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

Primary Care Corner with Geoffrey Modest MD: Chronic pain and dementia

8 Jun, 17 | by gmodest

by Dr Geoffrey Modest

​​The MMWR recently published the finding that the age-adjusted rates of Alzheimer’s deaths have increased in 41 states and the District of Columbia from 1999 to 2014 (see ).


— Alzheimer’s disease is the 6th leading cause of death United States, accounting for 3.6% of all deaths in 2014

— 93,541 Alzheimer’s deaths occurred in the US in 2014, at an age-adjusted rate of 25.4 deaths per 100K population. This represents a 54.5% increase over 1999, when the rate was 16.5 deaths per 100K (44,536 deaths).

— Most deaths occur in nursing home or long-term care facility (though this was a decrease from 67.5% in 1999, to 54.1% in 2014). The number of deaths in a hospital declined from 14.7% in 1999 to 6.6% in 2014,; the percent who died at home increased from 13.9% in 1999 to 24.9% in 2014. In addition, 6.1% died in a hospice facility in 2014.

— the specific Alzheimer’s death rate per 100K population increased among those to age 75 to 84 from 129.5 to 185.6 per 100K population, as well as in those >85 years old, from 601.3 to 1006.8 per 100K population.

— Overall, the death rates were higher among all age groups, both sexes, all races/ethnicities, urban/rural, and in almost every state. But they were particularly high for women and non-Hispanic whites

— age-adjusted rates for the 50 states and DC range from 7.0 to 29.8 per 100K in 1999, and from 10.7 to 43.6 per 100K in 2014


— this MMWR report is based on death certificates, which are not necessarily accurate and may be influenced by the prevailing culture and the US system of reporting.  In particular, the death certificate-based data derive from what the patients’ clinicians decide is the major cause of death, which in my experience is often a best-guess after the fact. and this is likely influenced by several nonclinical factors, including the acceptability of Alzheimer’s disease itself as a cause of death (which may have increased in this time period, at least partly explaining the noted increase), or even the influence of the funeral home directors. In addition, there may be very real issues with deciding if Alzheimer’s is truly the responsible condition: the definitive diagnosis is in fact histopathologic, though there are clinical definitions of reasonable accuracy, but many clinicians may be hesitant to use this label (is it truly accurate??) and avoid coding it; also, at what degree of cognitive impairment does it reach the threshold to go on a death certificate? (and that threshold is likely very different from one clinician to another)

— And, this increased incidence of Alzheimer’s also raises the question of potentially modifiable factors in the development of Alzheimer’s, which may have changed over time. For example, obesity, diet, physical activity, perhaps medications, and levels of education have been found to be associated with Alzheimer’s. See blogs 

— one profound implication of an increasing incidence of Alzheimer’s disease is the tremendous effects on others, and especially the paid and unpaid caregiver burden. And this stresses the need for additional resources to educate and train caregivers in the care of Alzheimer’s patients, as well as providing respite care and case management



another article just came out suggesting an association between persistent chronic pain and memory decline/dementia, in a longitudinal cohort of elders (see doi:10.1001/jamainternmed.2017.1622).



–10,065 community-dwelling older adults in the Health and Retirement Study, who were at least 62 years old in 2000 and answered questionnaires on pain and cognition; they were followed through 2012 with sequential questionnaires

— 60% female, median age 73 at baseline.

— 10.9% had persistent moderate-to-severe pain at baseline (n= 1120), 76% female (versus 58% female in the nonpain group), 86% white/8% black/5% Hispanic, 58% ever-smokers, 45% nonmarried/non-partnered

— comorbidities (all with differences with p<0.001): hypertension (61% in pain group, 48% comparison)/diabetes (20% pain, 14% comparison) lung disease (14% pain, 8% comparison), heart disease (37% pain, 25% comparison), stroke (13% pain, 8% comparison), depression (50% pain, 20% comparison), alcohol use (less in pain group overall, 68% reporting none versus 53% in comparison), difficulty doing more than 1 ADL (46% pain, 12% comparison)

— primary outcomes included memory score (a compilation of several cognitive tests, by the patient or proxy) and dementia probability score (a different cognitive test which estimates the chance that an individual would reach the DSM3 or 4 diagnostic criteria for dementia

–the fully adjusted model controlled for: demographic measures (age, sex, race/ethnicity), education level, current or former tobacco use, medical comorbidities of self-reported hypertension/diabetes/cancer/chronic lung disease/heart disease/stroke, household financial status, marital status, current alcohol use, and depressive symptoms


— of the elderly who reported pain on the initial intake evaluation, 69% reported pain subsequently (persistent pain)

— after covariate adjustment, persistent pain was associated with a 9.2% (2.8% – 15.0%) more rapid memory decline as compared to those without persistent pain

— after 10 years, there was a 15.9% higher relative risk of inability to manage medications, and 11.8% higher relative risk of inability to manage finances independently

— adjusted dementia probability increased 7.7% faster (0.55%-14.2%) in those with persistent pain, an absolute 2.2% increase in dementia probability


— the prevalence of chronic pain increases with age, affecting 25 to 33% of older adults. Also the recovery from chronic pain is less likely in older people. Older people with chronic pain are also more likely to be consumed by their chronic pain, reporting pain as affecting their subjective health status

— pain in the elderly has been reported to be more likely to be associated with falls, functional impairment, and cognitive decline/dementia in prior studies. But these are largely cross-sectional, observational studies at a point in time, and not a longitudinal study as above.

–although there are some very important advantages of the current longitudinal study, there are some evident limitations in terms of the extent of the data collected:  lack of data on the source/site of the persistent pain, or even of  the degree of the medical or psychological comorbidities. Also, there were very significant differences in the baseline characteristics of those with and without persistent pain, both medical and psychosocial (especially for depression and impaired ADLs, but also for all of the major medical comorbidities), or even the medications or other therapies prescribed. So, even though they adjusted for many potential confounders, they really did not have the database to control well for them, and there are likely other unknown (and therefore uncontrolled) factors involved.

— there were also no data specifically on the types of pain medications given. There is concern that use of opioids, for example, might affect cognition, perhaps exacerbated by concomitant disability, depression, etc. However, another study of patients with chronic pain getting NSAIDs found about the same amount of cognitive effect as with opioids. I would however add that several of the adjuvant pain medications (gabapentin, pregabalin, tricyclics, etc) may have significant effects on cognition. and SSRIs or SNRIs themselves, if used to treat depression or pain, can also affect cognition (studies are mixed on this)

–there is a plausible mechanism for the association of persistent pain and cognitive decline. Stress, depression, and even sleep deprivation have been found to affect cortisol secretion, and perhaps through that, can lead to neurologic changes, including decreased neuronal density/function in the hippocampus, with decreased ability to form new memories.

–and there may be other associated changes in those with persistent pain which were not accounted for but could affect cognitive health: eg, physical exercise, diet (ability to shop, cook food, etc), limitations in social interactions/increased social isolation


so, though these studies are both significantly flawed, it is clear that dementia is really prevalent, that it is likely increasing overall and especially so as the population ages, that there are very significant medical and social costs (especially on the caregivers), that many lifestyle measures likely provide benefit (as per prior blogs, see below), and that chronic pain by itself may be associated with the increased incidence and progression of dementia.  In this latter case, I think it is important, especially in the elderly, to treat pain but preferably with nonsystemic medications, avoiding the multitude of toxic effects of NSAIDs (cardiac, renal, GI, etc, to which the elderly are likely more susceptible), opiates, and the typical adjuvants (gabapentin, tricyclics, etc). Perhaps it is best to treat with physical therapy and psychosocial therapies (cognitive behavioral therapy, mindfulness therapy, stress reduction therapies, etc) and/or with injections or topical medications. and, as I have mentioned many times, injections into joints, trigger points, muscle spasms, etc often really help the pain and improve the function, often for a very long time.



blog on article finding that prescribing opiates to the elderly is associated with subsequent opiate dependence

article on cognitive behavioral therapy for low back pain

study on tai chi and another on mindfulness for chronic pain

Primary Care Corner with Geoffrey Modest MD: Low-dose hypertension therapy

7 Jun, 17 | by gmodest

by Dr Geoffrey Modest

A recent meta-analysis assessed 42 trials comparing low-dose (one-quarter dose) antihypertensive therapy versus placebo, finding some efficacy and essentially no adverse events at this low dose (See DOI: 10.1161/HYPERTENSIONAHA.117.09202).


— 42 trials were identified involving 20,284 participants

— 37 comparisons evaluated quarter-dose therapy versus placebo

— 7 comparisons were of dual quarter-dose therapy versus placebo

— 1 comparison of quadruple quarter-dose therapy versus placebo

— standard full-dose therapy for some of the commonly prescribed medications were:

— atenolol 50 mg

— metoprolol 100 mg

— carvedilol 25 mg

— amlodipine 5 mg

— verapamil 240 mg

— lisinopril 20 mg

— enalapril 20 mg

— valsartan 80 mg

— candesartan 8 mg

— hydrochlorothiazide 25 mg



— quarter-dose therapy versus placebo led to a 4.7/2.4 mmHg improvement (p<0.001)

— dual quarter-dose therapy was associated with a 6.7/4.4 mmHg improvement versus placebo (p<0.001), as effective as standard-dose monotherapy

— quadruple quarter-dose therapy versus placebo (only one study) found a blood pressure reduction of 22.4/13.1 mmHg (p<0.001)

— adverse events in both the single and dual quarter-dose therapies were not significantly different from placebo and had significantly fewer adverse events compared to standard-dose monotherapy



— the studies used in this meta-analysis were quite old, the average being published to 17 years ago, and 85% had eligibility criteria based on diastolic blood pressure alone.

— It is important to realize that this study only looked at office-based blood pressure targets, not clinical outcomes. For example, beta-blockers are now out of favor because of a few meta-analyses showing little benefit in terms of clinical outcomes. And some commonly used meds are not included (eg losartan).

— Some of the older JNC recommendations, perhaps based on some of these studies, did raise the issue of combining low-dose medications versus increasing a single medication, given the synergy in blood pressure lowering of the combination and the significant decrease in adverse events

— there was also an interesting meta-analysis from 2009 (see Wald DS. Am J Med 2009; 122: 290), looking at 42 trials with 10,968 participants, finding that doubling the dose of thiazides, beta blockers, ACE inhibitors, calcium channel blockers, and the summation of all of these classes of antihypertensives led to only 22% of the equivalent incremental effect of adding a 2nd agent (though calcium channel blockers did a little better than others, especially ACE inhibitors). This study really reinforces my own anecdotal experience, that doubling the dose of lisinopril, amlodipine, etc seems to have much less blood pressure lowering effect than adding a 2nd agent (e.g going from lisinopril 10 mg to 20 mg has pretty consistently a lower incremental value than adding low-dose hydrochlorothiazide to the lisinopril 10 mg, which has the added benefit of being available as a single combination pill). Sort of similar to statins where the most significant cholesterol-lowering is at the lowest dose (I have a patient on 1/3 of atorvastatin 10mg, ie 3.3 mg, with impressive lipid lowering). There was also another study finding that half-dose therapy was about 80% as effective as compared to full standard-dose therapy.

In a prior blog on the increased efficacy of chlorthalidone over hydrochlorothiazide, I lamented the fact that chlorthalidone was only available as 25 mg in the United States, that that dose was associated with significant hypokalemia, and that the pill was too small to cut easily. One of my readers commented that he actually took one quarter of the pill, did not find it difficult to cut the pill, and his blood pressure was well-controlled with that.


So, I do realize these meta-analyses are based on old studies, but it does seem to have validity, and is reinforced by my own clinical experience. Based on the study comparing doubling the dose versus combination medications, my practice has been to double the dose of the medication only if there is a compelling other reason (such as trying to decrease the level of proteinuria by maximizing ACE inhibitor dose) or if the medication is well-tolerated and the patient is pretty close to the target blood pressure.

Primary Care Corner with Geoffrey Modest MD: Nature vs nurture: studies on lipids and NAFLD

6 Jun, 17 | by gmodest

by Dr Geoffrey Modest

2 recent articles found significant genetic associations with 2 common clinical conditions: hyperlipidemia and non-alcoholic fatty liver disease (NAFLD). I bring these up to make the point that, though there is often a genetic association with disease, usually this is not determinant and there is an important role for a healthy lifestyle both in preventing the disease from happening and treating it if it does happen.

***Whole genome sequencing of two populations on the island Crete found some genetic variants which were either completely new, or rare in other populations, including one associated with increased HDL levels, and  another with decreased triglycerides and VLDL levels (see DOI: 10.1038/ncomms15606)



–945 patients from Crete, and specifically from Mylopotamos and Pomak villages, had genome-wide association studies done to look for common-frequency variants with small-to-modest effect sizes, evaluating 29 million single nucleotides variants in the Pomak and Mylopotamos cohorts. These results were imputed and validated for the larger populations in the areas

–there were 17 genome-wide significant independent signals found, including 2 novel cardiometabolic associations:

— chr16:70790626 for increased high-density lipoprotein levels (explaining 3.24% of the phenotypic variance in the Mylopotamos cohort)

— rs145556679, associated with decreased triglycerides and VLDL, so far only found in the Mylopotamos cohort and no other worldwide cohort.

— rs13382259, associated with decreased diastolic blood pressure in the Pomak population


–one advantage of looking at an isolated population is that low-frequency alleles will be amplified by the more limited gene pool

–though not commented on in the article, people from Crete are among the longest-livers in the world: Mylopotamos is called “Village of Long Life”. Of note, there is a similar genetic variant in isolated Amish populations in the United States. Interestingly, the people from Mylopotamos do eat a largely Mediterranean diet though their diet is high in animal fat, but with less atherosclerotic disease than would be predicted by their diet

***5-8% of lean people have evidence of NAFLD, despite the strong association with obesity. This study looked at 3 groups of patients stratified by BMI (see DOI: 10.1038/ajg.2016.318).  This is an Austrian study limited to Caucasian people; the prevalence of lean patients with NAFLD has been well-documented to be higher in Asian people but Caucasians had not been assessed previously. I am not sure about lean African Americans, though their overall prevalence of NAFLD is lower than either Latinos or white patients. This study was a subsample of the 2500 people in the Salzburg Colon Cancer Prevention Initiative study.


–187 nondiabetic patients were assessed by BMI groupings and had ultrasound evaluation:

–lean healthy: BMI <25, no steatosis, n=71

–age 56, 25male/46 female, bmi 22.7, whr 87 cm (waist-to-hip ratio), ALT 17, diabetes 0, impaired fasting glucose 13%, HDL 67/LDL 120/TG 86, A1c 5.4, adiponectin 13.8 [adiponectin has many functions but higher levels are associated with more insulin sensitivity], FIB4 score 1.2 [a noninvasive measure of hepatic fibrosis]

–lean NAFLD: BMI <25, steatosis, n=55

–age 61, 26male/29 female, bmi 23.6, whr 91 cm, ALT 21, diabetes 31%, impaired fasting glucose 20%, HDL 58/LDL 136/TG 117, A1c 5.5, adiponectin 10.4, FIB4 score 1.3

–obese NAFLD: BMI>30, steatosis, n=61

–age 63, 29male/32 female, bmi 32.7, whr 113 cm, ALT 28, diabetes 36%, impaired fasting glucose 32%, HDL 45/LDL 139/TG 169, A1c 5.8, adiponectin 8.9, FIB4 score 1.1

–all had oral glucose tolerance test,  serum metabolome, and genotyping for single-nucleotide polymorphisms (SNPs) associated with NAFLD


–lean NAFLD patients had fasting insulin levels similar to lean healthy ones, but did have markedly impaired glucose tolerance

–lean NAFLD patients  had a higher rate of the mutant PNPLA3 CG/GG variant vs lean healthy, 59% vs 35%.

–serum adiponectin levels were decreased in both NAFLD groups vs lean healthy (p<0.001 for both groups)

–metabolomics study showed variations in lysophosphatidylcholines, phosphatidylcholines, lysine, tyrosine and valine (several of these may have anti-inflammatory effects and increase insulin secretion, which could explain some of the differences in glucose tolerance between the groups)

–clinically no difference in lifestyle habits between the 2 lean groups, though the obese group did have more people consuming fast food at least once a week and had lower physical activity


–NAFLD is certainly much more prominent in those with obesity than in lean patients, though this does raise the question as to whether the prognosis is the same in these 2 BMI groups.  They did not do transient elastography (Fibroscans) or get liver histology in this study, though the lean NAFLD group did have the highest FIB4 score (suggesting more hepatic fibrosis, with a score of 1.3 being the cutpoint, suggesting that NAFLD is likely to be deleterious in this group)

–lean NAFLD patients were in between lean healthy and obese NAFLD in terms of metabolic derangements, though they did have much more impaired glucose tolerance than the lean healthy ones, and about 30% did have diabetes as defined by either fasting plasma glucose, impaired glucose tolerance, or A1c levels. The lower adiponectin levels in the two NAFLD groups further confirm their relative insulin resistance.

–the rate of the PNPLA3 allele were similarly high in both NAFLD groups, and much higher than in the lean healthy ones  (p=0.007). This does suggest that this genetic variant does seem to be associated with NAFLD independent of BMI (though not determinant, since 35% of the non-NAFLD people also had this allele). Unfortunately, there were no granular data presented to see if it were just this SNP which was responsible for the NAFLD, or if this subgroup happened to have a less healthy lifestyle. Would also be interesting to know if there were differences in their microbiomes as well, which is also affected by diet and exercise. See the multitude of prior blogs on NAFLD in the category, or just type NAFLD in the search window.

The reason I am evaluating these articles is to raise the general points that:

–clinically, we should still be suspicious of these underlying medical issues in patients who do not seem to fit the phenotype: a younger thin and athletic person who has a bad lipid profile, a thin person with an increase in their ALT levels on liver function testing, or evidence of fatty liver disease or glucose intolerance.  All of these people may have underlying medical problems (respectively, atherosclerotic disease, NAFLD) which might well affect their longer term morbidity/mortality

–and some anecdotes to fill this in, reinforcing the genetic components:

–one of my residents a few decades ago was curious about the strikingly high incidence of ASCVD/MIs in thin, athletic men in their 40’s in his Indian community near Boston. These men had surprisingly low HDL levels

–and lots of people who are morbidly obese with A1c in the mid 4 range, yet many who are much less obese (and some normal weight) who are either pre-diabetic or diabetic; I have found glucose intolerance in several thin patients with family history of diabetes

–as another example showing that risk factors may be discordant from the clinical picture: though obesity is the dominant risk factor for obstructive sleep apnea, it is found in about 11% of men with normal weight and about 3% of women (overall it is 2-3 times more common in men)

–so, we should think about risk factors when we see patients, but not be swayed too heavily and keep open mind to these diagnoses. Genes may be the dominant player in some patients.


On the other hand, there is usually some interplay between nature and nurture. Helping patients develop healthy lifestyles may well counteract the potentially deleterious effect of the genes, or potentially lead to more aggressive preventative approaches (eg, earlier uses of statins to prevent cardiac disease, etc)

–one interesting and I think poignant example of the nature/nurture interplay is that of the US Pima Indians in Arizona, who reportedly have the highest rates of diabetes and obesity in the US (see Schulz. Diabetes Care 2006; 29: 1866). In 1890 their water supply was taken over by white settlers, making it impossible to continue their agrarian lives. The US government subsidized  the food supply with sugar and white flour, and the previously lean population developed obesity and diabetes. It turns out that a group of Pimas, speaking the same language and from the same genetic stock, live in remote areas of rural Mexico and continue to practice their traditional agriculture. These Mexican Pimas have similar lower levels of obesity as other non-Pimas living in the same remote areas (around 7% in men and 20% in women), and have rates of diabetes close to the other non-Pima Mexicans (6.9% vs 2.6%). But the US Pimas have diabetes rate of 38% —  more than 5-fold higher!!! (and 68% in men older than 45, 70% of women older than 45 and 82%of women older than 55), which tracks with prevalent obesity (mean BMI=34), as well as their dramatically lower levels of physical activity there. A pretty clear example of a strong genetic predisposition to diabetes which manifests itself in the setting of a nonhealthy lifestyle/obesity….

so, I think all of this points to the fact that there are often important genetic factors leading to disease,and we as clinicians dealing with individual patients should still be attuned to patients who do not exhibit the typical profile (eg, those with high predisposition to heart disease or have abnormal LFTs from NAFLD but appear to be young and fit, or those who have OSA but are also of normal weight). And healthy lifestyles may not just help with these targeted diseases but are generally beneficial (collateral benefit… see blogs showing likely benefit from cancer, Alzheimers etc etc)

See here  for projected decreased cancer risk from exercise

See here for projected cardiovascular benefit from cardiovascular fitness

See here  for dietary influences on Alzheimers

Or, a slew of blogs on the benefits of Mediterranean diets

Primary |Care Corner with Geoffrey Modest MD: Spironolactone helps heart failure with preserved EF

5 Jun, 17 | by gmodest

by Dr Geoffrey Modest

One concern about patients having heart failure with preserved ejection fraction (HFpEF) is the dearth of effective treatments for long-term outcomes in this quite common problem. One of the most influential studies was the TOPCAT study, which found a nonsignificant benefit for spironolactone. A subsequent post hoc analysis (see below), however, found a significant outcome benefit ​ from spironolactone in those patients enrolled from the United States, Canada, Brazil, and Argentina (the Americas) but not in those from Russia and Georgia.

Details, in brief, of TOPCAT:

— 3445 participants with symptomatic HFpEF (EF>45%) were randomized to spironolactone vs placebo, mean dose 28mg/d, followed 3.3 years, around 80% also on a diuretic, ACE-I/ARB, and/or b-blocker, finding no difference in the primary outcome of the combination of cardiovascular death, aborted cardiac arrest, or hospitalization for heart failure. But the rate of hospitalization, as an isolated outcome, was improved a bit (12% vs 14%)


But in 2015 there was an article looking more directly at the regional variations in outcomes of the study, spurred on by the fact that the event rate in the placebo group in Russia/Georgia was so much lower than in the Americas (see Pfeffer MA. Circulation. 2015;131:34), finding:

— there was an unusually large, greater than fourfold, increase in primary outcomes in the patients in the Americas vs those in Russia/ Georgia, being 11.5 per 100 patient-years in the Americas vs 2.4 per 100 patient-years in Russia/Georgia.
— further breakdown showed that even in the placebo group, the respective numbers were 12.6 per 100 patient-years in the Americas and 2.5 per 100 patient-years in Russia/Georgia; and in the spironolactone group were 10.4 versus 2.3 per 100 patient-years– in the Americas, the primary composite event rate was actually a significant 18% lower in those on spironolactone, HR 0.82 (0.69-0.98); though in Russia/Georgia was nonsignificantly higher at HR 1.10 (0.79-1.51)

— review of all of the different cardiovascular outcomes assessed found a highly significant difference in the spironolactone groups in comparing the regional differences of these 2 areas, with almost all having p<0.001. Of note, there was also a much more significant increase in creatinine above 3.0 mg/dL, and lack of hypokalemia with potassium <3.5 mmol/L , in the Americas only (suggesting they were taking more spironolactone there)​

–My understanding is that these regional inconsistencies in the TOPCAT study, with the significant benefit in the Americas wing and suspicions about the accuracy of the Russia/Georgia group, led to the recent AHA guidelines endorsing the use of spironolactone in HRpEF (a Grade IIb recommendation, moderate level of evidence)​


And, a new analysis just came out assessing the serum canrenone levels (metabolite of spironolactone) in the stored serum specimens (see de Denus, S. N Engl J Med 2017; 376: 1690):

–206 patients from the US and Canada, and 160 patients from Russia had serum samples to assay; these people were representative of the overall TOPCAT populations from the different regions.

–of the patients assigned to spironolactone who reported taking the drug at the 12-month visit (76 of 101 patients from the US/Canada; 66 of 70 patients from Russia):

–canrenone concentrations were undetectable in 30% of the Russian group and 3% of the US/Canada (p<0.001)

–a significant correlation was found in the US/Canada between doses of spironolactone that people said they were taking and the canrenone concentration; no correlation for the Russian group

​–only those who had canrenone detected from both areas had significant increases in serum potassium and aldosterone levels (as anticipated)


So, all of this shows a few things:

–it is undoubtedly quite difficult to have equal quality control at different sites of care, and this is no doubt aggravated in different countries/cultures. Which means that we, as readers of these studies, should maintain a healthy skepticism regarding the results, especially when they are so discordant from the different sites, as in this study

–spironolactone seems to actually modify disease progression/associated cardiovascular effects, which is really important since not much else seems to help. The TOPCAT study tried to deal with the often difficult problem of making sure the patients’ symptoms were actually from the HFpEF (as opposed to being from the comorbidities, which it can be hard to differentiate from those related to those from HFpEF itself), by enrolling patients who had a recent hospitalization predominantly for heart failure or those with high natriuretic peptide levels (and those specifically who had high natriuretic peptide levels on subgroup analysis did show benefit from spironolactone). And diuretics overall help the heart failure symptoms, as well as treating underlying associated conditions (hypertension, atrial fib, myocardial ischemia, hyperlipidemia)

–my experience is also that spironolactone is a good augmenter of loop diuretics (I am treating a 93 yo patient with pretty severe symptomatic HFpEF, on torsemide 50mg but still symptomatic and with BNP of 639; I switched his diuretic therapy to torsemide 30mg and spironolactone 25 mg with quite dramatic clinical improvement and rapid decrease in BNP to 461, with further therapy limited by low blood pressure).

these studies therefore pretty much confirm that low-dose spironolactone is an important med in patients with symptomatic HFpEF, both in terms of symptomatic improvement (decreasing hospitalizations) and decreasing cardiovascular mortality, confirming guideline recommendations. And, it raises even further my respect for spironolactone — also seems to be a really potent agent in those with refractory hypertension


Primary Care Corner with Geoffrey Modest MD: Statins in the elderly?

26 May, 17 | by gmodest

by Dr Geoffrey Modest

​A post-hoc secondary analysis of data from the ALLHAT-LLT trial assessed the value of pravastatin, 40 mg per day, versus placebo in older patients, finding no benefit (see doi:10.1001/jamainternmed.2017.1442)


— the ALLHAT-LLT study, a community-based study which included 10,355 patients who had a fasting LDL level of 120 to 189 mg/dL and fasting triglycerides level < 350 mg/dL. 4546 were excluded for being younger than 65 and an additional 2942 because of baseline atherosclerotic disease, resulting in 1467 people in the pravastatin group and 1400 in usual care (UC)

— The intervention was open label pravastatin 40 mg per day versus UC. Mean follow-up 4.6 years

— mean age 71, 51% female, 40% white/31% Latino, 25% on aspirin/90% on anti-hypertensives/10% of women on estrogens, 22% current smokers, 51% diabetics, BMI 30, mean blood pressure 147/83

— baseline LDL was 148 mg/dL, decreasing to 109 mg/dL in those on pravastatin and 129 mg/dL in UC. The proportion taking statins was 78% versus 29%, respectively, by year 6 [ie, pretty large cross-over of patients]; this was more pronounced for participants 65 to 74 years old, where 32% of the UC group were taking a statin, though in those >75 yo, 15% were.


— all-cause mortality in pravastatin vs UC was non-significantly different for all adults 65 and older (HR 1.18, p=0.09), including nonsignificant HR of 1.08 those age 65 to 74, and 1.34 for those 75 years old or greater [ie nonsignificant trend to worse prognosis in group on pravastatin]

— coronary heart disease event rates were also not significantly different among the groups. No change with multivariable regression. No significant interaction between the treatment group and age.


— it is true that there are not rigorous data on older people and statin use, especially for adults greater than 75 years old, with much of the existing data being post-hoc analyses.

— One concern with the ALLHAT trial is that it was not a truly randomized controlled trial, with the control group being usual care. As a result, the 29% taking statins in the UC group might have been a very high risk group and thereby dilute the difference with the pravastatin group; and conversely 22% of those assigned to pravastatin did not take it (were they self-selected as healthier and less likely to have an atherosclerotic event??).  So ultimately, there was only a 20 mg/dL difference between the groups (and, pravastatin is one of the weaker statins). Another issue which may have disadvantaged the pravastatin group: among people 75 years and older, the mean systolic blood pressure was 150.6 mmHg in the pravastatin group and 147.5 mmHg in the UC group (p=0.01). Their finding of nonsignificant increases in overall mortality as well as in cardiovascular and stroke mortality, and heart failure event rates (all of which are strikingly contrary to the statin studies in younger people which show marked benefit of statins) suggest to me that intrinsic biases of the current ALLHAT study may have played an important role (though coronary heart disease rates were nonsignificantly lower in the pravastatin group)

–other issues: this was an open label study (with its attendant potential biases), did not include nonpharmacologic therapies of diet/exercise (and did those on statins stop doing these healthful behaviors because perhaps they found out from their clinicians that their lipids were better and they thought they didn’t need to continue the lifestyle changes, as was found in another statin study??), and those already on a statin were excluded from the study (those likely at higher cardiovascular risk, leaving the actual cohort at lower risk).

— There is reasonable theoretical benefit for using statins in the elderly: in people older than 65, about one half of all deaths are from atherosclerotic disease. And, statins reveal their effectiveness pretty quickly, within 6 months to a year.

— In general, although the relative risk of hyperlipidemia causing atherosclerotic disease in the elderly is lower than in younger people, the absolute risk in fact increases a lot with age.

— See blog , which reviews some of the data showing benefit of statins in the elderly, including a relatively recent meta-analysis of 8 trials, finding for example a 39% relative risk reduction for MI in those on statins, with number needed to treat of 24 for one year to prevent one MI. Relatively similar numbers for stroke prevention. Though these are mostly post-hoc subgroup analyses of bigger studies.


So, there really should be high quality randomized controlled studies of statin use in the elderly. The pathophysiology of atherosclerosis and the mechanics of statin use would suggest significant benefit for the elderly, even within 6 months of initiating statin therapy, and with minimal likely adverse effects. However, the current information is basically from post-hoc secondary analyses. Given our aging population and their high mortality rates from atherosclerotic disease, it really makes sense to have large rigorous studies. In the meantime, I will continue to treat the elderly with statins on an individual basis, having had (I think) good success: several elderly patients with severe CAD who have lived for decades on statins, dying in their mid 90s, and only one patient with an adverse effect/intolerance.

Also, as I mentioned in a recent blog, I am really concerned that the wrong message is getting out to clinicians: Physician’s First Watch and NEJM Journal Watch stated “Pravastatin Doesn’t Improve Clinical Outcomes in Seniors” . I think these sound-bite type analyses do in fact dumb-down the process of critical analysis of potentially important clinical articles, reinforcing (by omission) that the results of this study sound pretty definitive and thereby perhaps convincing some clinicians to stop prescribing statins. This approach is ahistorical (ie, omits the wealth of prior data suggesting benefit), and tends to reinforce the conception that the newest study trumps (perhaps bad term) all of the older studies (ironically, in this case, this study was based on a post-hoc analysis of a really old study, though I do think that in general newer studies tend to be accepted by us disproportionately, and inappropriately in some cases)


Primary Care Corner with Geoffrey Modest MD: Cellulitis treatment

25 May, 17 | by gmodest

by Dr Geoffrey Modest


Given the emergence of methicillin-resistant Staphylococcus aureus (MRSA), there are concerns about what empiric therapy to prescribe for skin infections, including cellulitis. A recent article in JAMA assessed the utility of adding trimethoprim-sulfamethoxazole (TMP-SMX) to cephalexin in treating uncomplicated cellulitis (see doi:10.1001/jama.2017.5653).


— 5 US emergency departments participated in an outpatient double-blind study of 500 patients >12 years old with cellulitis but without wound, purulent drainage, or abscess, seen from 2009-2012.

— Median age 40, 58% male, 57% white/35% black/1% Asian, symptom duration 3 days, 20% with a history of fever in the week prior to enrollment, 56% of infections were in the lower extremity/24% upper extremity/9% trunk abdomen or back/6% head or neck/4% groin or buttocks, 11% diabetics; mean length and width of erythema was 13 x 10 cm

— Bedside ultrasound was used to exclude abscess

— patients were randomized to cephalexin 500 mg 4 times a day plus TMP-SMX 160 mg/800 mg twice a day for 7 days, vs cephalexin plus placebo for 7 days

— primary outcome was clinical cure,  defined as an absence of fever, increase in erythema >25%, swelling, or tenderness at days 3-4; no decrease in erythema, swelling, or tenderness at days 8-10; and the presence of no more than minimal erythema, swelling, or tenderness at days 14-21.


— in a per-protocol analysis, which was limited to those patients who had a physical follow-up at 14-21 days after enrollment and took to at least 75% of medication doses: clinical cure at 14-21 days occurred in 182 (83.5%) of those assigned to cephalexin plus TMP-SMX, versus 165 (85.5%) in those on cephalexin alone, nonsignificant

— in the modified intention-to-treat population (participants took at least one dose of study medication, had an in-person or telephone assessment at the test of cure visit, and included those who withdrew from the trial or were lost to follow-up), clinical cure occurred in 189 of 248 patients (76.2%) in the combination group and 171 (69.0%) in the cephalexin group, a difference of 7.3% (-1.0% to 15.5%) p=0.07, almost but not quite significant. However in the more specific modified intention-to-treat analysis including those who took at least one dose of medications and had an in-person follow-up evaluation at any time during the study (ie, not including those who were lost to follow-up etc), the clinical cure rates were 83.8% in those on combined therapy and 82.8% of those with cephalexin alone, nonsignificant

— Adverse events, including the secondary outcome of overnight hospitalization, recurrent skin infections, and similar skin infections in household contacts through weeks 7-9, did not differ significantly. However, one patient on TMP-SMX did have acute-on-chronic kidney injury that resolved.

— 36 patients had treatment failure with cephalexin plus TMP-SMX: 10 (28%) were found to have an abscess the time of clinical failure and 9 (25%) developed an opening of the skin and purulent drainage

— 60 patients overall had treatment failure with clinical evidence of infection and had material available for culture: 41 (68%, and 10% of the per-protocol population) had MRSA, 8 (13%) had MSSA, and 3% streptococcal species, with no difference between treatment groups in the proportion having MRSA during follow-up.

— Post hoc subgroup analyses showed no difference between the groups if the patients had a history of fever or not, had diabetes, or by the size of the erythema.


— cellulitis is a common outpatient issue, and a difficult one because it is usually impossible to find the causative organism. b-hemolytic strep is often considered the cause, and the 2014 guidelines from the Infectious Diseases Society of America suggest choosing an antibiotic against Streptococci if there is no evidence of systemic signs of infection, penetrating trauma, evidence of MRSA elsewhere, or injection drug use. However, it is common for clinicians to cover MRSA infection in patients with cellulitis, given how widespread these infections are and how standard antibiotic therapy as with cephalosporins do not cover these MRSA infections.

— it is not so surprising that TMP/SMX does not add much to the treatment of cellulitis overall, since it does not provide much coverage for b-hemolytic strep, the presumed major causative organism

— One concern with this type of ED-based study is the higher likelihood of lower levels of medication adherence. They found that of the patients who took at least one dose of medications, only 52% were 100% adherent to the full regimen and 25% took 76-99% of their doses. This nonadherence was 3-fold higher in the cephalexin only group, for unclear reasons. But, the per-protocol analysis (ie those who took the meds) showed no difference between the outcomes of the treatment groups. I should add that some in this per-protocol group might have taken as little as 75% of their prescribed antibiotics, and this seems to me to be an arbitrary number. Which brings up the issue that some of these patients, in their most aggressively treated group, may actually have had insufficient antibiotic therapy. It would have been useful to have more granular data here assessing outcomes by finer gradations of medication adherence.

— Of note, of 36 patients who had treatment failure in the combined antibiotic group, over half had an abscess or skin opening/purulent discharge; and 71% of 28 in those on cephalexin alone. 10% of the patients who had clinical failure but had taken at least 75% of medication doses did have MRSA, which does suggest that even in patients with no clinical or ultrasound evidence of an abscess, there might well be MRSA (though, again, was this MRSA created or selected-for by taking insufficient quantities of antibiotics to cure the infection??) And this number is likely to be higher in many clinical settings, where cellulitis is defined exclusively clinically and without the use of ultrasound. It is somewhat reassuring that the treatment failures were similar in that both treatment groups had MRSA, though it seems that the numbers were too low to draw definitive conclusion.

–clearly from this study, there is a fine line between cellulitis and abscess formation. Even many of those who did not have an ultrasound-detectable abscess at the initiation of the study ultimately did develop fluid collections or skin opening with purulent discharge. And the proportion of patients with unknown fluid collections in our health center, for example, is undoubtedly higher than in this study since we do not have a bedside ultrasound machine (which limits the generalizability of this study in these settings). Again, it would have been interesting to have more granular data about the baseline characteristics of those who developed abscess/purulent discharge in this study and if there were any predictive models for them, since those who have abscesses are much more likely to have MRSA and should be treated differently. For one thing, those who are basically healthy (no immunocompromise, lots of comorbidities) who have no systemic symptoms and especially with small abscesses (eg <2cm) do just as well with incision and drainage and without antibiotics at all. And if antibiotics are indicated, they should cover MRSA.


so, this article, though not obviously generalizable to all settings (eg no ultrasound in many settings, perhaps higher medication adherence in primary care setting), does suggest reasonably strongly that TMP/SMX does not add much to the treatment of clinical cellulitis. Though imperfect, the per-protocol analysis does support this conclusion. Therefore, I think it is very reasonable to treat patients with cellulitis but without systemic symptoms/immunocompromise/etc just with a cephalosporin, thereby avoiding the unusual but occasionally severe adverse reactions of TMP/SMX, but with the caveat that there be close followup. Other studies have found that 90% of patients with cellulitis have significant clinical  improvement within 3 days, so that might be a reasonable target for a return visit. On the other hand, it seems reasonable to me to use the combo therapy in a patient who has a low probability for coming for follow-up, especially if one cannot rule-out a small abscess for lack of an ultrasound, where I would imagine that MRSA is more likely…


Primary Care Corner with Geoffrey Modest MD: Steroid knee injections: do they help??

23 May, 17 | by gmodest

by Dr Geoffrey Modest

A recent article in JAMA found that regular injections of intra-articular steroids was associated with decreased knee cartilage volume and no real improvement in pain in patients with knee osteoarthritis (see doi:10.1001/jama.2017.5283).


–140 patients with symptomatic knee osteoarthritis as well as synovitis by ultrasound (evidence of effusion synovitis, with suprapatellar pouch depth >2mm) were randomized to receiving intra-articular 1cc triamcinolone 40mg vs 1cc saline every 3 months for 2 years, both without local anesthetic

–mean age 58, 54% women, BMI 31, 65% white, mean hemoglobin A1c=6%, CRP 0.5

–all patients had radiographic evidence of Kellgren-Lawrence knee OA grade 2 or 3 (grade 2= definite osteophytes and possible joint space narrowing on anteroposterior weight-bearing radiograph; grade 3= multiple osteophytes, definite joint space narrowing, sclerosis, possible bony deformity)

–knee MRI was done at baseline and then annually


–there was greater cartilage loss with injected steroids (volume loss of 0.21 mm vs 0.10 mm with normal saline), though the amount of superficial fibrillations (fraying of the articular surface) was more common in the saline group (34% vs 13%)

–no significant difference between the groups in pain scores, or functional activities such as the 20 meter-walk time or the chair-to-stand time (these were all measured after asking patient to not take pain meds for 2 days prior to their evaluations)

–adverse events: overall more significant in saline group (63 vs 52, p=0.02), though no difference in what was considered treatment-related.  Cellulitis in one patient in the saline group, also hemoglobin A1c actually decreased significantly in the steroid group (-0.1% vs increase of 0.2% in the saline group, and this was controlling for BMI, radiographic DJD classification, sex). No difference in hypertension


–As noted in a recent blog on the lack of benefit of arthroscopy in patients with degenerative knee disease (see here​ ), knee DJD is remarkably common and a leading cause of  disability (and medical costs, largely for procedures)

–the physiologic rationale for intra-articular steroid injections is that DJD is typically associated with synovitis, with its associated elaboration of biochemical mediators having the potential for causing further joint destruction (collagenases, aggrecanases, cytokines). And local steroids might decrease the inflammation and this destructive cycle. Animal studies have supported this hypothesis. This study, utilizing MRI to assess the steroid effects on cartilage, seems better designed than prior studies which have used xrays, given how insensitive xrays are to assessing the radiolucent cartilage.

–so, how can one reconcile the conclusions of this study (negative impact on cartilage and no effect on pain) with the other studies finding pain improvement in the 4 weeks after the injection, an older but smaller study of 68 patients with the same basic protocol as in this study finding some benefit for pain, and with the huge anecdotal experience of benefit (steroid injections are done increasingly commonly)???

–these patients had pretty mild DJD, especially in terms of baseline symptoms, with a WOMAC pain score of 8.3. This score is based on 5 items (pain during walking, using stairs, in bed, sitting or lying, and standing upright, each with a score of 0-4, ranging from  None (0), Mild (1), Moderate (2), Severe (3), and Extreme (4); ie total maximum score of 20, with an average score of 8, as in the above study, being between mild and moderate.

–one issue with knee OA is how to define it or its progression objectively. The Framingham Study found a poor correlation between radiographic knee OA and symptoms (see Hannah MT. J Rheumatol 2000; 27: 1513, for example). And there is no accepted minimal clinically important difference for MRI cartilage measurement, as duly noted by the authors of this study. Also, I am concerned about the increase found in cartilage fibrillation found in the non-steroid group, since this early splitting of the tangential cartilage surface might harken more severe and clinically important cartilage changes over the somewhat longer term

–they only assessed pain relief at the 3-monthly evaluation, with no data on how patients fared in the first 1 or 2 months [and, in my pretty extensive experience with knee injections, probably around 1000 over the years, the vast majority of patients getting relief for the first few months, some much longer, and that relief translates into dramatic improvement in function and pain relief; ie they can walk and actually do things they couldn’t do before]

–there are even some literature (a meta-analysis of 38 studies) supporting saline injections as  helping with pain relief [ie, their control injections were not necessarily sham injections; saline itself may have some benefit. Which is an important difference. There seems to be a more profound placebo effect with injections than pills, so perhaps the real control for this injection study should be a needle in the joint with no meds injected???]

–and, in terms of generalizability of these results, it is important to stress that these patients had clinically mild knee OA at baseline, but still received injections every 3 months [not necessarily common clinical practice for those with mild-to-moderate symptoms], so their results might not apply to many patients who are actually getting knee injections for more severe, functionally limiting pain despite exercise/physical therapy/etc

–what about the decrease in cartilage thickness?? This is certainly concerning, though perhaps there are non-measured countervailing processes going on: are patients getting a lot of early pain relief [a good thing], but then using their knees more [walking, etc] which leads to more cartilage destruction through wear-and-tear???  And, though small, does the relative improvement in A1c in the steroid group reflect the patients’ ability to do more exercise?


So, how should this study affect clinical practice??

–my non-rigorously-tested finding, through loads of knee injections, is that 90+% of patients have much less pain and are able to function much better after injections, and I will continue to do injections

–that being said, injections should be accompanied with aggressive patient education around the importance of quadriceps strengthening exercises, which often help a lot [there were older studies suggesting this may not be true in patients with misaligned knees, perhaps from more severe DJD, where the patella does not track correctly and quad strengthening might exacerbate knee symptoms, but my sense is that this is relatively uncommon].  And some patients need a knee injection in order to do more exercise or physical therapy…

–other therapeutic options are sparse. Arthroscopic meniscal repair or joint lavage seems to do nothing . Physical therapy is important, but does not help many patients much (especially those who are frail, have advanced DJD, are unable to do the necessary home-based exercises,…). NSAIDs have a wide array of undesirable adverse effects, especially in the population with symptomatic knee OA, since they are typically older and have lots of comorbidities (and in this study, unlike NSAIDs, steroids were not associated with hypertension, for example)

–I also use the equivalent of 40 mg triamcinolone with 2cc of 1% lidocaine. This might have better efficacy (unknown to me) than just 1 cc of triamcinolone alone, since the added volume of the anesthetic may help the steroid reach more areas of the inflamed knee joint, and perhaps the anesthetic improves the pain relief beyond the steroid itself.

–after I have done a few knee injections, especially if there are diminishing returns (the first injections working for much longer than subsequent ones), I do discuss and recommend consideration of surgical management (usually knee replacement surgery)

–but I am certain that I will continue having patients, especially older ones, who often have serious medical comorbidities, who adamantly refuse surgery and really want repeated knee injections (even every 2-3 months) in order to function.  This study will change my practice in that I will discuss the issue of potential cartilage harm more forcefully than previously.

–one important general issue is my concern about the quick summaries of potentially clinically very important articles:  the one-line synthesis of this study was “Intra-Articular Corticosteroids Show No Benefit in Knee Osteoarthritis” in Physician’s First Watch/NEJM Journal Watch, and there was little more added in the few summary lines. I am very concerned that this type of analysis may undercut an important therapeutic modality for many patients, perhaps leading to fewer injections even though the patient may achieve very important pain relief and improved functioning/quality of life.  This brings up one of the reasons I do these blogs: we in primary care clinical practice are inundated with new articles (mostly drug-company sponsored) and new guidelines (often done by specialty societies whose members directly or indirectly are involved with drug companies, etc) on a daily basis. It is essentially impossible to keep up with the information onslaught. The summary services such as Journal Watch are really helpful in scanning the literature and alerting us to new articles/guidelines that might affect our clinical practice. But they may well have the very negative effect of dumbing down the literature to quick quips (sound bites?) that really make it impossible to figure out if a certain article or guideline really should apply to the patient sitting in front of us. My hope with these blogs is to look at a few of these articles that might well affect practice, give sufficient (and accurate) summaries of the methodology, types of patients involved, procedures done, and their results; then briefly put in my sense of how this article fits in with older literature and our model of disease physiology; and provide some specific concerns, if any, which might affect its clinical utility. This way, the reader can decide what they think about the article (or guideline), be able to review the specifics of the study​, even use my link to see the study itself for more details, and then figure out how or if they will integrate it into their practice

Primary Care Corner with Geoffrey Modest MD: Against arthroscopy for DJD of knees

16 May, 17 | by gmodest

by Dr Geoffrey Modest

The BMJ just published a systematic review comparing knee arthroscopy versus conservative management in patients with degenerative knee disease (see doi:10.1136/bmjopen-2017- 016114), an update of a prior review, adding ten new studies.


— 13 RCTs and 12 observational studies were included

— studies were diverse: the analysis included those with symptomatic degenerative knee disease, defined as persistent knee symptoms that affect quality of life and does not respond to conservative treatment, but with or without osteoarthritis and, in those getting arthroscopic surgery, “including any or all of debridement and/or partial  meniscectomy”.  Those with acute trauma were excluded.


— knee arthroscopy led to a very small reduction in pain in the first three months (mean difference 5.4 on 100 point scale), and very small or no pain reduction up to two years (mean difference of 3.1), high-certainty evidence. The MID (minimally important difference) being 12 points.

— knee arthroscopy led to a very small improvement in function in the short term (4.9 on a 100-point scale) and very small or no improvement at two years (difference of 3.2), moderate-certainty evidence. The MID being 8 points.

— very low probability of serious complications after knee arthroscopy, low-quality evidence. The most common serious adverse effect was venous thromboembolism, at a rate of 5/1000, followed by infection at 2/1000.



–symptomatic degenerative knee disease is remarkably common after age 45, affecting about 25% of people, from osteoarthritis of the knee joint lining and/or menisci.

–arthroscopic knee surgery for DJD (degenerative joint disease) is the most common ambulatory orthopedic procedure in the US (and 9th most common of all ambulatory procedures), associated with transient improvement in pain but requiring activity restriction for 2-12 weeks.

–so, this study found  evidence of minimal benefit from knee arthroscopy at 3 months (not considered clinically significant), which decreased further over the next 2 years, for both pain and function.

–one recent study was cited as an impetus for this systematic review, which included 140 adults (mean age 50; duration of pain 15 months). Though 96% did not have definitive radiographic evidence of OA (osteoarthritis), 91% had MRI-documented meniscal degeneration grade 3a or 3b, which is the worst grade. Patients were randomized to exercise therapy alone vs arthroscopic partial meniscectomy (see For the exercise group:  progressive neuromuscular and strength exercises over 12 weeks, 2-3 sessions/week. For those getting surgery: instruction for home-based exercises 2-4x/day to regain knee ROM  and reduce swelling. For the primary outcomes, no difference in pain at 2 years, using the KOOS (knee injury and osteoarthritis outcome score) subscales on pain, other symptoms, function in sports and recreation, and knee related quality of life (also no difference at 3 months; but the exercise group had had greater improvement in all muscle strength variables at 3 months). On secondary analysis the meniscectomy group had lower pain scores at 12 months that was considered clinically significant (though the exercise group had greater muscle strength at that time). [However, this study does not really rule-out the possibility that meniscectomy with aggressive PT afterwards is superior to both of their interventions, since the therapy post-surgery was just the suggestion to do home-based exercises].


so, the current study comes to the same conclusions as prior analyses, but by adding 10 more studies reinforces those conclusions that arthroscopy is not generally indicated in those with degenerative knee disease. And, I think, it does apply well to primary care, since it did not limit patients to specific MRI or xray findings (and I get MRIs of patients with typical chronic knee pain only quite rarely; and in general the utility of xray itself is questioned, since there is a pretty poor correlation with symptoms per the Framingham Study), or even to specific clinical findings, but seems to apply simply to those with undifferentiated chronic knee pain not responsive to conservative treatment, but affecting quality of life.

Primary C|are Corner with Geoffrey Modest MD: Postmarketing adverse effects of drugs

15 May, 17 | by gmodest

by Dr Geoffrey Modest

A recent article in JAMA, which made it into the popular press (see​ ), assessed post-market safety events of new drugs approved by the FDA between 2001 and 2010, finding a large number of serious adverse events after the drugs were approved and on the market (Downing NS. JAMA 2017; 317(18): 1854)


— all new drugs and biologics approved by the FDA between 2001 and the end of 2010 were assessed, excluding diagnostic agents, sunscreens, and drugs not intended for use in the United States, with follow-up through February 28, 2017. They did not include labeling changes and dosage form discontinuations.

— 222 novel therapeutics were approved, 183 pharmaceuticals and 39 biologics.

— 47 (21.2%) were for cancer treatment and hematology

— 37 (16.7%) for infectious diseases

— 26 (11.7%) for cardiovascular disease diabetes and hyperlipidemia

— 77 (34.7%) of these drugs received priority reviews and 28 (12.6%) received accelerated approval

— 62 (27.9%) were designated as orphan products

— median total FDA review time was 311 days, but was less than 200 days for 54 novel therapeutics, and greater than 400 days for 34.2%. 52 (23.6%) were within 60 days of their regulatory deadline approval dates.


— 123 new post-market safety events occurred, including 61 boxed warnings, during a median follow-up of 11.7 years, and affecting 71 of these new drugs (32% of all new meds).

— Three medications (valdecoxib, tegaserod, efalizumab) were withdrawn from the market

— median time from approval to first post-market safety event was 4.2 years

— post-market safety events were significantly more frequent among biologics vs meds, incidence rate ratio (IRR) 1.93 (1.06-3.52, 0.032), p=0.03.

— Within the therapeutics, those for psychiatric diseases had the highest IRR of 3.78 (1.77-8.06), p<0.001. Drugs for the treatment of cancer and hematologic disease had the fewest post-market safety events, with safety events at 10 years being reported in 60.0% for the psychiatric drugs and 21.4% for the cancer/hematologic ones

those medications receiving accelerated approval had IRR= 2.20 (1.15-4.21), p=0.02, as well as those approved within 60 days of their statutory decision deadline with IRR= 1.90 (1.19 -3.05), p=0.008

— no difference in post-market safety events among those drugs which were first-in-class versus the look-alikes


— as a baseline during this period, the FDA assessment of safety of novel therapeutics seems tilted toward approval: the majority of key (“pivotal”) trials, which served as the basis of FDA approval, enrolled fewer than 1000 patients and had follow-up of only six months or less; and approval seemed to happen more often just before the deadline for that decision

— one issue that is not commented on, at least that I have seen, is that several pivotal studies are terminated early, prior to their expected date, because of evident benefit. Sometimes this benefit is highly statistically significant, though the absolute benefit is not so great. And, by stopping the studies earlier than anticipated, there’s even less time for adverse events to manifest themselves.

— The above article highlights the problems with relying on postmarketing drug surveillance (which obviously needs to continue), finding that 32% of medications had postmarket safety events, and half of these were boxed warnings (ie, were considered severe). An even bigger issue is the poor compliance of drug companies and medical device companies to do and report postmarketing findings, despite these being specific FDA conditions for their approval. A blog on FDA-approved medical devices found that <20% of FDA-required post-marketing studies were actually done. A JAMA letter confirmed similar numbers for drugs (see Fain K. JAMA 2013; 310 (2): 202)

— BUT, the biggest concern now is that it seems that the FDA is poised for major changes which would lead to even more accelerated drug approval overall (the above study found that accelerated approval was itself associated with a higher incidence of postmarketing adverse events). Of the many many issues concerning the current Trump administration, I would like to highlight two:

— Trump  is fundamentally anti-scientific and anecdotal in his approach. This is concretely manifested by his ready denial of climate change and replacing 5-9 scientists on the EPA board with representatives of industry, since they “understand the restrictive nature of regulations”, and is planning a 40% reduction in funding for the scientific component of the EPA. He thereby shuts himself off from the potential to analyze policy from a scientific instead of a political or “gut” viewpoint. (see  for example).  In addition, his impetus to change the FDA, such that it leads to more rapid approval of medications, is grounded in an anecdotal example of a young girl with Pompe disease, a rare inherited disorder (which is even stranger since her father was in a position to found a company to develop enzyme replacement therapy). This rather extreme anecdote seems to him to justify really quick approval of drugs by the FDA, with less rigorous scientific evidence, and with the potential for real harm to lots of people.

— Trump just recently approved Scott Gottlieb as head of the FDA, with a commitment to get new drugs to the market sooner. Gottlieb is an advisor to several large pharmaceutical companies: “an unprecedented web of Big Pharma ties. He has spent most of his career dedicated to promoting the financial interests of the pharmaceutical industry”, per Dr. Michael Carome director of the Public Citizen’s Health Research Group ( ); the article also pointed out that “Gottlieb would be tasked by Trump with eliminating some of the regulatory burdens at the FDA”, getting new drugs to the market sooner.


So, I think that this postmarketing JAMA study and others suggest that the FDA may already be approving some drugs too quickly (and faster than their European counterpart), and this approval process is already slanted towards the drug companies (the studies are usually designed by the drug companies to get the outcomes they want through their own analyses of the results (see prior blogs, including a blog on empaglifozin and one on ezetimibe as cases in point ). And the current Trump administration changes are likely to make things a whole lot worse.  Even if there is a significant class action lawsuit and settlement against the drug company because of a severe adverse event found after FDA approval, even large settlement sums pale in comparison to the drug company profits from those drugs (at least in the cases I’ve seen).


Primary Care Corner with Geoffrey Modest MD: Osteoporosis treatment guideline

11 May, 17 | by gmodest

​by Dr Geoffrey Modest

The American College of Physicians just updated their clinical practice guideline on the treatment of low bone density/osteoporosis to prevent fractures in women and men (see doi:10.7326/M15-1361).


–treatment options

–bisphosphonates: high quality evidence that they reduce vertebral, nonvertebral and hip fractures in postmenopausal osteoporotic women (specifically shown with alendronate, risedronate, and zolendronic acid; ibandronate shown to reduce radiologic vertebral fractures; zoledronic acid reduces vertebral fractures in osteoporotic men). [BUT, these differences in bisphosphonates probably reflects the studies that have been done and may not indicate true differences between these meds. eg, we tend to assume that all ACE inhibitors are similar without specific studies showing true equivalence. Though it may not be true…..]. Adverse events: low-quality evidence that they can be associated with atypical subtrochanteric fractures (FDA issued warning), though very uncommon, in 1.78/100K in women taking bisphosphonates for <2 years, and up to >100/100K if >7 years (though these numbers pale in relation to fractures prevented in high risk patients). also, low quality evidence for osteronecrosis of the jaw (also rare). older concerns about atrial fibrillation were not found in newer studies. similarly with MI. but high-quality evidence of association with mild upper GI symptoms; hypocalcemia, flu-like symptoms, uveitis, and episcleritis with zoledronic acid; myalgias/joint symptoms with ibandronate and zoledronic acid. uncertain risk for cancer

​–denosumab (monoclonal antibody which prevents the development of osteoclasts through RANKL inhibition): high quality evidence that it reduces radiographic vertebral, nonvertebral and hip fractures  in post-menopausal women. Adverse events: high-quality evidence for mild upper GI symptoms; moderate quality evidence of increased risk of infection, (eg bacterial cellulitis, though no increase in serious infections); rash/eczema. also rate atypical femoral fracture and osteonecrosis of the jaw through 8 years of therapy

–teriparatide (hormone fragments of PTH):  high quality evidence that ​it reduces radiographic vertebral and nonvertebral fractures in post-menopausal women. Adverse events: high-quality evidence of association with mild upper GI symptoms, headache, hypercalcemia. also renal effects, hypercalciuria.

–SERMs (selective estrogen receptor modulators): high quality evidence that ​raloxifene reduces vertebral fractures in osteoporotic women, but non-statistically-significant decrease in risk of nonvertebral or hip fractures. Adverse events: high-quality evidence of hot flashes, thromboembolic events (incl pulmonary embolism, cerebrovascular death)

–estrogens: moderate quality evidence that ​it does not reduce fractures in postmenopausal women (a change in recommendations, based on newer data on postmenopausal women with established osteoporosis). Adverse events: high-quality evidence of association with cerebrovascular and thromboembolic events; higher incidence of breast cancer

​–calcium and vitamin D: moderate quality evidence​ that calcium or vitamin D alone has uncertain effect on fracture risk. Adverse events: probably no increase in MI with calcium (though shown in one analysis). increased risk of hypercalciuria with vitamin D

–physical activity: insufficient data to show conclusively that there is benefit for preventing fracture risk. no studies looking at comparative benefit of physical activity vs other interventions [though data show that physical activity can regulate bone maintenance and stimulate bone formation, increasing mineral content, and improving muscle strength (which protects the bone)]

–comparative benefits of above meds: insufficient head-to-head studies. and network meta-analysis to assess likely differences between meds did not find any


​–offer pharmacotherapy for women with known osteoporosis with alendronate, risedronate, zoledonic acid, or denosumab, to reduce risk of hip and vertebral fractures. strong recommendation, high-quality evidence

–treat osteoporotic women with pharmacologic therapy for 5 years. weak recommendation, low-quality evidence​​. the studies to support this excluded patients with severe osteoporosis (total hip BMD in the beginning T-score worse than -3.5),  or whose total hip BMD at 5 years was lower than their baseline. and post-hoc analysis found that those with pre-existing fractures or T <-2.5 after 5 years of therapy may benefit from continued treatment

–offer pharmacologic treatment with bisphosphonates to reduce the risk for vertebral fracture in men who have clinically recognized osteoporosis. weak recommendation, low-quality evidence​

–do NOT do bone mineral density (BMD) monitoring during the 5-year drug treatment period for osteoporosis in women. weak recommendation, low-quality evidence​.

–do NOT use estrogens or raloxifene in post-menopausal women in the treatment of osteoporosis. strong recommendation, moderate-quality evidence​

–make decision whether to treat women >65yo with osteopenia who are at high risk for fracture based on discussion, fracture risk profile, and benefits/harms/cost of meds. weak recommendation, low-quality evidence​ (there are some studies on post-hoc analysis suggesting those with T scores near the osteoporosis range finding that treatment with risedronate significantly reduced the risk for fragility fractures by 73%



–osteoporosis is exceedingly common: 200 million people worldwide, and 54 million in the US have osteoporosis or low bone density (50% of Americans >50yo are at risk for osteoporotic fractures). also, huge economic impact on health care system: $25.3 billion/yr by 2025.

–risk factors (per the article): age, female sex, menopause, hypogonadism or premature ovarian failure, low body weight, history of parental hip fracture, ethinicity (white people at higher risk), prior fracture (clinical or not), prior fracture with minimal trauma (“fragility” fracture), rheumatoid arthritis, smoking, alcohol (>2 drinks/d), low BMD, vitamin D deficiency, low calcium intake, hyperkyphosis, falling, immobilization, long term use of some meds (glucocorticoids, anticoagulants, anticonvulsants, aromatase inhibitors, cancer chemotherapies, gonadotropin-releasing hormone agonists). by the way, older studies have suggested that the compilation of risk factors of body weight/BMI, smoking, alcohol, and ethnicity only account for aoubt 15% of the risk variability (ie,  patients with lots of these risk factors  may well have normal BMDs and vice versa)

–BMD definitions: T-score in postmenopausal women and men >50yo is 2.5 SD below the young female adult mean (reported as worse than -2.5). BUT only 1/2 of the people with osteoporotic fractures have this low a BMD (which is probably because the BMD is a quantitative assessment of the amount of calcium in a cross-section of bone and does not reflect the internal qualitative structure of the bone). And, on the other side, osteoporotic patients on bisphosphonates with no improvement in BMD still have large reductions in fractures. There is also the Z-score, which compares people to those of the patient’s age and sex, and those better than -2.0 are “within the expected range for age”. The FRAX score (see and has been used about 5 million times, per the website) combines clinical risk factors, which can include BMD but does not need it, though data on benefit of FRAX are lacking. the threshold to treat varies by health care system and by clinical judgment, but rough guidelines, eg from the North American Menopause Society (see osteoporosis NAMS 2010 in dropbox, or DOI: 10.1097/gme.0b013e3181cdd4a7) is at least 20% for major osteoporotic fracture or at least 3% for hip fracture.

–in terms of physical activity, there are a slew of observational studies showing benefit, reasonable mechanistic support (the more one stresses one’s bones, the stronger they are; and conversely the less stressed, as through immobilization or prolonged bed rest esp in the elderly, the weaker). But it turns out that drug companies, the major sponsors of studies these days, are less interested in this nonpharmacologic modality. and, practically, it is probably unethical to randomize people to exercise vs not and follow them for years, given the multitude of exercise benefits, physically and mentally… and there are no real comparative effectiveness trials of the different meds because, I would guess,  drug companies choose the low bar of placebo control for their studies instead of comparing to another active med in order to maximize the likelihood for benefit and FDA approval.

–also, there really are limited data on treatment of men

— I would also add to the list of risk factors: HIV and some HIV meds (eg tenofovir disoproxil fumarate) and renal failure, as well as other medical conditions (hyperparathyroidism, monoclonal gammopathy and myelomas, etc)

–in terms of stopping therapy, I have been repeating the BMD after 5 years of therapy and continuing with bisphosphonates if the T-score is worse than pre-treatment or if it is more than about -2.5 to -2.7, after discussion with patients.  I am concerned that some clinicians are just stopping the bisphophonates at 5 years without reassessing the patient with another BMD, based on not knowing the exclusions or posthoc analysis of the relevant studies, which may put some patients at significant risk (see blog)


so, seems to me to be pretty reasonable recommendations overall. I do have lots of women on bisphosphonates without adverse effects, including into well-advanced age. The rationale here is that the risk of falls increases with age, as does the risk of hip osteoporosis. And the benefit of bisphosphonates in terms of fracture protection is evident within 11 months, and 8 months in those >70yo (see DOI 10.1007/s40266-016-0344-7).


see prior blog for a more detailed review of the studies on stopping bisphosphonates after 5 years.

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