Top 5: POEM's To Win Over Any Patient


I'm still alive.

Yeah, you. That single person who reads this blog with any morsel of consistency. 

I'm still alive, thanks Mom.

I have found myself giving talks more than blogging them so I apologize for the lack of content in the past 3 months. Well, absolutely no content for the past 3 months. For some reason, my talks given on outpatient COPD and hypertriglyceridemia induced pancreatitis don't translate well to blog form. But I'm back to give you my q3month blog. And it's inspiring.

Well, inspiring for myself.

I just got back from the AAFPNC in Kansas City where I was using all of my dirty college football coach tricks to recruit spectacular medical students to the Pacific Northwest. (Only a small amount of endorsement money was passed, I promise!) While at the conference, I was able to go to a few talks this year. Most of them were pretty mundane to be honest but there was one that I really loved. Dr. Steven Brown was the main speaker and he basically discussed the top POEMs (patient oriented evidence that matters) of the past year. He did it in a way that fluidly showcased not only his knowledge of the actual POEMs and studies but insight into methodology, medicine, and practicality. So with that inspiration, I give you the second installation of the "Top 5" series. Top 5 POEMs in the past year... or so.


1. Augmentin has no effect in Acute Bronchitis, except to give you diarrhea AND normal coughs last for 18 days.

I haven't got passed the first POEM and I already cheated by combining two in one. I really like this one because it's something we could probably use nearly every day in the clinic. Who hasn't had a patient coming in with a cough that "now has been 8 days!" In early 2013, there was a study [1] that took nearly 500 patients in Georgia and randomly called them to ask about their expected duration of a cough. They then compared this to observational and RCT studies found in the literature and found that the average published duration was 17.8 days while the patient median expected duration from the patient survey was 5-7 days. It's not revolutionary by any means but it's some nice data to back up your patient education visits.

"But what about the antibiotics?!"

Another difficult area of patient education as I'm sure many of you are all too familiar with. But now with even more data to help you explain why antibiotics are not a panacea. In October of last year, there was a RCT [2] with roughly 400 patients with a respiratory tract infection less than 1 week duration who were randomized to either ibuprofen (600mg TID), Augmentin (500/125mg TID) or placebo for 10 days. They had the patients use a diary card to record the duration of cough. Study found no significant difference in length of cough between ibuprofen (9 days) and Augmentin (11 days) or placebo (11 days). On top of that, 1 out of every 8 patients had a negative side effect from Augmentin, typically gastrointestinal related. 


2. JNC8 was a big deal: Older folks finally get more leeway, initial treatment is a free for all, and diabetics get cut some slack.

Family medicine physician are a different breed. We don't get too excited about new medical devices or the latest medication on the market that costs twice your mortgage. What we DO get excited about is stuff like JNC8. It's the crack cocaine of my people. JNC 8 did not disappoint either, it gave a nice, clean high. First off, they finally said we could loosen the blood pressure reigns on our older patients. Now the goal for those greater than 60 years old is 150/90 instead of 140/90 previously. We were already doing this before but now we can point to JNC8 and get some much needed back up.

      Favorite gif of this year

      Favorite gif of this year

They also gave providers more options to start with when starting an antihypertensive. Now the recommended first line in nonblack populations is thiazide, calcium channel blocker (CCB), angiotensin-converting enzyme (ACE) inhibitor or angiotensin receptor blocker (ARB). That's pretty much everything except beta-blockers. In black populations, they recommend thiazides or CCB. They also said not to use ACEI and ARBs together, likely stemming from studies in the past year showing either no benefit in mortality [4] or mixed results [5, 6]with regards to nephropathy progression.

Finally diabetics can join the rest of us at 140/90 instead of needing stricter blood pressure goals. Hurrah!


3. New anticoagulants basically the same as warfarin in atrial fibrillation.

The new anticoagulants remind me of the new kid who'd come to your school in the middle of the year. They'd have some spunky name, in this case something like dabigatran. But you know my friends call me Pradaxa. They'd be dressed different. They'd act different. You wouldn't know what to make of them. Would they become the new "cool kid"? or someone who would be sitting alone at lunch? Well the answer is "we still don't know".

A meta-analysis [7] was released this past year that pitted the new anticoagulants versus warfarin. The authors searched RCTs and pooled roughly 72,000 patients who had atrial fibrillation and received either a newer anticoagulant or warfarin. The study ended up finding that there was a small but significant improvement in all strokes (3.1% vs 3.8%, RR=0.81, NNT=148) but not difference in rates of MI or ischemic strokes. The newer agents also had a small but significantly higher rate of GI bleeds (2.6% vs 2%, NNT=129). Overall, the numbers aren't astounding and there were several concerns about the authors being affiliated with the companies and obscure methodology. So I took from this that I will treat the two classes of medication the same and will use cost and ease of use as my main branching points. 


4. Aspirin - disappointing you as I get older.


Created over 100 years ago, aspirin was the King Kong of medicine at one point. It was the pharmacological "apple a day" and was used for anything from colon cancer to being laced in gum for children after tonsillectomy (who knew they would end up bleeding more). However, as years go on aspirin has been found to be less and less effective and the very real side effect of GI bleeds has become more prominent. 

I cheated for the second time now and took a POEM from 2012. It was too good to pass up on. There was a meta-analysis [8] 2 years ago that looked at over 100,000 patients without clinically apparent cardiovascular disease. Doses of aspirin ranged from 100-500mg qday and found that the composite outcome was 3.86% in aspirin arm and 4.16% in the placebo arm. In layman's terms the difference would be 1 in 253 patients treated for roughly 7 years. Now remember, composite outcomes can be misleading and this study is no different because no individual outcomes (MI, stroke, death) were significant. When you pool insignificant results you can sometimes create a composite outcome that is indeed significant. There was also a significantly higher risk of hemorrhagic stroke and major GI bleeds in the treatment arm with a NNH of 261. 

So no more daily aspirin for healthy grandpa. If he has cardiovascular disease or diabetes then that's another discussion.


5. Mammograms - numbers to help with your fit.

Personalized numbers are all the rage these days. Whether it's the firmness of your Sleep Number bed, your sleep or steps on your Fitbit or the degree loft on your driver (my personal favorite), people love customizable numbers. I've chosen this last POEM because again, it's something we do all the time in primary care - breast cancer screening. What's better than to have some hard numbers to give your patients? We love numbers!

The authors in this study [9] from March 2014 did an analysis of the most recent data (RCT, epidemiologic) on breast cancer screening and outcomes. They specifically looked at breast cancer death avoidance, overdiagnosis, and false positives and then divided them into age categories. The numbers came out as such:

Among 1,000 40-Year Olds:

  • Breast Cancer Death Avoidance: 0.1 to 1.6
  • Overdiagnosis: 11 
  • False Positive: 510 to 690 with 60 to 80 to get biopsies.

Among 1,000 50-Year Olds:

  • Breast Cancer Death Avoidance: 0.3 to 3.2
  • Overdiagnosis: 3 to 14
  • False Positive: 490 to 670 with 70 to 100 to get biopsies.

Among 1,000 60-Year Olds:

  • Breast Cancer Death Avoidance: 0.5 to 4.9
  • Overdiagnosis: 6 to 20
  • False Positive: 390 to 520 with 50 to 70 to get biopsies.

My one beef with this study is that they looked at yearly screening. I would have loved to have seen this data parceled out with biennial screening as well. We will see where USPSTF ends up with its new recommendation as they are in the middle of a large update.

In the end, I'm still going make breast cancer screening a personalized decision for my patients. We'll talk risk and benefits and go from there. For ease of remembering you could consider simply quoting 1-3 breast cancer deaths avoided, roughly 10 over diagnosed and about 500 false positives for every 1,000 women screened in general.

Who Wants To Live Forever: Mortality Benefits in the Outpatient Setting

It's been three months since the last post. I've been sitting here for the past 5 minutes trying to come up with some witty excuse of why it's been so long but apparently I'm running low on wit today. So no excuses, let's get wit it. 

There are a couple of questions I've found you can hang your hat on in medicine: "What is the patient's clinical presentation?", "Are they stable or not stable?", and today's topic "Will this affect their mortality?" Because in the end, prolonging quality living is something that most doctors and most patients can agree upon. Also, these are great phrases to say when you have no clue what you're doing. Next time you hear me ask one of these questions it's probably safe to assume I'm mentally regrouping. In lieu of this important question we frequently ask ourselves, I've decided to rehash what exact interventions decrease mortality in the outpatient "Big 3" (Diabetes, Cardiovascular Disease (CVD), and Hypertension). The goal here is to have more of a quick lesson and a point of reference in the future and less of a narrative. So you won't be getting any long drawn out stories today (lucky you).

*Credit goes to Mark Johnson as he has given talks like this previously. 

Data from Yudkin, JS, BMJ 1993; 306:1313.



  • Smoking Cessation
    • Males: All cause mortality for smokers increased significantly from 2.3% to 19.8% per 1,000 patient years. [1]

    • Females: All cause mortality for smokers increased significantly from 1.1% to 14.6% per 1,000 patient years. [1]

    • Female relative risk of mortality in IDDM smokers = 1.4. Dose dependent response seen. (Prospective cohort for 20 years in Nurses Health Study, 2001) [2]


  • Physical Activity
    • Walking 2 hrs/wk reduced all-cause mortality in diabetics by 39% (HR 0.61 [95% CI 0.48-0.78]). Mortality rates lowest for those who walked 3-4hrs/wk (HR 0.46). [3, cited in this review paper]
    • Physically inactive men with DM2 had 1.7x increased risk of premature death compared to physically active men. [4]
    • Systematic review (12 studies) found a HR of 0.6 (CI 0.49-0.73) for total mortality between the lowest and highest fitness levels in DM1 and DM2 patients. Looking further into the study it looks as though <2 hours/wk and >9 hours/wk were the ends of the ranges. [5]


  • Blood Pressure Control
    • ACCORD trial (2008): Found NO difference in mortality (in addition to macrovascular outcomes) when comparing goal SBP <120 and <140 in diabetics. [6]
    • ADVANCE trial (2007): Only major placebo-controlled RCT. Evaluated effects of ACE/Thiazide combo medication in moderate-high (<25% 10 yr risk) to very high (>25% 10 yr risk) CV risk diabetics regardless of history of hypertension. Mean SBP was 139 in mod-high risk and 151 in very high risk. No target BP's but end result was roughly 5 point reduction with combo medication compared to placebo. Final mean BP's were 135.74 and 140/76 when both groups combined. Other medications used at discretion of physicians. Found significantly lower rate of CV and all-cause mortality.
      • However, in my opinion the study is poorly made as the results are only applicable if they are pooled. Both the placebo and treatment arms in the mod-high risk groups were under new JNC 8 goal of < 140 mmHg. Only when you combine mod-high and very high risk do you get the mean averages of 135/74 versus 140/76. It's also probably a given that very high risk patients will already have hypertension. [7]
    • HOT trial (1998): Specifically looked at DBP in individuals. Only in diabetics, the RR of CV events was significantly reduced (RR 0.49) in DBP <80 compared to <90. [8]
    • UKPDS (1999): Means of two groups was 144/82 vs. 154/87. After 8-9 years the lower BP group had 32% reduction in deaths related to diabetes. Ten year follow up study showed for each 10 mmHg reduction in SBP there was 12% reduction in diabetic complications including sudden death. Affect maxed at 120 mmHg.
      • Of course this study didn't strictly look at the now standard SBP < 140 mmHg [9]


  • Glucose controlling medications BUT Metformin decreases myocardial infarctions (MI's)
  • Statins BUT they decrease MI's.
  • ACE inhibitors BUT they decrease renal disease in those with proteinuria and may decrease CV events compared to calcium channel blockers.


  • Aspirin 
    • 2003 - No difference in mortality or CV outcomes found in subgroup analysis of RCT (1,031 patients). Patients were > 50 years old. [10]
    • 2008 - Japanese RCT (~2,500 patients) found significantly decreased fatal CV and cerebrovascular events in low-dose (0.08%) versus non-aspirin (0.8%) groups. (NNT 139) However, no difference in all-cause death or other CV outcomes. [11]
    • Low dose aspirin not associated with any increased bleeding risks compared to non-diabetic patients. [12]

Summing up the above data. You would be an evidenced based doctor making clinically significant recommendations by telling your patient to walk at least 2 hours a week and ideally up to 9 hours+, quit smoking, and to keep BP somewhere in the <140/90 range. The BP data presented above is somewhat of a moot point when considering that the JNC 8 recommendations simply make the goal <140/90 across the board. No more of the nuanced guidelines for diabetics.

Aspirin is of course less clear when considering raw data versus major guidelines. Preceptors will frequently remind me "are they on a baby aspirin?" and of course I always feel bad for not remembering. But I am less convinced now that it would make a difference. As of 2013, the ADA recommends (p. 59) aspirin in patients with 10-year risk of events >10%. This recommendation pretty much includes all males/females > 50/60 years old with diabetes. Comparing that to USPSTF's recommendation of low-dose aspirin for all males/females > 45/55 years tells me that anywhere in the 45-50 and 55-60 age range is probably fine. I'll probably continue the low-dose aspirin for now while of course weighing risks and benefits but now I'll allow myself some leeway when I forget.

Now that was simply regarding mortality. Of course, there are other clinically significant outcomes that we care about that don't address mortality such as non-fatal macrovascular events and microvascular disease that can affect quality of life (blindness, dialysis). But if it is a matter of life or death for our diabetic patients then there are only 3 tools we can use as physicians. 

Part 2: Hypertension to be continued

Get In Mah Belly 2: Fun Size Food Facts



What originally started as a "fun size food facts" blog has quickly turned into "all you can eat buffet world eating champ facts" blog. I've combined calcium and vitamin D as you really can't talk about one without talking about the other and quickly found out that I bit off more than I could chew. A lot like real life eating for me actually. Much of this might be review for you (especially SFH residents) but it's always good to have a refresher. And since calcium-vitamin D is a pretty dry topic, I have decided to post random pictures of cheese for your viewing pleasure.

If there was a single vitamin or mineral that my mom used to back her opinions on food, it'd be calcium. Every healthy thing in our fridge or pantry was a "good source of calcium", especially foods I hated (looking at you, beans). I kind of think she stole it from the boxes of cereal that would advertise that same statement. In actuality, it's not that hard to get a decent amount of calcium on a traditional American diet. Cheeseburger here. Milkshake there. Metformin for dessert and you're done. But what about more calcium or vitamin D for either the general public or those at higher risk for osteoporosis? 

The now well known landmark study that helped propel the calcium-vitamin D discussion was the Women's Health Initiative (WHI) [1]. In short, the massive study looking at women aged 50-79 years old found a small benefit in terms of bone mineralization but no significant benefit in fracture rate in those taking 1,000mg calcium and 400IU vitamin D supplementation when compared to placebo. [2] The study did have some limitations as it was looking at a general population and not specifically at higher risk patients (decrease calcium-vitamin D in diet, known low bone mineralization) in addition to being a relatively low dose of both. But overall it gave some evidence to the thought that maybe the ol' calcium-vitamin D combo wasn't as potent as once thought.

There was also another study that also came out in 2006 but this one pointed to some benefit. It was a double-blind five-year RCT that looked specifically at elderly females (>70yo). They were randomized to take calcium carbonate (600mg BID) versus placebo and found that there was indeed a significant decrease in fracture incidence (10.2% vs 15.4%; HR, 0.66; CI 0.45-0.97). But the kicker? It was only seen in those were took their supplements greater than 80% of the time. [3] Surprise - you have to take your medications for them to actually work.

That was 2006. But what did 2007 bring? More evidence for the calcium-vitamin D tag team.

A meta-analysis of 29 RCTs looking at all adults greater than age 50 found a significant decrease in both all fractures (RR 0.88, CI 0.83-0.95; p=0.0004) as well as bone-mineral density. They also note again that compliance was a big component as there was a greater fracture incidence reduction (24%) in those with high compliance (>80% of the time taken). Dose was also a big variable as those taking less than 1200mg had less of a benefit than those taking more than that dose (0.80 vs. 0.94; p=0.006). So what do we take away from this? Again, dose and compliance seem to be major players in preventing fractures in older adults. [4]

So it looks like there's some growing evidence for calcium supplementation in older, compliant adults.

Now what about vitamin D?

Again, looking back at the WHI study. We saw that the combination of calcium and vitamin D yielded no benefit in terms of fracture incidence. One thing to remember is that they studied a relatively low dose of 400IU daily. Looking back at the 2007 meta-analysis above, they found a small but significant incidence reduction in fractures when patients supplemented with 800IU or more versus those with less than that amount (RR reduction 3%, p=0.03). So it might just be a dose dependent thing?

Unfortunately, the years 2007-2010 say no.

From 2007-2010, several studies came out that showed no evidence for fracture incidence reduction in those taking only vitamin D. [5,6,7] These studies mainly looked at all adults greater than age 50 and most studies included in these pooled analysis looked at vitamin D dosages greater than or equal to 800IU. The largest and most recent of these [7], had a mean age of 69.9 years and looked at over 68,000 patients. It found no evidence for either 10 or 20 microg doses of vitamin D (aka 400 or 800IU). Bummer.

But wait! The years 2011-2012 say.... maybe?

A meta-analysis was done for the USPSTF in 2011 that looked at the potential benefits and harms of vitamin D with and without calcium supplementation [8]. This study ended up showing a significant reduction in fracture incidence for institutionalized patients (RR=0.71, CI 0.57-0.89]) but not in community-dwelling patients when supplementing with calcium-vitamin D. They also found no significant reduction in incidence for any dose of vitamin D. This was the start to the 2013 recommendation by USPSTF, which we'll see in a little bit. On the flip side, an equally large meta-analysis in 2012 showed there was actually a benefit to fracture incidence reduction in higher vitamin D dose supplementation (>800IU) in those 65 years or older. [9]

Ok, let's regroup.

All this back and forth is great and stuff but there isn't a really reliable trend in all of this data, especially for vitamin D - leads us to February 2013, when the USPSTF came out and gave no recommendation for vitamin D and calcium supplementation (Meta-analysis). Specifically, they gave an "I" recommendation for insufficient evidence to adequately weigh the harms and benefits of calcium and vitamin D supplementation. So why the sudden change from previous recommendations?

A very long story short(er) is that there just wasn't enough robust data to trump the sporadic data showing potential harms. Specifically, those harms involve kidney stones and cardiovascular disease (CVD). Initially in the WHI, they proposed that they would actually see a benefit in terms of CVD outcomes. However, they ended up finding no benefits at all and some subsequent studies in 2010 and 2011 actually showed a potential increase in CVD outcomes in those with calcium and vitamin D supplementation. [10,11] One of these two studies actually reanalyzed the WHI data set to specifically look for those women who began supplementation after the start of the trial. They found a RR of 1.24 (p=0.004) for risk of MI's. Their rationale for why the WHI showed no change in CVD outcomes was that a vast number of participants were actually taking supplements prior to the trial, thus negating and difference seen in the control and study groups.

Even more recently, there have been studies in the past 2 years looking at calcium supplementation that showed potential excess CVD risk in adults aged 35-64 as well as men aged 50-71 but not women. [12,13] Combine this with the WHI also finding a small but significant increased risk of kidney stones (HR, 1.17; CI 1.02-1.34) and you have yourself a nice smear campaign going. But of course in the midst of all these negative studies there was also a meta-analysis that found no increase CVD risk with calcium supplements and even a small benefit (RR=0.90) for fracture incidence for vitamin D. [14] It's important to remember though that there has not been a single study yet (as far as I know) that has shown an increase in all-cause mortality. The only adverse studies have been with regards to CVD events.

You can now start to see that the once clear waters of calcium appear to be quite milky murky.

With regards to adverse effects of vitamin D, there aren't many unless you're taking outrageous doses (500,000IU single dose increasing fracture risk) or for some other reason have elevated levels in your body. One study looking at several different nutrients and vitamin levels (B12, folate, iron, vitamin D) in adults aged greater than 75 found that both low (<20ng/mL) and high (>30-50 ng/mL) vitamin D levels were correlated to mortality.  [15] That really doesn't mean much to me except don't take exorbitant amount of vitamin D.

So what did I take from all of this? To me, calcium is much like fish oil in that you'll never find a negative study on it as long as it's dietary. All of the main positive fish oil studies were dietary and all of the negative calcium studies have been supplements. So overall, I'd probably take a cup of milk or a slice of Gruyere over a calcium pill. If I had a patient with increased fracture risk and no cardiovascular history, I'd probably advise them to take 1200mg calcium and 800IU of vitamin D. If the opposite were true, I'd probably advise them not to take any calcium-vitamin D supplements. I'm less convinced either way with Vitamin D alone as there isn't really a concerning side effect profile but there's also not a ton of great evidence for it with regards to fracture risk. I'd probably tell my patients I'm fine with 800IU vitamin D either way, especially if they have a history or risk of falls.

In the end, the best medicine is probably not a supplement but to tell your patient to smile and say "cheese".