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: lung cancer screening results at the VA

28 Feb, 17 | by EBM

By Dr. Geoffrey Modest

A recent real-world study reported on the results of the implementation of the low-dose CT (LDCT) lung cancer screening in smokers at 8 VA hospitals (see doi:10.1001/jamainternmed.2016.9022​).


  • 93,033 primary care patients assessed: 4246 met criteria for screening, 2452 [57.7%, a pretty low number…] agreed to be screened: 96% men, mean age 65,
  • Of note, there was a large variation in the number of positive LDCT screens by site, varying from 31% to 85%. [This raises the issue of lack of consistency in radiologist interpretation of LDCTs, which is also found in mammography evaluation and for several other x-rays]


  • 1257 (60%) had lung nodules, of whom 1184 (56%) required tracking (solid nodules <8mm without suspicious features (irregular or speculated borders) and not known to be new or growing based on prior imaging, ground glass nodules >5mm, or mixed solid and round glass nodules of any size)
  • 42 (2%) required further evaluation but did not have cancer
  • 31 (1.5%) had lung cancer within 330 days of follow-up
  • False-positive rate of 97.5% !!!
  • 857 (41%) had incidental findings (e.g. emphysema, other pulmonary abnormalities, coronary artery calcification)
  • They calculated that 880,899 patients in the VA system would meet criteria for lung cancer screening


  • The recommendation for LDCT screening of smokers was largely based on the National Lung Screening Trial (NLST), but
    • There were significant differences in the demographics of these VA patients’ vs the NLST participants: more men, older group (53% were 65 or older), more current smokers (57% vs 48%)
    • The rate of positive screens was more than twice as high in this study (60%, vs 27% in NLST)
    • I have sent out many blogs on LDCT screening in the past (see below), but my concerns are several: the large number of false positives, the amount of radiation, the fact that one good trial (NLST) which lasted only 3 years generated a massive screening initiative (which can last up to 22 years, or 25 years if you go by the USPSTF guidelines!!), had very few lung cancers actually detected (despite their extrapolation which projected saving 3 deaths/1000 high-risk individuals screened), did not include some high risk patients and did include some low risk ones, and reinforced the false perception that the main problem with smoking is lung cancer.
  • The editorialists wrote a very powerful response (see doi:10.1001/jamainternmed.2016.9446), noting:
    • For every 1000 people screened:
      • 10 would be diagnosed with early-stage lung cancer (potentially curable)
      • 5 diagnosed with incurable advanced-stage lung cancer
      • 20 would undergo unnecessary invasive procedures (bronchoscopy and thoracotomy) because of the screening
      • 550 will have unnecessary alarm and repeated CT scanning, with its attendant radiation [which, as noted in my prior blogs, actually increases the average radiation exposure from the low-dose from the initial screen by 4-fold to that of a regular chest CT, given the follow-up requisite high-dose regular CTs, PET scans etc.]
    • They also point out that many of the anticipated problems from LDCT screening were articulated by the CMS advisory body MEDCAC (Medicare Evidence Development and Coverage Advisory Committee), noting that they had “low confidence” that LDCT benefits would exceed the risks, and “high confidence” that evidence gaps remained after the initial studies (NLST did find benefit, though 3 European trials found no benefit)

So, to me, this VA study suggested several things:

  • I think it reinforces that there really should be multiple studies done in different patient populations (include some “real-world” sites, where the recommendations will actually be implemented)
  • That it is a bit crazy to generalize from a 3-year study to guidelines which could potentially expose millions of people to 22+ years of radiation.
  • That all of this is especially true before we embark on a screening test which is so resource-intensive. Not just the cost (which is a lot, and could be used for many other social or medical issues which are underfunded), but also the intensity of resources (developing systems to track these patients, carving out time from the already time-limited primary care encounter to deal with shared decision-making, being sure that the patient qualifies for the study, doing the referral for the screenings over the years, devoting the time and resources of other office staff to dealing with all of this as well, and then doing all of the above for following up on the very common incidental findings (41% in this study), false positives (97.5%) etc….
  • And, by the way, another article in the same journal (see doi:10.1001/jamainternmed.2016.9016 ) found that from 2010 to 2015 (NSLT was published in 2011), there were large % increases in LDCT done in never-smokers and low-risk smokers, such that many more of these who actually do not qualify per the guidelines are getting LDCTs than those who do qualify, suggesting that this very low-risk group is pretty undoubtedly getting risk with almost no benefit, and that there is some collateral damage to having guidelines: either confusion on the part of the clinicians, or insistence on the part of patients who do not want to be denied this (???) potentially life-saving intervention……
  • And, speaking of collateral damage, one of the big concerns in primary care is that we are working in a quite litigious society, and we may be medico-legally responsible if a smoker who meets criteria for LDCT does not get one, even if logic is on our side…

Prior related blogs:


Primary Care Corner with Geoffrey Modest MD: Is Mammography Useful?

14 Feb, 17 | by EBM

By Dr. Geoffrey Modest

This blog will bring up 2 recent studies suggesting the lack of efficacy of mammography screening coupled with significant overdiagnosis.

  1. An article a couple of years ago looked at screening mammography in the US, with 10 year follow-up of breast cancer incidence and mortality (see Harding C. JAMA Intern Med 2015; 175: 1483).


  • This was an ecological study of 16,120,349 women 40 years of older who resided in 547 counties reporting in the Surveillance, Epidemiology, and End Results (SEER) cancer registries during the year 2000.
  • 53,207 had a diagnosis of breast cancer and were followed for 10 years.
  • The researchers looked at the extent of the screening in each county, and the results of both breast cancer incidence and mortality (the latter being defined as women diagnosed with breast cancer in the year 2000 who had died from the disease during the 10 year follow-up period). Overall, in the 547 counties, the overall 10-year incidence based mortality was 47.2 per 100,000 cases diagnosed in 2000.


  • There was a strong positive correlation between the extent of mammography screening and the breast cancer incidence (P<0.001)
  • But, there was no relationship between screening and breast cancer mortality.
  • Each increase of 10 percentage points in the extent of screening was accompanied by:
    • A 16% increase in breast cancer diagnoses, RR 1.16 (1.13- 1.19)
    • Not even a trend to a change in breast cancer deaths, RR 1.01 (0.96-1.06)
  • Analyzing by tumor size, screening led to a higher incidence of small breast cancers (<= 2 cm), but not with a decreased incidence of larger breast cancers (>2 cm )
  • Each increase of 10 percentage points in screening is associated with:
    • A 25% increase in the incidence of small breast cancers, RR 1.25 (1.18- 1.32)
    • A 7% increase in the incidence of larger breast cancers, RR 1.07 (1.02- 1.12)
  • The following figure shows that as the proportion of women had a mammogram in the past two years, the incidence of breast cancer diagnoses increased significantly yet the 10-year mortality did not budge


  • So, pretty powerful large-scale epidemiologic study, finding that mammography led to a large increase in the diagnosis of small cancers, but there was no decline in the detection of larger cancers. This may be the reason why there was no significant difference in the overall death rate from breast cancer by doing mammography screening.
  • What does this mean? It may mean that there are a subset of very aggressive small cancers which spread and cause clinical disease and mortality, and that screening is didn’t help for these. And that a very large number of small cancers that are picked up by mammography are in fact “overdiagnosed” (defined as: tumors that would not have become clinically evident in the remaining lifetime without screening).
  • One would have expected that if screening did pick up small tumors earlier, that over time the diagnosis of larger and less treatable cancers should decrease. It is quite concerning that the number of larger breast cancers in fact continued to increase over the study. And, of course, the goal of screening is to reduce mortality, which was not found in the study. One additional finding was that increased screening would lead to more breast conserving surgical procedures; however they found no evidence of a decrease in extensive mastectomies.
  • Without getting into a lot of detail, the authors present reasonable arguments that this is not just lead-time bias, or ecological bias (this latter happens when looking at group data and assuming that it applies to the individual who may or may not have had a mammogram). Also, there was no association between mortality rates even comparing those counties with much higher breast cancer incidence, reducing the potential bias of comparing counties with very different incidences of breast cancer. But, they also did not have data on women who had therapy or what the risk factors were for the women who developed breast cancer. Also, this was just a 10-year follow-up, and patients may well live more than 10 years with newer therapies, but I would have expected some evident benefit of screening by 10 years (and at least a trend to benefit…)
  • There have been several important changes in technology over the past several decades, some of which may make older studies less applicable now (these older studies are the ones on which current mammogram recommendations are based). On the one hand, the sensitivity of our screening methods is greater and we are picking up much smaller tumors; and, perhaps these smaller tumors are more likely to regress than the larger ones picked up previous, leading to increased overdiagnosis. On the other hand, treatments have improved a lot, and the risk/benefit equation may have changed some. Given the potential harms of overdiagnosis (including surgery, radiotherapy, and chemotherapy), we should be looking at the new balance. In addition, there are interesting advances in genomic profiling, which are helpful in determining how aggressive a tumor is likely to be as well as how intensive therapy should be

So, a large study like this offers interesting insights, especially when looking at likely overdiagnosis (which one cannot determine in an individual patient). As with all screening tests (e.g. PSA), it would be really useful to figure out how to risk stratify patients, with more aggressive screening in those at higher risk. That is much more likely to show benefit for screening then with screening the general population.


  1. A more recent article look specifically at breast cancer overdiagnosis by mammography screening in Denmark (see doi:10.7326/M16-0270)). The study looked at women aged 35 to 84, from 1980 to 2010.


  • Denmark had a perhaps unique opportunity to look at the results of mammography screening both because it has rigorous databases (the Danish Breast Cancer Group, DBCG, and the Danish Cancer Registry, DCR) as well as a 17-year screening program which involved 20% of the population aged 50 to 69. This differential access to mammography screening allowed for real-world comparisons to a large, essentially randomized non-screening populations. Clinical breast exams were not included.
  • The DBCG database included 90,665 women aged 35 to 94 who were diagnosed with invasive breast cancer, and 4267 diagnosed with DCIS
  • For the mammography group in DBCG, they divided tumors into two groups: large (>20 mm) and small (<20 mm), considering the large tumors as “advanced” because they are equivalent to T2 or greater in the TNM classification system
  • The screening routine, somewhat different from what we do in the US, was biennial screening with a 2-view mammography on the first round, with 1-view mammography at subsequent screens except for women with dense breasts who always received a 2-view mammogram.
  • The DCR provided individual data on tumor size in women with invasive breast cancer.


  • For women aged 35 to 69: in non-screening areas the incidence of advanced cancer increased throughout the observation period.
  • For women 70 to 84: in the non-screening areas the incidence of advanced cancer also increased throughout the observation period, and was most pronounced in the later years.
  • The incidence of non-advanced tumors increased in the screening versus prescreening periods, incidence ratio 1.49 (1.43- 1.54), i.e. a 49% increase
  • Looking just at nonadvanced tumors, there were 711 invasive tumors and 180 cases of DCIS that were overdiagnosed in 2010 (overdiagnosis rate of 48.3% including DCIS and 38.6% excluding DCIS
  • There was no reduction in the incidence of advanced cancers through mammography screening.
  • Their conclusion: “it is likely that one in every three invasive tumors in cases of DCIS diagnosed and women offered screening represents overdiagnosis (incidence increase of 48.3%)”


  • This study is quite remarkable since it is reasonably close to a really large-scale randomized controlled trial, in which for 17 years 20% of women got mammograms and the rest didn’t. And there was no difference in advanced breast cancers through consistent mammography screening
  • See below for other blogs on the poor utility of mammography screening, also showing almost no decrease in breast cancer mortality but large numbers of overdiagnosed mammogram-detected cancer.
  • A lot of the “overdiagnosis” is from DCIS (about 25% of all new breast cancer diagnoses), which the National Cancer Institute now classifies as a “noninvasive condition” (an observational study of 108,196 women with DCIS in the SEER registries found an overall breast cancer death rate of 3.3% over 20 years, similar to the general population: see Narod SA. JAMA Oncol 2015; 1(7):888
  • All of this reinforces the fact that early detection of breast cancer is fraught. For breast cancer, there are 230,815 diagnoses/year in women, 2109 in men; and 40,860 breast cancers deaths/year in women and 464 in men, and affects 1 in 8 women!!!; yet mammography screening has perhaps minimal benefit. Which really brings up the issue of prevention (which, it turns out, does not get much funding). As noted in prior blogs, one big unknown is the prevalence of industrial toxins (many of which are estrogenic, including pesticides, BPA, others used in plastics, etc etc) which are in our environment and may well be carcinogenic. Large numbers of new chemicals are being used and thousands of new ones are introduced each year with minimal attempts to look at potential toxicity. In addition, it is reasonably clear from the studies that healthy diet, weight control, and exercise are helpful. It seems to me that it would likely be much more useful to devote our national resources into preventing breast cancer by regulating environmental toxins and promoting healthy lifestyles than attempting early detection.

See which documents the quite remarkable discordance in radiologists’ reading of breast densities

See​ for the 25-year results from the Canadian National Breast Screening Study finding NO benefit from mammography screening but that 22% of mammography-detected breast cancers were overdiagnosed.

See for a 2014 meta-analysis, finding that mammography yielded very small changes in breast cancer mortality (e.g. screening women in their 50s would lead to 3-32/10,000 decrease in breast cancer mortality, but have 6130 false positive and 30-137 overdiagnoses)​. As mentioned above, these studies were older ones.

Primary Care Corner with Geoffrey Modest MD: Colonoscopy Screening in the Elderly?

10 Nov, 16 | by EBM

By Dr. Geoffrey Modest

A recent observational study of Medicare recipients found that those 70-79 years old seemed to benefit from colorectal carcinoma (CRC) screening (see doi:10.7326/M16-0758). Study sponsored by the NIH.


  • 1,355,692 Medicare beneficiaries (from 2004-2012) aged 70-79, who were of average CRC risk, assessing 8-year risk for CRC and 30-day risk for adverse events.
  • Average risk was defined as: no history of adenoma, IBD, colectomy, and no colonoscopy/sigmoidoscopy/fecal occult blood in the past 5 years; and no prior abdominal CT, diagnosis of anemia, GI bleed, other GI symptoms, weight loss within the past 6 months
  • Included were those who were “health-conscious”, defined as having received at least 2 of the 3 preventive annual Medicare serviced of annual wellness visit, influenza vaccine, and breast or prostate cancer screening


  • 70-74yo, 8-yr risk of CRC was 2.19% (2.00 to 2.37%) in the screening group vs 2.62% (2.56 to 2.67%), so absolute difference of -0.42% (-0.24% to -0.63%)
  • 75-79yo, 8-yr risk of CRC was 2.84% (2.54 to 3.13%) in the screening group vs 2.97% (2.92 to 3.03%), so absolute difference of -0.14% (-0.41% to +0.16%) – i.e. nonsignificant
  • 70-74yo, excess 30-d risk of adverse events with colonoscopy was 5.6 events per 1000 people (4.4 to 6.8)
  • 75-79yo, excess 30-d risk of adverse events with colonoscopy was 10.3 events per 1000 people (8.6 to 11.1)


  • The current guidelines, as in many guidelines, varies by who is writing them. The USPSTF currently recommends screening by any of several methods, from 50-75 yo in those at average risk (evidence grade “A”, with individualized decisions in those 76-84, though the evidence grade here was “C”, meaning that they recommend”offering or providing this service to individual patients based on professional judgment and patient preferences. There is at least moderate certainty that the net benefit is small”
  • There are several concerns about drawing major conclusions from this new NIH-sponsored study:
    • Although there are 132,000 new cases of CRC in the US per year and 50,000 CRC-related deaths, it is not clear to me that this proportion applies in the more elderly population. As noted in this study of “health conscious” elderly, there was much more morbidity found in the older 75-79 yo cohort (e.g. hypertension in 80.5% vs 74.9% in the 70-74 yo, ischemic heart disease in 45.3% vs 36.6%). This increased morbidity is likely to translate to more people “dying with the cancer than dying from the cancer”.
    • All of this data is from the Medicare database, which, my guess, does not have the most accurate detailed information, and does not even have the CRC-specific mortality, a pretty useful endpoint for this study…
    • I am not so sure of the assumption that people who are more “health conscious”, as they define it, are in fact healthier/qualify as “average risk”. My guess is that the threshold for colonoscopy screening in the elderly varies lots by who the provider is (some may well push continued screening either in the undocumented belief they are helping the patient, they are uncomfortable effectively saying “you are too old to continue screening”, etc.), and some patients I see request different screens even with considerable morbidity (either they do not want to deal realistically with death/their prognosis, they are pretty somatic and want to  search for problems, etc). And, I would not be surprised if a higher percentage of less healthy patients get flu shots more aggressively (one of their “health conscious” criteria), either because of provider or patient preferences (and the fact that they come in for health care more often, with more opportunities for vaccines). Only a well-designed prospective trial would work to sort this out.
    • This study was limited to colonoscopy screening, which has been documented in the past to work much less well in the elderly, with higher numbers of inadequate preps (leading to more colonoscopies with more intensive preps), and (also perhaps related) higher perforation rates, which can lead to major abdominal surgery in an older and higher risk population. So, perhaps not the screening method of choice…
    • The stage-shift found in screening (i.e., fewer cases of more advanced CRC lesions in the screened group) certainly is supportive of screening, but again, colonoscopy is not only very expensive but quite invasive, so it really is important to look at real clinical outcomes before making a screening decision (i.e., does this stage-shift to higher stage lesions really translate to more morbidity/mortality?)
  • It seems to me to be a tad disingenuous to conclude in the abstract that “screening colonoscopy may have had a modest benefit in preventing CRC in beneficiaries aged 70 to 74 years and a smaller benefit in older beneficiaries”, but then in the last paragraph, having basically the same sentence, but with the qualification “and a smaller (if any) benefit in those who are older” (my emphasis). The reality is that many busy clinicians rely on the accuracy of the abstract and may not read the whole article, especially in primary care practice which is not only really busy, but requires clinicians to read and assimilate literature from all of the specialties. The above article also tends to minimize the adverse effects, stating they were “low but greater among older persons”. But, the rate was twice as high, and I would not be surprised if the actual effect of these adverse outcomes, in terms of resulting functional impairments, increases in an older population (they just don’t bounce back as well even from less-than-severe adverse effects).
  • And, this is really my main criticism of the take-home message of this study: I would phrase the conclusion more like “there is no clear evidence that screening colonoscopy offers any significant benefit in those 75-79 years old, that the possible benefit in terms of decreasing CRC diagnosis may translate even less into real morbidity and mortality benefit in this age group, and that there was almost a doubling of adverse events in this pretty susceptible population.” I personally do think that a healthy 79 yo, who really does have a realistic life expectancy (e.g., the healthiest 25% of women aged 80 has a 17 year life expectancy, and men 13 years), might realize actual clinical benefit by diagnosing and treating CRC early, especially since treatment for early lesions is pretty benign, but I have adopted FIT testing as my preferred non-invasive CRC testing, which should help winnow the colonoscopies and their adverse effects to a much smaller exposed group, and one with a higher yield for benefit over risks.

Primary Care Corner with Geoffrey Modest MD: Cervical Screening Guidelines From ASCO

17 Oct, 16 | by EBM

By Dr. Geoffrey Modest

The American Society of Clinical Oncology just published guidelines for the secondary prevention of cervical cancer (see doi: 10.1200/JGO.2016.006577, or go to ). These guidelines were unusual in that they stratified the screening approach based on the country’s resources, reflecting a global initiative, and also had several differences from the current US guidelines.


  • HPV testing is recommended in all resource settings, though visual inspection with acetic acid may be used in countries with basic resources.
  • Frequency of testing:
    • For countries with maximal resources: should be from age 25 to 65, every five years if negative.
    • For countries with enhanced resources: age 30 to 65. If two consecutive negative tests at five-year intervals, then every 10 years. Stop at age 65 if consistently negative results for the past 15 years
    • For countries with limited resources: age 30 to 49 every 10 years
    • For countries with basic resources: age 30 to 49, 1 to 3 times per lifetime
  • Treatment options for patients with a positive screen:
    • In countries with more than basic resources: colposcopy, then loop electrosurgical excision (LEEP) if positive
    • For countries with basic resources: treat with cryotherapy or loop electrosurgical excision
  • Follow-up post-treatment:
    • 12 month follow-up is recommended in all settings
  • For HIV-positive women (also applies to women who are immunosuppressed for any reason):
    • Overall, screen with HPV testing twice as many times per lifetime as in the general population (as above). Screening should begin as soon as they get the HIV diagnosis [? when to start if they are born with the infection/or get it from a transfusion at age 8???].
      • In countries with maximal  resources: screen with HPV every 2-3 years
      • In countries with enhanced resources: screen with HPV at 2-3 year intervals; but if negative, every 5 years (approx 8 screenings in lifetime)
      • In countries with limited resources: twice as often as in general population (4-6 screenings per lifetime)
      • In countries with basic resources: begin screening with HPV if available, or with visual inspection with acetic acid, at age 25, then every 3 years if negative initially. [A bit unclear, since they then suggest it will be approximately twice per lifetime]. These recommendations are based on murky data…
    • Postpartum screening: overall no screening recommended during the pregnancy, partly because the normal immune changes in pregnancy can have increased HPV changes which subside after pregnancy.
      • Screen at six months in all countries other than those with only basic resources, where screen at 6 weeks since longer interval could lead to loss of follow-up [though this might apply to other countries, or areas in other countries as well…..]
    • No screening should be done in people who had a total hysterectomy for benign causes [though see blog, which would support general screenings in HIV-positive women who had hysterectomy]
    • And, in countries with basic resources without mass screening — infrastructure for HPV testing, diagnosis and treatment should be developed
    • Self-screening: there is evidence that women doing their own HPV sampling may improve screening coverage, though the pooled sensitivity and specificity are lower, especially for CIN2+. So, overall not suggested except in women who might otherwise not get tested at all [and the sensitivity and specificity are actually only a little lower].
    • Postulated effect of HPV vaccine: likely to decrease the incidence of HPV 16/18 cancers, and with approx 5 year later onset of disease as a result of decreasing these most-carcinogenic genotypes, so potentially can start screening later in life, and decrease screening to ages 30, 45, 60. Maybe no need to screen at all??? Or only once??  But all recommendations are pending actual data….


  • High-quality screening programs can lower the incidence of cervical cancer by up to 80%
  • HPV is the most frequent sexually-transmitted infection, with one study finding 43% of college women getting infected over 36 months (see Ho GYF. New Engl J Med 1998; 338:423)
  • As a point of reference, the US screening guidelines at this point are quite different from the above ASCO ones for countries with maximal resources (e.g. the USPSTF recommendations):
    • Begin at age 21, do cytology screening only until age 30, and then every 3 years if normal. HPV testing not recommended because of higher likelihood of unnecessary follow-up and procedures (HPV infection tends to be transient)
    • After age 30, either continue cytology only every 3 years, or do cytology/HPV co-testing every 5 years, if normal results
    • Stop at age 65 unless there is increased risk (history of abnormal screens, prior HPV-related disease, immunocompromise, DES exposure); and if there are 2 negative consecutive co-tests or 3 negative cytologies within prior 10 years; and no history of high-grade dysplasia or worse
  • Although ASCO cites the importance of HPV testing, they do not make formal recommendations about primary HPV testing vs co-testing, noting just that some countries and regions have moved towards adopting primary HPV testing (see )
  • The American College of Obstetricians and Gynecologists also just came out with their recommendations in 2016 (though, not sure what to make of this: but these guidelines were retracted from the January 2016 issue of their publication Obstetrics and Gynecology, and I could find only recommendations for HIV-positive women, though these were also retracted), with a few differences from ASCO. for HIV infected women:
    • Start screening within one year of onset of sexual activity, but no later than age 21
    • Screening should be continued throughout a woman’s lifetime and not stop at age 65
    • For women less than 30 years old:
      • Cytology screening (without HPV testing) should be repeated in 12 months (though some people feel it should be followed in six months)
      • If three consecutive cervical cytology tests and normal, follow-up cervical cytology should be done every three years
      • If ASCUS on cytology, and reflex HPV testing is positive, then colposcopy. If HPV testing results are not available, repeat cytology in 6 to 12 months if more advanced dysplasia is found, refer for colposcopy
    • For women older than 30 years old, do cervical cytology or co-testing:
      • If only cytology is done, follow-up is as for women less than 30 years old
      • If co-testing is done and negative, repeat at three years; if cytology is negative and HPV positive, repeat in one year (though if HPV 16/18 is present go directly to colposcopy). if either of the co-tests at one year is abnormal, colposcopy
    • The hope is that as HPV vaccination becomes more widespread, the incidence of cervical cancer will decrease significantly as well as the need to screen for it; though this is in the relatively distant future, given the high prevalence of HPV infections currently, and the vaccine does not help those currently infected or with abnormal cytology from infection
    • They do recommend starting screening at age 25, as is done in several countries in Europe for example, noting that there is lack of evidence of the benefit of decreased cancer risk in those under 25 (very uncommon), and potential harm or screening and overtreatment. The United States still recommends initiation of screening at age 21. And though HPV infections are remarkably common in women under 25 (as noted in study above), HPV infections clear spontaneously, and 90 to 95% of those with even LGSIL as well as many with high-grade lesions regress spontaneously
    • And they recommend HPV screening at age 25 in countries with maximal resources, different from the general recommendations in the US to start HPV screening at age 30
    • So, my guess is that the formal US recommendations will change significantly in their next iteration (the USPSTF recommendations date from 2012, with an anticipated update 2018). Perhaps internal controversy led to ACOG retracting their guidelines??  But it is pretty clear that recently the approach to cervical cancer screening has changed significantly in other countries.

Primary Care Corner with Geoffrey Modest MD: Radiologist Variability in Mammography Readings

13 Oct, 16 | by EBM

By Dr. Geoffrey Modest

A recent article revealed the dramatic variability in radiologists’ interpretations of mammographic breast density (see Sprague BL. Ann Intern Med 2016; 165: 457). Determining breast density accurately is certainly important because increased breast density leads to difficulty in reading mammograms and is an independent risk factor for breast cancer. In this light, one prerequisite for us in primary care is that the radiologic determination of breast density is consistent and accurate. But, details, from an NIH supported study:

  • Data from 216,783 screening mammograms from 145,123 women aged 40 to 89 were included, from 30 radiology facilities within three breast cancer screening research centers of the Population-based Research Optimizing Screening through Personalized Regimens (PROSPR) consortium.
  • 83 radiologists were involved; each interpreted at least 500 screening mammograms from 2011-3, using the BIRADS reporting system, along with patients age, race, and BMI


  • 9% of mammograms were rated as showing dense breasts
  • Across radiologists, the finding of dense breasts ranged from 6.3% to 84.5% (median 38.7%, interquartile range 28.9% to 50.9%). !!!!
  • Variation in breast density assessment was pervasive in all but the most extreme patient age and BMI combinations
  • Among women with consecutive mammograms interpreted by different radiologists, 17.2% had discordant assessments of breast density.


  • One of the scariest issues to me as a clinician is that I need to rely on an accurate interpretation of medical tests in order to inform my patient management. The sheer magnitude of the variation in breast density assessment is quite striking.
  • There are also other studies, mostly 10-20 years old, showing that the general radiologic interpretation of mammograms has considerable variability as well.
  • There are certainly other tests that have significant variability andhighlight this issue more broadly — for example finding significant spine MRI abnormalities in totally asymptomatic patients:
    • One study (see Jensen MC. N Engl J Med 1994; 331: 69) looked at 98 people without back pain, where their MRI scans were interpreted by two experienced neuroradiologists at the Cleveland Clinic, finding that 52% had a bulge in at least one intervertebral disc, 27% had a protrusion, and 1% had an extrusion. 38% had multilevel abnormalities. Only 36% had a normal MRI.
    • A systematic review (see Brinjikji W. AJNR2015 36: 811-816) found dramatic MRI or CT changes in asymptomatic people, which increased with age.  For example, the prevalence of disc degeneration went from 37% at age 20 to 96% at age 80, bulges went from 30% at age 20 to 84% at age 80, disc protrusion from 29% age 20 to 43% at age 80, annular fissure from 19% age 20 to 29% at age 80.  so, lots and lots of impressive disc changes even in asymptomatic 20 year olds……
  • Another issue, which we tend to understand more intuitively, is that of ultrasounds, which are clearly operator-dependent. But we had a patient with chronic hepatitis B, who had a “normal” screening RUQ ultrasound for hepatocellular cancer, but a CT revealed a 9cm cancer!! I spoke with a trusted hepatologist who commented that he used a CT to scan to screen for really high-risk patients because of the variability of ultrasounds (though that is not exactly a clear-cut, or generally accepted algorithm….)
  • One major concern about over-reading breast density (as well as potentially scaring patients that they might be at higher breast cancer risk) is that this findingoften leads to further studies such as ultrasound, digital breast tomosynthesis, and MRI examination (though there is minimal evidence to support these tests, and they may well lead to unnecessary biopsies, more radiation exposure, etc. And the USPSTF formally gives these procedures an “I” rating, for insufficient evidence)
  • And, another issue:  half of the United States has legislation currently requiring disclosure of mammographic breast density, in some cases advising women to discuss supplemental screening tests with their providers if they have dense breasts (again without supportive medical evidence). And even theFDA is considering a legislative requirement to report breast density information to patients. I think there is a real concern about non-medical legislators enacting medical legislation, where legislators may be swayed by patients pleading for unproved treatments, perhaps with the support of an “expert witness”. Or, perhaps the legislature decides to require a certain treatment based on small or flawed studies, writes the treatment into law, but then new and better studies contradict this legislative imperative. One recent example is in Massachusetts, where a law was passed requiring insurance to cover long-term antibiotic therapy for chronic Lyme disease, though several studies, including a new one (see Berende . N Engl J Med2016;374:1209-1220), have not found benefit from long-term antibiotics. Or, in the past, there has been legislation supporting the availability of bone marrow transplants for women with breast cancer, but without any evidence of benefit (and pretty clear harm). I do realize that there have been egregious, inappropriate treatment denials by some health insurers in the past which has led to some of this legislation and public/medical community outrage. But legislating medical diagnostics and therapies is fraught…..
  • So, this inconsistency/unreliability in breast density interpretation may subject many people to potentially dangerous interventions. I think it is really important that we as clinicians understand that many procedures we order are subject to large variability, as above. So, what can we do??
    • Whenever possible, we should interpret these “objective” data in the context of the clinical situation of the patient, and not always reflexively respond to the test results (it is just another piece of data, such as from the history or physical, which should be put in the overall gestalt of what is going on with the patient). Of course, some of these objective findings, even unsuspected, may be very important and not dismissed (e.g., the incidental finding of early pancreatic or renal cancers).
    • Maybe we should consider getting second opinions more often than we currently do, to assess interobserver agreement
    • Perhaps there should be triggers in place for certain findings (such as dense breasts on mammogram), requiring a blinded read by another radiologist, or?? always having mammograms automatically re-read by another radiologist??, or having an automatic second-read whenever a radiologist comments “should be repeated in 3-6 months by another test”, which puts the medicolegal imperative on us in primary care to do yet another test with potentially more radiation exposure, cost, possible unnecessary procedures, etc.
    • Perhaps there needs to be much more transparency in the system overall, maybe requiring those reporting on these results to have regular standardized testing themselves and posting the results (sort of like requiring hospitals to report their C-section rates).
    • Perhaps we need a good computer program????
  • I realize this blog is more tangential than others, but i do think this issue of inconsistency in mammography reading does bring up a slew of general issues in clinical medicine…..

Primary Care Corner with Geoffrey Modest MD: Cervical Cancer Screening Less Frequently?

12 Oct, 16 | by EBM

By Dr. Geoffrey Modest

A recent review of cervical HPV screenings in the Netherlands found that those with negative screening could potentially be screened less frequently than every 5 years (see


  • 43,339 women aged 29-61 with a negative HPV and/or cytology were randomly assigned to HPV and cytology co-testing (intervention group) or cytology testing alone (control group); with 3 screens: at baseline, 5 years and 10 years; and with followup at 14 years. Those in the cytology only group also got HPV testing but this was blinded to all.
  • Mean age 43
  • Their triage approach (different from US recommendations):
    • For intervention group (cytology plus HPV):
      • Normal HPV and cytology: repeat in 5 years
      • At least moderate dyskaryosis on cytology: colposcopy
      • HPV positive, and neg or borderline/mild dyskaryosis (eg ASCUS or LGSIL) on cytology: repeat HPV/cytology at 6 and 18 months. refer to colposcopy if continued HPV positive or cytology worse
    • For control group (cytology only)
      • Normal cytology: cont routine screen
      • At least moderate dyskaryosis on cytology: colposcopy
      • Borderline/mild dyskaryosis (eg ASCUS or LGSIL) on cytology: repeat cytology at 6 and 18 months. refer to colposcopy if cytology same or worse


  • Co-testing group: 20,490 of 21,623 women had double negative HPV/cytology, 764 had pos HPV/neg cytology, 369 pos cytology/neg HPV
  • Cytology only group: 20,533 of 21,716 had negative cytology, 814 had pos HPV/neg cytology (the HPV results were blinded), 369 pos cytology/neg HPV
  • During 14 years of followup:
    • Co-testing: 149 CIN2, 152 CIN3 (including 5 adenoca in situ), 8 squamous cell and 6 adeno carcinomas
    • Cytology only: 126 CIN2, 169 CIN3 (including 5 adenoca in situ), 17 squamous cell and 10 adeno carcinomas
  • Breakdown of the 14 year followup according to cytology and HPV status (again, HPV results were blinded for the control group)
    • Cancer:
      • Cytology neg/HPV neg: 7 in intervention, 12 control; 3.3 vs 5.7/100,000 women, incidence ratio 0.58 (0.23-1.48), nonsignficant
      • Cytology neg/HPV pos: 4 in intervention, 15 control; 55.4 vs 190.9/100,000 women, incidence ratio 0.29 (0.10-0.87)
      • Cytology pos/HPV neg: 3 in intervention, 0 control; 79.7 vs 13.4/100,000 women, incidence ratio 5.97 (0.30-119.22), nonsignficant [but they had to use 0.5 instead of 0 for the cancer count, in order to do the math]
    • CIN3+ (the combination of cervical cancer and precancer):
      • Cytology neg/HPV neg: 74 in intervention, 86 control; 35.0 vs 40.7/100,000 women, incidence ratio 0.86 (0.63-1.17), nonsignficant
      • Cytology neg/HPV pos: 82 in intervention, 94 control; 1135.1 vs 1196.1/100,000 women, incidence ratio 0.95 (0.71-1.28), nonsignificant
      • Cytology pos/HPV neg: 10 in intervention, 16 control; 265.7 vs 427.1/100,000 women, incidence ratio 0.62 (0.28-1.37), nonsignficant
    • The cumulative incidence of cervical cancer 14 years after the initial negative cytology/negative HPV screen in the co-testing group (0.09%) was the same as in the cytology negative patients in the cytology-only group after 9 years
    • The cumulative incidence of CIN3+ was 0.56% 14 years after the initial negative/negative screen in the co-testing group, but 0.69% in the cytology negative patients in the cytology-only group after 9 years
    • Combining both groups, the incidence of CIN3+ was 72.1% lower (60.5-80.4%) in women >40 years old vs younger; no statistically significant difference in cervical cancer


  • Several studies have supported using only HPV screening without cytology (primary HPV screening) for detection of cervical dysplasia/cancer (g., see from BMJ or Ronco G. Lancet 2014; 383 (9916): 524); the latter study found that there was 60-70% better protection with primary HPV screening over cytology screening. And primary HPV screening might avoid over-referral to colposcopy and biopsies. And decrease the number of screens done/longer intervals between screenings. Several countries now do primary HPV screening including Australia, Italy, Netherlands, New Zealand, Sweden and the UK. The current study looked not just at cervical cancer, which may take years to manifest itself, but also to high-grade precancerous lesions (CIN3+) to try to ascertain if the longer screening interval could miss women with evolving cancers (which it didn’t: those with combined screening had the same incidence at 14 years as the cytology only group at 14 years).
  • So, this study suggests several things:
    • It confirms the superiority of HPV/cytology screening over cytology alone
    • The very low incidence of CIN3+ in the overall combined groups (including the blinded HPV testing of the cytology-only group) who had negative HPV testing (independent of cytology) was quite low: 84 events in 20,859 patients (e.g., as compared to those who were HPV positive but cytology negative, with CIN3+ in 82 of 764 patients), affirming that HPV testing is superior to cytology testing
    • The study also confirmed the utility of testing more than just the highest risk HPV 16/18 types, since there were 30 of 501 patients with CIN3+ who were HPV positive/cytology negative and HPV 16/18 negative
    • And the big conclusion was the very low risk of CIN3+ and cervical cancer itself in patients who were >40yo and had dually negative initial HPV/cytology
    • Putting this all together, in 2017 the Netherlands will implement the strategy of every 10-year screening for HPV negative women at least 40 years old
  • So, there really seems to be increasing data suggesting that primary HPV is a superior screening test (adding cytology seems to add more false positives than providing real clinical benefit), though i would imagine there need to be more studies in different populations to see what the optimal screening interval should be.

Primary Care Corner with Geoffrey Modest MD: Non-alcoholic Fatty Liver Disease 1

7 Sep, 16 | by EBM

By Dr. Geoffrey Modest

There have been several articles recently on non-alcoholic fatty liver disease (NAFLD) in a recent special issue of the journal Digestive Diseases and Sciences, as well as a recent release of NAFLD clinical management guidelines by the European Assn for the study of NAFLD. Since NAFLD is so common throughout the world, since it is amenable to lifestyle interventions, and since there was so much interesting info on NAFLD but so many unresolved questions, I will devote 3 blogs to this:

  1. Natural history of NAFLD
  2. Review of therapies, with more detail on a couple of topics (e.g. the role of the microbiome and of specific dietary components, esp. fructose)
  3. A review of the EASL guidelines for NAFLD


NAFLD Natural history (see Goh G. Dig Dis Sci 2016; 61: 1226)

  • NAFLD (nonalcoholic fatty liver disease) was first defined only in 1980, which is rather surprising given that it is: incredibly common (in the US its prevalence has increased from 5.5 to 11% between 1988 and 2008), the most common liver disease in the world with estimated prevalence of 20-30%, the most rapidly rising indication for liver transplantation, and likely to be the number one indication for liver transplantation by 2020. The fact that the prevalence is increasing, however, is not so surprising, since NAFLD is so closely related to insulin resistance, obesity and metabolic syndrome. It is important to keep in mind that with all of these statistics, there are real issues of differing definitions and ascertainment bias overall. For example, in one study the prevalence of ultrasound-diagnosed hepatic steatosis with normal liver enzymes was 16.4%, but the prevalence of hepatic steatosis with abnormal LFTs was 3.1% (i.e., determining NAFLD by ultrasound vs abnormal LFTs as the NAFLD yields very different prevalences).
  • NAFLD is formally defined as the accumulation of >5% fat in the liver, not attributable to alcohol, drugs or other secondary causes, and represents the spectrum from NAFL (non-alcoholic fatty liver, or steatosis) to necroinflammatory changes of NASH (non-alcoholic steatohepatitis), advanced fibrosis, cirrhosis and HCC (hepatocellular carcinoma). The biopsy may be indistinguishable from alcoholic steatohepatitis.
  • Long-term prognosis: the most common causes of death are cardiovascular (the number one cause: see Mantovani A. Dig Dis Sci 2016; 61: 1246), malignancy and liver disease. And it seems likely that a large % of those diagnosed with “cryptogenic cirrhosis” actually have NAFLD. The extent of these outcomes varies in different studies, from not much of an increase to being equivalent to hepatitis C. However, it seems that most studies find the mortality increases significantly as one goes from NAFL (steatosis only) to NASH, and seems overall to be much worse with more severe stages of NASH (e.g. one study with 18.5 years of follow-up found that liver-related mortality increased from 3% in non-NASH to 18% in those with NASH). A meta-analysis found that mortality was not much higher in those with simple steatosis vs the general population, but in those with NASH there was still an 81% increase in overall mortality  and 471% increase in liver-related mortality.
  • Progression of NAFLD: in a meta-analysis of 133 patients with simple steatosis, 39% developed progressive fibrosis, 53% remained stable and 8% improved. This translates to an average annual progression rate of 1 fibrosis stage over 14 years. Also, NASH progresses: a study of 221 patients found that 37% had progressive fibrosis on repeat biopsy over 5 years. Overall, it seems that in patients with NASH and no fibrosis, there is a 1 stage progression of fibrosis over 7 years. BUT there are a small group who have much more rapid progression.
  • 10-25% of patients with NASH progress to advanced fibrosis/cirrhosis. In a small Australian study comparing patients with NASH cirrhosis to those with hepatitis C, about 40% of each group developed liver-related complications over 7 years, though other studies have found lower mortality rates than hep C (but with more cardiovascular mortality). There are some data suggesting that higher serum ferritin levels (> 1.5x upper limit of normal) is associated with a higher likelihood of NASH and more advanced fibrosis.
  • BUT, one counterintuitive point that makes it difficult to rely simply on noninvasive testing: there is no relationship with the height of serum transaminases and the degree of hepatic inflammation or fibrosis. And, there can be significant hepatic inflammation without increased transaminases.


  • These data make it very difficult to figure out what is best to do with patients who have increased ALT levels (really common). In general, we screen for other causes of increased LFTs, especially for viral hepatitides (esp hep B and C, but I also check to make sure either immune to hep A naturally or by immunization, and immunize against hep B if nonimmune), autoimmune hepatitis (e.g. ANA, anti-smooth muscle, anti-liver-kidney microsomal antibody-1), iron overload (iron, TIBC, ferritin). And get an ultrasound.
  • One concern with NAFLD as an entity is that it really is a diagnosis of exclusion. That makes it more likely that NAFLD is not a single condition: there could well be unknown causes of fatty liver that are lumped together in “NAFLD”; there seem to be a variety of predisposing conditions (though insulin resistance is the most common, there are many people with NAFLD without that); and there are such variable prognoses (some never progress, some regress, some advance, and some advance very quickly).
  • Should we be screening for NAFLD (not currently recommended)? From the NHANES data, its incidence has increased from 5.5% around 1990 to 11% in 2008, and the % of cases of chronic liver disease attributable to NAFLD has increased from 47% to 75% during this time. (i.e., much more common than for other causes, which we do screen for). And there are interventions that help (see later blog). I personally do screen with LFTs in obese kids and all adults, and (not surprisingly, given the frequency of NAFLD, have found many cases, much more than hepatitis C). And I have had some success in convincing patients to lose weight and do more exercise based on these results. But should we be doing more inclusive screening with an ultrasound, to pick up the many cases where the LFTs are normal (and we know that LFT changes can be transient, and may never be found, despite the possibility of significant hepatic inflammation)????? I am not doing that, but it certainly seems reasonable….
  • What is the best way to follow those with steatosis on ultrasound or raised ALT levels suggestive of NAFLD? Should we be following ultrasounds routinely to look for progression? And if we do serial ultrasounds, how often? Biopsy is currently considered the only method to really see if there is active inflammation or fibrosis. Hopefully in the not-so-distant-future we will have reliable non-invasive tests: e.g. transient elastography which might helpdifferentiate NAFL from NASH and also track its progression (some small studies found it was able to differentiate degrees of steatosis, and another finding a stepwise increase in liver stiffness that correlated with the degree of biopsy-proven hepatic fibrosis). And there are potential serum markers: e.g. FIB-4 (a calculation involving age, AST, ALT and platelet count), which correlates well with the degree of hepatic fibrosis; and other markers look promising (e.g., cytokeratin-18, which reflects hepatocyte apoptosis). More studies are needed on these, but there are glimmers of hope that we can avoid biopsies.
  • And if we decide to do a liver biopsy, how often should they be done, in light of the pretty high progression rate to NASH and fibrosis? What about trying to pick up those who are rapid progressors?
  • One other consideration: since NAFLD is so common and has an attendant increased risk of cardiovascular disease, I have a low threshold to prescribe a statin, especially in middle-aged and older patients. See blog which reviews several of the studies, in patients with NAFLD, hepatitis B and C, finding clinical benefit. In NAFLD, there are some data showing that statins both reduce cardiovascular mortality, but also have some benefit in improving NAFLD histology or its future complications (advanced fibrosis, etc.), as well (also see Mantovani A. Dig Dis Sci 2016; 61: 1246 for more info).

Primary Care Corner with Geoffrey Modest MD: USPSTF Recommendations on Lipid Screening in Adolescents

30 Aug, 16 | by EBM

By Dr. Geoffrey Modest

The USPSTF just came out with their lipid screening recommendations for children and adolescents (see JAMA 2016; 316(6): 625), giving them an “I” rating (current evidence insufficient to recommend screening). For full supporting documents of the USPSTF recommendations, see JAMA 2016; 316(6): 645). Their points:

  • They consider both dyslipidemia from genetic heterozygous familial hypercholesterolemia, FH, (1 in 200-500 people in North America and Europe), where there are really high cholesterol levels (LDL>190 mg/dl, often 2-3x that of unaffected people) and evidence of increased cardiovascular risk (though not typically until age 30: 1 in 6 men and 1 in 10 women have ischemic heart disease by age 40, 25% of women and 50% of men by age 50), as well as multifactorial dyslipidemia, MD, (may have genetic component as well as environmental, esp obesity, where the LDL is lower but >130). Several studies show that there is moderately good tracking of childhood hyperlipidemia into adulthood. But they comment that there are no long-term data that adolescent hyperlipidemia from MD leads to clinical cardiac events in adults. Or that treatment of the MD changes those hard clinical outcomes [i.e., no RCTs have been done, and it is hard to imagine that they ever will be. Would require lots of kids randomized into different treatment wings, maintaining those treatments for decades, and then being able to follow them until they were 50+ years old when more clinical events happen]
  • NHANES data suggest that 7.8% of children 8-17 yo have total cholesterol >200 mg/dl, and 7.4% of those 12-19 have LDL >130 [i.e., pretty common]
  • For those with FH there are data that short-term (<2 year) treatment leads to decreased LDL and decreases the early surrogate marker of carotid intima-media thickness [which actually seems to be a good predictor of future atherosclerotic disease, especially in kids]. There are no compelling data on the long-term benefit of statins through a randomized-controlled trial, since in those with FH it is considered unethical not to use statins. However, a retrospective analysis looked at children with FH who were put on statins for a mean of 10 years starting at age of 14, and compared them to the adults with FH not on statins (they predated the use of statins); they assessed outcomes by age 30 , finding that the adults had many more cardiovascular events at a younger age than the kids (at the age of 30, the cumulative CVD survival was about 90%, vs 100% in the kids, even though 28% of the kids were smokers. See Braamskamp MJ. Am J Coll Cardiol 2016; 67(4): 455). It was notable in this study that the youngest parent with an MI was 20 years old and died at age 23.
  • Harms of screening. Also inadequate evidence, though mostly concerns about medicalization (anxiety about diagnosis, labeling, potentially harmful therapy). Statins are well tolerated with transient adverse effects in kids (increased liver enzymes). And though cardiovascular disease is still the number one killer, there is concern about overdiagnosis (i.e., some people treated may never develop cardiovascular disease)
  • Therapy: statins are typically used, given the adult data. But no consensus on when to start in kids with FH (some say age 8-10, some age 20). No data on long-term use in kids [though there are potential issues concerning taking statins with pregnancy, which they did not mention….]
  • They do acknowledge the rationale for screening kids:
    • Atherosclerosis is a known progressive process which starts in kids, with autopsy studies showing that [my data, not in their document]:
      • Everyone has fatty streaks in their coronaries by age 15-34
      • Advanced atherosclerotic lesions are found in 2% of men and 0% of women aged 15-19, but
      • Advanced atherosclerotic lesions are found in 20% of men and 8% of women aged 30-34
      • The Bogalusa Heart Study found that in adolescents dying mostly from trauma at median age of 19.6, there was a strong correlation between the levels of cardiovascular risk factors (BMI, lipids, and BP) and the degree of atherosclerosis
    • Lipid levels in kids are associated with the extent of adult atherosclerosis
    • Familial hypercholesterolemia, FH, is associated with premature ischemic cardiovascular disease
    • Short-term treatment of patients with FH leads to substantially lower LDL levels and some evidence of improvement in atherosclerosis (see below)
    • Abnormal lipid levels in adults is strongly associated with coronary heart disease events
    • Early identification and intervention in adults can prevent such events
    • And, I would add, prevention of more advanced lesions upfront not only decreases mortality (still a significant number of patients die with their first MI, and more are chronically disabled), but also helps long-term, since statins do not reverse atherosclerosis (just stabilizes it and makes the plaques less likely to rupture), but the residual recurrent cardiac risk in those who sustain an MI (secondary prevention) even with statins remains much higher than in those on statins who are just at elevated cardiovascular risk (i.e., primary prevention)
  • Although there are some potential benefits of identifying and treating those with FH, there are no good data for those with multifactorial dyslipidemia. An Ohio universal screen program (n=6500) looked at nonfasting total cholesterol screening, finding elevated levels >200 in 8.5%. This cohort then had fasting lipids, finding 5.8% had LDL>130. The USPSTF review found that those with the highest likely yield for hyperlipidemia are kids with obesity (12.3%), overweight (8.9%), children 9-11yo (7.2%) and those 16-19 yo (7.2%) [There is an unexplained typical 10-15% decrease in cholesterol levels during puberty].


  • When reading guidelines, it is important to remember that USPSTF overall is an independent governmental agency (so less influenced by drug companies, etc., than some of the professional society recommendations), is very focused on rigorous data (so will not make recommendations if the issue has not been studied well. They say the same thing about blood pressure screening in kids: no data that screening leads to decreased future cardiovascular events, so an “I” rating). In fact, for pretty much the same reasons (lack of clearcut studies), USPSTF does not recommend screening average risk males until age 35 and women till 45. All of these USPSTF recommendations are much less aggressive than the Am Acad of Pediatrics (screen at age 2-10 in all obese kids or if family history of dyslipidemia, etc.), NCEP (screen at age 20), Am College of Physicians, etc.
  • And, as a perspective, we all screen for phenylketonuria in newborns, with a prevalence of 1 in 10,000, but not for familial hypercholesterolemia, with 50x the prevalence (>500,000 born each year)… though I do realize the rigor of studies showing benefit is different)
  • It is pretty clear that targeted screening of those considered to be high risk is not successful: in a few pediatric epidemiologic studies, 1/2 the kids would be missed by relying on parental information/ family history of dyslipidemia or premature cardiovascular disease (e.g., see Ritchie SK. Pediatrics 2010; 126(2):260). And, pretty surprisingly, only about 1/4 of FH patients receive their appropriate diagnosis by middle age (see Neil HA. BMJ 2000; 3212:148). From other organizations (e.g. Am Acad of Pediatrics), screening has been recommended as universal. Universal screening is also much easier to integrate into care than targeted screening.
  • I do definitely think that discussing diet and exercise is really important at all age groups (though adolescents may be the most resistant of them….). But I also do think, from my clinical experience, that this discussion is more effective, at least in some people, if there are personal specific markers that suggest that the individual may be at higher risk. And knowing the lipid levels of adolescents may be very helpful in changing behavior. For example, I think that knowing a person smokes helps individualize their treatment recommendations and is likely to be more effective than just telling everyone they should not smoke (of course, finding out that someone smokes is a bit less invasive than doing a blood test).
  • So, to me, the rationale to do testing wins out. One may pick up the relatively unusual cases of familial hypercholesterolemia, and it seems pretty intuitive that they need meds as supported by our understanding of their physiology and supplemented by some limited data. But one will mostly pick up the pretty common multifactorial hyperlipidemia, and I personally would pursue these patients aggressively with diet and exercise, based on logic but without definitive studies to prove it. I.e., I would use the found high lipid levels to talk with the patient (motivational interviewing) about diet and exercise and suggest much more aggressive follow-up than on their non-dyslipidemic peers. I would not start meds in this group, but if they were able to improve their lifestyle, would track their lipids and give them feedback. And, it turns out that most adolescents do not have optimal lifestyles (see next study)


Not so surprisingly to those in clinical practice with kids/adolescents, there are very impressive data that adolescents have pretty bad lifestyles in terms of cardiovascular health. A recent AHA scientific statement (see DOI: 10.1161/CIR.0000000000000441) noted:

  • 91% had poor diets, 9% intermediate diets: specifically, overconsumption of sodium, sugar, solid fats, refinded carbohydrates; under consumption of fruits, vegetables, whole grains, dairy, dietary fiber
  • Only 10% of boys and 5% of girls aged 16-19 years old had the recommended 60 minutes of moderate-to-vigorous exercise per day (including muscle-strengthening and bone-loading activities at least 3 days/week. [By the way, my untested hypothesis is that kids can use their cellphones, which seem to be pretty ubiquitous even in poorer communities, to track their “numbers of steps” they take each day, in order both to quantify an important aspect of exercise and give direct feedback to them]
  • 27% of12-19 year olds are obese (BMI>95th percentile)
  • This document does promote “early identification and control of dyslipidemia, including heterozygous familial hypercholesterolemia, throughout youth and into adulthood can reduce clinical cardiovascular disease risk beginning in young adult life”. Also it is typical that total cholesterol decreases 10-15% around puberty, independent of diet (suggesting that we should check lipids when kids are 9-11 yo, before these pubertal changes). Non-HDL is more predictive than any single lipid marker. NHANES data suggest that 26-35% of adolescents have “intermediate or poor” levels of total cholesterol overall (worst in Mexican-Americans, best in non-Hispanic whites)
  • Blood pressure. Use sex/height-specific percentile charts. 9-12% have intermediate or poor BP status
  • Fasting glucose <100 mg/dL. Part of the AHA’s Strategic Impact Goal Through 2020 and Beyond, though they note that this metric is not currently used in pediatric practice, misses the boat in obese children since hyperinsulinemia is the first sign of metabolic derangement, and does not adhere to the Am Diabetes Assn definitions (which are the same as for adults). By the fasting blood glucose <100 metric, 20-38% of 12-19 year olds have intermediate or poor levels in the NHANES study.
  • Smoking (twice as common in adolescents exposed to second-hand smoke: reinforcing the need to get parents/caregivers not to smoke, at least in the house). NHANES data suggest that about 1/3 of 12-19 year olds have tried cigarettes within the prior 30 days.


So, this all reinforces the approach of lipid screening in adolescents, despite the lack of clinical outcome data. To me, the reasons to screen younger people are not just to pick up the extreme cases of familial hypercholesterolemia (which typically requires meds) but to use the results as a concrete means to educate adolescents/families about the need for lifestyle changes, with a focus on diet, exercise and obesity, for anyone with hyperlipidemia.




A Response to: USPSTF Does Not Back Lipid Screening in Adolecents

See the response below, sent around with Holly’s permission. Her study (hyperlink below) in 3 sets of adolescents/young adults aged 17-21 and their parents (including some with familial hypercholesterolemia, obesity, and generally healthy) posited different cholesterol screening scenarios, finding that in each of these 3 groups, both the adolescents and the parents saw worse cholesterol results as signifying poorer health, with several commenting about the need to change their lifestyles. This provides support for universal cholesterol screening and using the results to try to influence behavior.  Even in adolescents.



Awesome summary Geoff about an issue near and dear to my heart (pun intended!).

I am interested in whether knowing about heart health/heart disease risk changes teens behavior.  Probably not but certainly gives us something to anchor our counseling on as you note.  You might find this qualitative study we did asking teens about their hypothetical response to lipid screening interesting –

Thanks as always for such an awesome blog!!!

Holly Gooding, MD, MSc

Division of Adolescent/Young Adult Medicine, Boston Children’s Hospital

Division of General Internal Medicine, Brigham and Women’s Hospital

Harvard Medical School, Boston, MA


Primary Care Corner with Geoffrey Modest MD: USPSTF Diabetes Screening Misses Most People

26 Jul, 16 | by EBM

By Dr. Geoffrey Modest

A study looked at the sensitivity/specificity of the current USPSTF guidelines for diabetes screening in a community setting, finding over half the cases are missed (see ). The USPSTF in 2015 recommended diabetes screening for those aged 40-70 and who are overweight/obese.


  • Retrospective analysis of electronic health record data of 50,515 adult primary care patients seen between 2008-2010 in 6 health centers in the Midwest and Southwest, followed for up to 3 years (median 1.9). [This screening was prior to the 2015 USPSTF guidelines]
  • 18,846 (37%) were >40 yo; 33,537 (66%) were overweight or obese; 39,061 (77%) were racial/ethnic minorities (35% Black, 334% Hispanic, 9% other)
  • They excluded patients with dysglycemia (glucose intolerance or diabetes) at baseline
  • They then compared the actual findings of dysglycemia (by the usual fasting or post-prandial sugar/A1c) criteria with who would have not been screened if adhering to the later-published 2015 USPSTF guidelines


  • 29,946 (59%) had a glycemic test within 3 y of follow-up
  • 8,478 of them developed dysglycemia (78% by using the A1C criteria)
  • 12,679 (25%) of the 50,515 patients overall would have been eligible for screening per the 2015 USPSTF guidelines
    • Overall sensitivity of the guidelines was 45.0% (43.9-46.1%)
    • Overall specificity was 71.9% (71.3-72.5%)
    • PPV was 38.8% and NPV was 76.8%
    • Subgroup analysis:
      • Compared to normal weight: overweight people had 31% more dysglycemia; obese 145% increase
      • PCOS 124% increase
      • Dramatic increases in dysglycemia in those with increasing number of diabetes risk factors (the Am Diabetes Assn risk factors as noted below)
    • On multivariate analysis, significant associations with the development of dysglycemia: age >40, overweight/obese, nonwhite race/ethnicity, hypertension, PCOS, history of gestational diabetes, family history of diabetes
    • Dysglycemia cases in racial/ethnic minorities were significantly less likely to be eligible for USPSTF-guideline based screening, though they had higher odds for developing dysglycemia (OR for Black patients 1.24; Hispanic 1.46). The sensitivity for different racial/ethnic groups was:
      • White: 54.5%
      • Black: 50.3%
      • Hispanic/Latino: 37.7%
      • And the lower sensitivity in racial/ethnic minorities reflects the greater proportion of patients who developed dysglycemia at a normal weight and under 40 yo (e.g., 20% of Hispanic/Latino patients of normal weight developed dysglycemia, as well as 31% of those <40yo)


  • Diabetes is really common: from the USPSTF document: “approximately 86 million Americans aged 20 years or older have IFG or IGT. Approximately 15% to 30% of these persons will develop type 2 diabetes within 5 years if they do not implement lifestyle changes to improve their health”
  • Many studies over the last several decades in several different countries show that intensive lifestyle interventions can prevent or at least delay the development of diabetes, with reasonable argument that this would significantly decreased the associated micro- and macrovascular morbidity. This strongly supports the likely utility of screening/potential for earlier intervention.
  • It should be noted that the Am Diabetes Assn has much more expansive guidelines (see , p16) for testing asymptomatic adults:
    • All overweight (BMI >25, or >23 in Asian-Americans) with at least one additional risk factor: physical inactivity, first-degree relative with diabetes, members of high-risk ethnic group (African American, Latino, Native American, Pacific Islander), women who had baby >9 # or had gestational diabetes, hypertension, HDL <35 or triglyceride >250, other clinical condition associated with diabetes (e.g. acanthosis nigricans), history of CVD
    • In absence of above, everyone at age 45
  • The NICE guidelines in the UK focus on those at high diabetes risk, independent of obesity (see )
  • There are clear limitations of this study. This was not a prospective study of all-comers to assess the prevalence of dysglycemia, along with individual risk factors. And even in the PLoS analysis, they do not disaggregate the issues of age<40 and normal weight (i.e., it is not entirely clear if normal weight people who also are <40 yo have a significant incidence of dysglycemia). [On my brief search, I was quite surprised at how little epidemiological data was available on the prevalence of glucose intolerance in the US population, including different ethnicities/other subgroups.]
  • My experience absolutely reflects the results of the PLoS study: we find many people who have either glucose intolerance or diabetes who would not qualify under the current USPSTF guidelines (I have been using the Am Diabetic Assn guidelines, though I should add that in our health center, a very large % of our patients under age 40 do qualify for testing by these guidelines). However, though this study is not definitive, it really raises the ante: it seems that a fair number of patients under 40yo or with normal weight have dysglycemia. And it seems to me that the benefits (early reinforcement of intensive lifestyle changes) far outweigh the risks….

Primary Care Corner with Geoffrey Modest MD: Lung Cancer Screening for Smokers, an Individual Risk-Based Approach

20 Jul, 16 | by EBM

By Dr. Geoffrey Modest

The USPSTF strongly recommends low-dose chest CT (LDCT) annual screening for ever-smokers with >30 pack-year smoking history aged 55-80 or until they are 15 years after stopping smoking, based on the 3-year National Lung Screening Trial (NLST). JAMA just published an article evaluating the use of risk models/individual risk-based strategies to help focus the LDCT intervention (see  doi:10.1001/jama.2016.6255). By looking at individual lung cancer risk beyond the criteria of NLST, they actually found relatively higher risk in some patients with a low risk by NLST (and therefore no screening done in NLST or offered by USPSTF) but a low risk for many included in the USPSTF guidelines. Of note, there is no currently accepted validated risk tool for lung cancer for population screening.


  • They looked at 3 databases: the CXR-only wing of the NLST (2002-2009), the ever-smokers control group of the Prostate, Lung, Colorectal, and Ovarian cancer screening trial (PLCO, 1993-2009), and the National Health Interview Survey (NHIS, 1997-2001)
    • PLCO: 155K US men and women 55-74 yo had 4 annual CXRs and found no benefit from screening CXR in smokers
    • NLST randomized 53.5K smokers aged 55-74 with at least 30 pack-years smoking to LDCT vs CXR and found 20% decrease in lung cancer mortality by LDCT
    • NHIS: 87.5K people followed in annual cross-sectional US group from 2004, with linkage to the National Death Index. This database from 1997-2001 was used to validate the lung cancer death model. Then the model was applied to a more contemporary US population (NHIS 2010-2012) looking at all ever-smokers aged 50-80, which included 18,643 people, 52% male, 72% white/15% black/9% Hispanic, 55% post-high school education, 32% with BMI>30, 36% current smoker and 26% <10 pack-yrs, 21% 10-20, 15% 20-30, 14% 30-40, 24% >40; 67% had quit >15 years; 98% no family history, 7% had emphysema.
  • They assessed an array of characteristics from these databases (including age; education; sex; race; smoking intensity, duration, and quit-years; BMI; family history of lung cancer; self-reported symptoms of emphysema) to develop risk-based models for lung cancer incidence and deaths, then applied that to the later NHIS database.


  • Hazard Ratios for lung cancer incidence (will only mention the really significant ones)
    • Age, very highly correlated with lung cancer incidence (HR 80) and lung cancer death (HR 432)
    • Pack-years smoking:
      • 30-40: HR 1.63 for lung cancer incidence; 1.74 for lung cancer death
    • 90% of the CT-preventable lung cancer deaths are likely preventable by screening only 49% of US ever-smokers aged 50-80
    • One remarkable finding: in the risk-based model to USPSTF, 36% of the USPSTF-eligible smokers would not be screened (5-year lung cancer risk, 1.3%; NNS, 647), and would be replaced by 36% high-risk smokers (5-year lung cancer risk, 3.2%; NNS, 226) who did not meet the USPSTF criteria (mostly because they were African-American, lower BMI, less educated; 22% smoked <30 pack-years but tended to be longer-term smokers (>45 years, but 61% smoked <1/2  pack-per-day), and 14% quit > 15 years ago but were high intensity smokers, almost all having >30 pack-years and 53% >45 pack-years ).e., following the USPSTF recommendations, instead of an individual risk-based approach (incorporating more than just the smoking history), would both over-screen many low-risk people and not screen some of the high risk ones.
  • The lung cancer incidence model was validated by the chest x-ray groups of the NLS and PLCO of ever-smokers; the lung cancer death model was validated in the 1997-2001 NHIS and in the PLCO x-ray group of ever-smokers
  • Lung cancer mortality, by this model, was 24% lower than expected in the NLST x-ray group
  • Based on the NHIS 2010-2012 data, there was an estimated 43.4 million ever-smokers aged 50-80 in the US.
  • Screening the 9 million ever-smokers eligible by the USPSTF criteria for LDCT: the estimate was to prevent 46,488 deaths over 5 years
  • But screening 9 million of the highest risk group by the risk-based population would prevent 55,717 deaths (9229 more)
    • So, in this risk-based model NNS (number-needed-to-screen to prevent a death) was 162 vs 194 with USPSTF, with fewer false-positives (116 vs 133), all with p<0.001
    • And if one used the USPSTF NNS of 194 and applied that to the highest risk of the risk-based group (with risk >1.9%) [i.e., increasing the screening by the risk-based analysis to equal the NNS from USPSTF], then 3.1 million more people would be screened and then 62,382 deaths would be prevented. This increase to 12.1 million would also not have an increase in numbers of false positives.


  • So, this study did a good job developing risk models and validating them through several US cohorts, thereby suggesting that they are pretty robust and transportable. The models seemed to be more efficient than the USPSTF recommendations in terms of identifying higher risk individuals, decreasing lung cancer incidence and mortality, and decreasing the false-positive rates from LDCTs.
  • In the NLST itself, 88% of CT-prevented lung cancer deaths occurred in the 60% of those at highest risk and there were 64% with false positive results; only 1% of the lung cancer deaths occurred in the 20% at the lowest risk (see Kovalchik SA. N Engl J Med 2013; 369: 245). In fact, if one looks at those with normal LDCT in the initial NSLT, there was a dramatically lower risk of lung cancer development or death (see blog at end), to the point that I am strongly considering stopping LDCT after a couple of negative scans.
  • A few limitations of the risk-based study:
    • There is the assumption that the 20% decreased mortality in NSLT-eligible would apply to the NSLT-ineligible that would be included in the risk-based strategy.
    • The NHIS cohort only has data on lung cancer mortality not incidence, so the above risk-based model for mortality was validated only by the 2 research studies (PLCO and NLST x-ray only groups)
    • Confining the LDCTs to the high risk only, by using the risk-model, may be associated with more complications from procedures and surgery (these people are generally at higher surgical risk), so might distort the risk:benefit calculations from NLST
  • The risk-based approach does have some improvement in decreasing lung cancer deaths from 46488/9 million screened to 55717/9 million, which is increasing absolute numbers from 5.2 to 6.1/1000 screened. Still pretty small numbers.
  • So, as per my prior blogs on the USPSTF recommendations, I think they way-over interpreted NLST by extending a 3 year study (with decreasing pickup of incident lung cancers by the 3rd year), to the potential for 25 years of annual screening and the attendant high radiation exposure, to extending the upper age limit from 74 in NLST to 80, and then to codifying a single but good study into routine practice though the absolute risk reduction was relatively small (62 deaths per 100,000 person-years). The current individual risk-based approach suggests that there are many high risk patients who would not be screened by the USPSTF criteria, and many on the USPSTF list who are actually quite low risk and likely do not benefit much from the screening. I hope (and sort of expect) that the current USPSTF guidelines will be reviewed and reconsidered at some point in the next couple of years… And, it is important to remember that, as pointed out in the other blogs, lung cancer is not even close to being the primary killer associated with smoking (heart disease from smoking being much more prevalent, and COPD, etc. rating pretty high as well). And focusing on lung cancer screening may dilute the bigger message (i.e., it would be pretty awful if patients who had a normal LDCT felt that smoking was not really so bad for them, and it was okay to continue smoking…) see older blogs below for more on this.

See: which analyzes a retrospective study from NLST finding that those with an initial normal LDCT had about 35% lower rates of lung cancer incidence and mortality is a critique of NLST, especially its perhaps overenthusiastic acceptance and extensions by USPSTF

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