbial colonization in early life can affect hippocampal function)

Details of this study:

  • Both germ-free (GF) mice (ie, no intestinal microbiome) and specific pathogen-free (SPF) mice (ie, all having the same intestinal microbiome) developed equally large increases in their serum corticosterone levels when submitted to the stress of maternal separation shortly after weaning.

  • The microbiota of those SPF mice changed a lot when subjected to the stress of maternal separation, having only 55-70% similarity to those not exposed to stress. And this “early-life dysbiosis” in their microbiome did not change as the mice reached adulthood.

  • Later anxiety levels were higher in the mice exposed to the maternal separation, but only in the SPF group. there were 3 assessments of anxiety: a step-down test (going down from an elevated platform), light preference test (amount of time in illuminated compartment) and tail suspension test (how long they were immobile).

  • The altered microbiota was transferred from the SPF mice with maternal separation into healthy GF mice, finding that these GF mice did not maintain this abnormal microbiome and did not have abnormal responses to the 3 anxiety tests. Since those mice with early trauma developed a long-lasting dysbiotic microbiome but just implanting this dysbiotic microbiome was not long-lasting and did not alter them clinically, this suggests that there are a combination of stress-induced microbiome changes in addition to host factors that affect later anxiety.

So, my interest in the microbiome is that it provides an insight mechanistically into at least one physiologic way that animals (including humans, though it is much easier to study mice) are changed by their environment. There are several of my blogs dealing with different diets, food additives (eg artificial sweeteners) and their effect on the microbiome, along with the attendant effects on heart disease, glucose intolerance etc. But I really liked the one on diabetic mice showing that those on metformin improved their glycemic profile (as expected), but there was a significant increase in the bacterium Akkermansia. In a parallel experiment, just increasing this bacterium (in the absence of metformin) also enhanced glucose tolerance and decreased adipose tissue inflammation, suggesting that an additional mechanism of action for metformin may be through its effect on the microbiome (see )  — and this blog even looks at a study in humans, finding that meat-eaters have changes in their intestinal microbiome which lead to the production of an atherogenic chemical (TMAO), which was not found in vegetarians given meat (before their gut flora could be changed).  In the current mouse study, it is impressive that subsequent anxiety is not simply a genetic/epigenetic response to a traumatic early childhood event, but that the putative mechanism for this later anxiety seems to be mediated through a combination of microbiome and host changes. This study, of course, begs for followup ones looking at the outcome of potential microbiome improvements after the initial stress insult, such as through diet, exercise, or even meds. Interesting stuff…

Primary Care Corner with Geoffrey Modest MD: Blood Pressure Variability and Heart Disease

10 Aug, 15 | by EBM

By: Dr. Geoffrey Modest

The Annals of Intl Medicine recently published an article on the relationship between visit-to-visit variability of blood pressure and coronary heart disease (CHD), stroke, heart failure and mortality (see doi:10.7326/M14-2803). This was a post hoc analysis of the large database from the ALLHAT study (Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial).  ALLHAT was an NIH and NHLBI funded study, and one of the best studies from a primary care perspective in that it really was community-based, with 623 clinical sites in the US, and included a truly ethnically diverse population (45% non-Hispanic white, 30% non-Hispanic black, 15% Hispanic white and 3% Hispanic black) to look at the effects of antihypertensives (one of its wings) in adults >55 yo with at least one CHD risk factor. Details:stestoskop-64276_640

  • 25814 patients with at least 7 visits during the study, noting the visit-to-visit variability (VVV) of blood pressure on clinical outcomes. Blood pressure was assessed in a structured and consistent way (though there were lots of sites and there might have been some differences).
  • Those without a clinical CVD (cardiovascular disease) event during the first 28 months were then followed until the end of the study, for mean of 2.7-2.9 more years with max of 5.7 years.
  • Results:
      • Total clinical events: 1194 fatal CHD or nonfatal MI, 1948 deaths, 606 strokes, 921 heart failure events
      • Higher SBP standard deviation (SD) was found in older (67.4 vs 65.7 yo) and non-Hispanic black persons (39.9 vs 25.5%), less likely to be men (47.1% vs 55.3%), less likely to take aspirin (32.1 vs 36.2), less likely to have an HDL<35 mg/dL (9.1 vs 14.3%), did have a higher mean SBP (144.2 vs 132.7 mmHg) and higher pulse pressure (64.0 vs 54.4 mmHg).
      • Those in the highest quintile also were more likely to have low medication adherence (18.1 vs 10.3%), be on more antihypertensive meds (2.0 vs 1.4), have more changes in medication classes (64.1 vs 27.3%) and be more likely to be randomized to lisinopril (30.0 vs 20.9%), and less likely to get chlorthalidone (41.2 vs 49.6%) or amlodipine (22.8 vs 29.5%)
      • After multivariate adjustment, including controlling for mean systolic blood pressure (SBP), Hazard Ratio comparing those with the highest vs lowest quintile of standard deviation of SBP (which was >=14.4 mmHg vs <6.5 mmHg):
        • Fatal CHD or nonfatal MI: HR 1.30 (1.06-1.59)
        • All-cause mortality: HR 1.58 (1.32-1.90)
        • Stroke: HR 1.46 (1.06-2.01)
        • Heart failure: HR 1.25 (0.97-1.61)
      • Higher VVV in diastolic blood pressure was also associated with CVD events and mortality
      • On review of their figures of cumulative incidence of events, overall they are splaying apart as time goes on (ie, more of an effect)


      • A pretty quick and dirty secondary analysis of a complex study, finding major differences between the quintiles of VVV. Although their subgroup analysis did find the highest hazard ratio for low medication adherence (HR around 2, and statistically significant, for both fatal CHD/nonfatal MI and for all-cause mortality), they still found a higher CVD and mortality risk after adjusting for this.
      • There is a reasonable putative mechanism for the VVV/CVD relationship: there is a higher VVV of blood pressure when there is more arterial stiffness (lower aortic distensibility), endothelial dysfunction, and (probably) subclinical inflammation.
      • There was a whole issue (pretty unusual) of the Lancet in 2010 devoted to blood pressure variability and CVD (4 articles, for example see Lancet 2010; 375: 895–905), finding that high VVV was associated with an adjusted HR for stroke in the UK-TIA aspirin trial of 4.84 (comparing highest vs lowest decile), 4.29 for the ASCOT-BPLA atenolol group and 4.39 amlodipine group, 1.78 in the European Stroke Prevention Study, and 3.35 in the Dutch TIA study (see their table 2). These are all post-hoc analyses, as is the current study, but they do provide pretty strong support.
      • There are several studies (including ALLHAT) finding an increased incidence of stroke in patients on ACE inhibitors.
      • Studies of 24-hour efficacy of antihypertensives have found that amlodipine and diuretics have the least variation during 24 hours, and b-blockers and ACE-inhibitors have much more (eg, see  Lancet 2010; 375: 867–869​). That being said, there are data suggesting that HCTZ 12.5 mg is not so great, but that chlorthalidone or HCTZ at higher dose, such as 50mg, does decrease 24 hour variability. and, it might be significant that the quintile with highest VVV in the current study were disproportionately on lisinopril. (One argument around strokes is that Lisinopril has waning effects in the wee hours of the morning, and stroke is more common at that time). The potential connection between 24-hr blood pressure variability and SBP variation in different visits is that patients may take their pills at different times, and the med with the least consistent 24-hour effect is more likely to have more VVV, and this may also be especially true when their blood pressure checked at different visits scheduled at different times of the day.
      • So, VVV in blood pressure probably predicts a less good outcome. We might use that as surrogate marker of higher risk and treat risk factors even more aggressively (eg lipids, smoking, etc, and reinforcing lifestyle issues).

Latest from Evidence-Based Medicine

Latest from EBM