Monday, October 29, 2018

PSA screening: USPSTF recommendations changed, but the evidence did not

Comparing the 2018 U.S. Preventive Services Task Force (USPSTF) recommendation statement on prostate cancer screening in the October 15th issue of American Family Physician with its previous recommendation, the first question family physicians ought to ask is: what new evidence compelled the USPSTF to move from recommending against PSA screening in all men to determining that there was a small net benefit for screening in some men? Did another major randomized trial show a reduction in all-cause or prostate cancer-specific mortality in men invited to screening? Did other systematic reviewers re-analyze the evidence and find a mortality benefit where none previously existed? Have urologists or radiation oncologists developed new treatments for localized prostate cancer that no longer cause erectile dysfunction, urinary incontinence, or infections?

No, no, and no.

One of the Top 20 Research Studies of 2017 for Primary Care Physicians, the only U.S. trial of PSA-based screening for prostate cancer, reported that after a median followup of 15 years, there were still no differences in mortality between the two groups. In 2018, a large U.K. randomized trial of a single PSA screening also reported no effect on prostate cancer mortality after a median followup of 10 years. In both trials, more prostate cancers were diagnosed in the groups assigned to routine screening, but treating these cancers did not lead to improved health outcomes.

Last month, the authors of a 2010 Cochrane review of PSA screening (previously summarized in AFP's Cochrane for Clinicians) published an updated meta-analysis in the BMJ that incorporated the U.K. trial findings and extended followup of the U.S. and European screening trials and concluded that "at best, screening for prostate cancer leads to a small reduction in disease-specific mortality over 10 years but does not affect overall mortality." They also estimated that "for every 1000 men screened, approximately 1, 3, and 25 more men would be hospitalized for sepsis, require pads for urinary incontinence, and report erectile dysfunction, respectively." Another U.K. trial comparing active surveillance for localized prostate cancer with immediate surgery or radiation therapy found higher rates of clinical progression in the active surveillance group, but no differences in health-related quality of life or mortality.

Representing the views of American Academy of Family Physicians (AAFP), Drs. James Stevermer and Kenneth Fink explained in an editorial why "the AAFP believes that the net benefit [of PSA screening] does not justify routine screening or routinely offering shared decision making." The AAFP took the unusual step of declining to endorse the USPSTF recommendation statement and instead writing its own clinical preventive services recommendation that emphasizes the harms of routine screening. Men who bring up the topic of PSA screening should engage in shared decision-making with their physicians about the benefits and harms of screening and express a clear preference to be screened before undergoing the test.


This post first appeared on the AFP Community Blog.

Monday, October 15, 2018

We should be alarmed that congenital syphilis is on the rise

The Tuskegee Syphilis Study will always be a black mark on American medicine. Designed to record the natural history of untreated syphilis in a population of illiterate African American men in rural Alabama, it continued for 25 years after penicillin became widely available and an accepted cure for the disease. Participants were deceived into believing that they were receiving effective treatment and never informed that they had the option of leaving the study. Although the ethics of these practices were occasionally questioned, the study did not end until it had received widespread negative publicity and Congressional hearings had been called.

In 1972, the last year of the Tuskegee study, I hadn't been born yet. But when I was a medical student doing my Internal Medicine clerkship in early 2000, an older African American man was admitted to our hospital service with memory loss, confusion and difficulty walking and was eventually diagnosed with tertiary neurosyphilis. Based on what he told me about his sexual history and prior symptoms (genital ulcers, rashes, muscle and joint pains), he had likely been infected decades before. Though he was a U.S. citizen by birth, he did not have a family physician and had rarely received health care due to not having insurance or the ability to pay. I don't remember what happened after we started penicillin, but much of the damage already done to his nervous system was irreversible. This patient's illness wasn't quite as outrageous as doctors knowingly withholding antibiotics for decades, but I have little doubt that it could have turned out differently if he had had access to primary care earlier in his life.

Last month, the Centers for Disease Control and Prevention (CDC) announced that the number of reported cases of congenital syphilis in the U.S. rose from 362 in 2013 to 918 in 2017, paralleling increases in syphilis infections in reproductive-age women during this time period. From 2016 to 2017, congenital syphilis cases rose from 16 to 23 per 100,000 live births. Although two-thirds of affected infants have no symptoms at birth, congenital syphilis is associated with increased neonatal mortality and a variety of early (through 48 months of age) and late complications.

The first line of prevention against congenital syphilis is screening for syphilis in all pregnant women at the first prenatal visit, a well-established standard of care that the U.S. Preventive Services Task Force (USPSTF) recently reaffirmed. Although some cases occur in infants whose mothers receive no prenatal care, about one-third of women who delivered a baby with congenital syphilis in 2016 were screened during their pregnancies.

The CDC, the American Academy of Pediatrics, and the American College of Obstetricians and Gynecologists all recommend repeating syphilis screening in women at high risk for syphilis at around 28 weeks of gestation and at time of delivery. Women at high risk include those living in higher-prevalence communities or geographic areas; those living with HIV infection; those with a history of incarceration or commercial sex work; and those exposed to a sexual partner with confirmed syphilis infection. Early penicillin treatment of infected pregnant women reduces the risk of congenital syphilis.

It is alarming that nearly a thousand American babies born last year were afflicted with congenital syphilis. Medically speaking, preventing this condition through detection and treatment of syphilis is straightforward. This is a population health failure, resulting from an underfunded public health infrastructure and a fragmented health system that makes it hard for women at high risk to access timely prenatal and primary care. Our country can do much better.


Parts of this post first appeared on the AFP Community Blog.

Saturday, October 6, 2018

What works and what doesn't for chronic sleeplessness

Can't sleep? Then spend your extra awake time reading the latest installment of Implementing AHRQ Effective Care Reviews in the September 1 issue of American Family Physician, on management of insomnia disorder in adults. This evidence review, which supported an American College of Physicians practice guideline, examined the effectiveness of behavioral therapies and medications for adults with insomnia disorder, defined as "poor sleep quality or quantity that causes distress or dysfunction and lasts for longer than three months."

The most beneficial sleep intervention overall is cognitive behavior therapy for insomnia (CBT-I), which produced sustained improvements for at least 6 months. CBT-I consists of cognitive therapy, sleep restriction and stimulus control, and sleep hygiene education. Medications that have sufficient evidence demonstrating improvement in short-term (3 months or less) sleep outcomes include eszopiclone, zolpidem, and suvorexant; there was insufficient data to evaluate benzodiazepines or over-the-counter sleep aids (diphenhydramine, doxylamine, or melatonin). For most patients, medications should not be prescribed for longer than five weeks.

Physicians commonly prescribe antipsychotic medications off-label to treat insomnia in older persons. The Practice Guidelines in the September 15 issue summarized a Canadian guideline for deprescribing antipsychotics for behavioral and psychological symptoms of dementia and insomnia, produced by the Deprescribing Guidelines in the Elderly Project. Due to the potential harms of these medications and the lack of evidence of benefits (a single randomized trial with 13 participants found nonsignificant differences in sleep latency in patients taking quetiapine), the guideline recommends that antipsychotics prescribed for primary or secondary insomnia in which comorbidities are under control be discontinued without tapering, regardless of treatment duration.

AFP's sister journal, FPM, recently published an article on deprescribing unnecessary medications that featured a four-step process (review current medications; identify inappropriate, unnecessary, or harmful medications; plan deprescribing with the patient; and regularly re-review medications) and links to additional resources on medication reconciliation and deprescribing.


This post first appeared on the AFP Community Blog.

Monday, October 1, 2018

Setbacks and a milestone for evidence-based medicine

For my students in 2018, it's difficult to imagine an era when there was no such thing as evidence-based medicine (EBM). When I started medical school in 1997, though, the term had only been in use for six years, having been introduced by Dr. Gordon Guyatt at McMaster University (though EBM's intellectual origins can be traced to several key individuals). When I tell trainees how recently EBM began, they often ask, "Well, then, what kind of medicine did physicians practice before?" The answer is, we largely practiced "eminence-based" medicine (but this tongue-in-cheek article offered some equally poor alternatives).

Although it may be well established, the primacy of EBM is more fragile than it seems. In the September 15 issue of American Family Physician, my longtime mentor and editor emeritus Jay Siwek, MD reviewed common misconceptions, barriers, and practical solutions. For example, evidence can be distorted by financial conflicts, misinterpreted though the lens of one's preexisting beliefs, or ignored by those who deride evidence-based practice guidelines (incorrectly) as "cookbook," "one-size-fits-all" medicine. A recent essay in BMJ also described threats to evidence-informed policy making driven by ideological decisions on both ends of the political spectrum:

We tend to alight on examples of harmful interventions that fit our own political preferences. For example, ... public health researchers leaning towards the left might cite evidence that abstinence only sex education is more likely to lead to increased sexual risk behavior than comprehensive sex education. ... But only referring to examples where the evidence aligns with our own preferences risks suggesting to those on the left that they do not need evidence to know what does not work (as it is just obvious), and to those on the right that evidence informed policy is a liberal conspiracy.

EBM has experienced serious setbacks in the past few months. One, which I discussed previously, was the shuttering of the National Guideline Clearinghouse, a vital repository of evidence-based guidelines that was maintained by the Agency for Healthcare Research and Quality. Another is the Sept. 14 resignation of health services researcher Dr. H. Gilbert (Gil) Welch from Dartmouth College. Described as "an internationally recognized expert on the effects of medical screening and overdiagnosis" in his official Dartmouth biography, Welch literally wrote the book on overdiagnosis, which makes the university's determination that he plagiarized ideas in a 2016 paper puzzling at best. Welch resigned in response to Dartmouth's demands that he make his accuser the first author on the paper and stop teaching at the school. He has denied the plagiarism accusation, and the New England Journal of Medicine has declined to retract the article, viewing it as an authorship dispute rather than a breach of ethics. Although I hope that Dr. Welch's work will continue, his resignation is not only a huge loss for Dartmouth, but for all who have followed and benefited from his seminal work on the downsides of screening, including me.

On the same day as Dr. Welch's resignation, another giant in evidence-based medicine, Dr. Peter Gotzsche, was unceremoniously dismissed from the governing board of the Cochrane Collaboration and expelled from membership in the group. Previously the director of the Nordic Cochrane Center and author of 17 Cochrane reviews, Dr. Gotzsche called his explusion from the organization emblematic of a "moral governance crisis" and accused Cochrane's executive team of sacrificing scientific rigor and open debate in a "growing top-down authoritarian culture and an increasingly commercial business model." Cochrane leadership had been annoyed by Dr. Gotzsche's co-authorship of a BMJ Evidence-Based Medicine critique of a Cochrane review of HPV vaccines, which some felt was overblown and might shake public confidence in the vaccine. But their action sent the chilling message - antithetical to the democratic values at the heart of EBM - that open dissent would not be tolerated.

Thankfully, the news isn't all bad for EBM. An article by fellow AFP Deputy Editor Mark Ebell and colleagues in the September Annals of Family Medicine celebrated the top 20 POEMs (summaries of studies of patient-oriented evidence that matters) of the past 20 years. Since 1998, this group has systematically reviewed more than 100 clinical medical journals for such studies. My EBM teaching favorites from the top POEMs list include those from 2002 (hormone replacement overall is not beneficial), 2009 (prostate-specific antigen screening does not reduce mortality from prostate cancer), and 2013 (fasting is not necessary before measuring lipid panels). Congratulations to the "POETs" for their past contributions, and for continuing to do the yeoman's work of bringing medical evidence to the point of care in primary care.