To screen or not to screen

US Preventive Services Task Force's draft guidelines add to prostate cancer screening debate

Authors

  • Carrie Printz


Editor's note: This is Part 1 of a 2-part series.

When the US Preventive Services Task Force (USPSTF) issued its new draft guidelines on screening for prostate cancer last fall, they added further controversy to the ongoing debate over the effectiveness of screenings.1

“I was considered a radical until October 2011,” says Otis Brawley, MD, chief medical officer of the American Cancer Society (ACS). “The task force turned me into a moderate.”

The ACS guidelines state that men should make an “informed decision” with their physician about whether to undergo prostate-specific antigen (PSA) screening, because research has not yet proven that the potential benefits of screening outweigh the potential harms of screening and treatment. Dr. Brawley says he “mildly disagrees” with the way the USPSTF guidelines were written, but that he understands what they were attempting to say. “I think the way they worded their statement comes across stronger than they intended,” he says. “If you read their entire statement, they recommend against the ‘routine’ use of PSA [screening]. There are people doing mass screenings in vans, malls, and health fairs to get patients to facilities that treat prostate cancer. We have little data that this is good for patients.”

The Trials

The USPSTF, which is sponsored by the Agency for Healthcare Research and Quality (AHRQ), is the leading independent panel of private-sector experts in prevention and primary care. Its members conduct rigorous, impartial assessments of the scientific evidence for the effectiveness of a broad range of clinical preventive services, including screening, counseling, and preventive medications.

We did a systematic review of harms and concluded that the benefit of the screenings was small to none.—Michael LeFevre, MD, MSPH

Illustration 1.

Their draft guidelines, which were published in the Annals of Internal Medicine, were updated based on new findings from 2 large screening trials: the US-based Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial and the European Randomized Study of Screening for Prostate Cancer (ERSPC).

After considering an extensive literature review of these and other studies considered to be less strong, the panel issued new guidelines, giving a “D” recommendation to PSA screening for men younger than 75 years. Their prior guidelines stated that they did not have sufficient evidence to determine whether this population should undergo the screenings (screenings are not recommended for men older than 75 years).

“We don't think the benefits outweigh the harms,” says Michael LeFevre, MD, MSPH, USPSTF co-vice chair. “We did a systematic review of harms and concluded that the benefit of the screenings was small to none.”

The major problem the USPSTF found with PSA screening was the risk of overdiagnosis and overtreatment. Of 1000 men treated with prostatectomy, 5 will die of postoperative complications, between 10 and 70 will have a serious complication such as heart attack or stroke, and 200 to 300 will experience side effects such as erectile dysfunction, urinary incontinence, or both, Dr. LeFevre says.

New to the guideline process is a public comment period, during which the task force received more than 2000 responses. At press time, task force members were processing these responses and were expected to issue their final recommendations, which are also likely to be published in the Annals of Internal Medicine. Dr. LeFevre anticipates publication sometime in late summer or early fall 2012.

Mixed Reaction

Whether these guidelines are a step in the right direction or a step backward depends on whom you ask and how they interpret the 2 main studies considered by the USPSTF. In reviewing what they termed the 2 “fair quality” studies, task force members found that PSA screening was associated with no reduction in prostate cancer–specific or all-cause mortality in the PLCO trial and with an absolute risk reduction in mortality of 0.07% in the ERSPC trial. (See “Two Large Trials Show Prostate Screening Has Limited Reduction in Mortality.”) The latter study concluded that 48 men must be treated to prevent 1 prostate cancer–specific death.

Clinical trials on prostate cancer screenings have limitations, so you have to do modeling, and some people think modeling takes a leap of faith.—Ruth Etzioni, PhD

Illustration 2.

Several experts interviewed by “CancerScope” agreed that both the PLCO and ERSPC studies have flaws, although the degree to which they are flawed and what the results actually indicate are still subjects of debate.

William Catalona, MD, director of the clinical prostate cancer program at Northwestern University in Chicago, Illinois, and a pioneer of PSA screening who discloses he has received funding from companies that manufacture PSA assays, believes the PLCO study is deeply flawed. Among the many flaws he notes are:

  • Before the trial, at least 43% of the participants had undergone screening, and 4% to 5% previously underwent biopsy, removing some high-risk tumors from the population.

  • The greatest flaw was that at least 52% of controls were screened during their 6 years in the trial, causing the data to be contaminated.

  • The PLCO protocol, involving screening for 6 years with PSA using a ≥4 ng/mL cutoff plus digital rectal exam for 4 years, does not represent how screening has evolved today, when many more assessment tools factor into treatment decision making, including measuring baseline PSA at a young age, PSA velocity, percentage-free PSA, and other risk assessments.

  • Only 4% to 5% of African Americans were included in the study despite the fact that they have a 50% higher incidence rate of the disease.

Dr. Catalona asserts that the ERSPC and Göteborg study (a Swedish trial included in the ERSPC that showed an even greater benefit to screenings) are far better studies and had less contamination. “The ERSPC is the biggest population trial and the Göteborg study, the best, was population-based,” he says. “The task force gave much more weight to PLCO despite the fact that it is much more flawed, and when they combined the good and bad trials, it diluted the mortality benefits from the good trials.”

What we all should be realizing is that we have to make a guess. While we know that screening finds cancer, we also know it finds cancer that doesn't need to be treated.—Otis Brawley, MD

Illustration 3.

However, Dr. LeFevre says that although the contamination of the PLCO trial likely affected the results, it would not have been enough to reduce the benefits of screening to zero, which is what the study found. He also had some criticism of the ERSPC, noting that there were some treatment differentials and preconsent problems. “Even if the benefit was as high as was shown in the European trial, we still believed the benefit of screening was small to none,” he says. “That trial found that per 10,000 men screened, you would save about 6 lives.”

Importance of Follow-Up

Ruth Etzioni, PhD, a biostatistician at Fred Hutchinson Cancer Research Center in Seattle, Washington, is regarded as an expert on the role of PSA testing, overdiagnosis, quality of life, and mortality. She notes that she looks at the guidelines “neutrally.” She says, “I have no reason to champion or criticize PSA screening, and I will not take positions. I'm just trying to figure out the numbers and use them to guide my recommendations.”

Dr. Etzioni says that the latest follow-up to the PLCO study, published in the Journal of the National Cancer Institute, showed that annual prostate cancer screening did not reduce the number of deaths from the disease, even among men in their 50s and 60s and in those with underlying cancer conditions.2 The ERSPC trial has not yet published updated results, she says.

The ERSPC trial combined multiple trials, many with different start times, protocols, number of screenings, and length of screening intervals, whereas the PLCO trial was a true, multisite trial, she says, adding that this is why the USPSTF found the ERSPC study less rigorous.

Although the ERSPC found 20% fewer deaths in the screened arm, Dr. Etzioni notes 2 caveats: First, the studies showing this benefit were not based on the entire study sample—the 20% reduction was observed in the 55- to 69-year-old age group. Second, there were very few deaths in this study. “If you look at the fraction of men who actually had their lives saved, it was … less than 1 in 1000,” she says.

Dr. Etzioni notes, however, that scientists currently have poor estimates of overdiagnosis. She believes the ERSPC's 48:1 ratio is too high, and that the US overdiagnosis estimate is closer to about 25%. “The harms are too high, and the number of lives saved depends on how long you have follow-up in the trials,” she says. “That 0.7 per 1000 lives saved figure will increase with time. When we correct for that, it's at least half of the 48:1 benefit ratio.”

The ERSPC had 9 years of follow-up, which is not long enough in prostate cancer, she says, adding that more lives would be saved if the studies had 20 to 25 years of follow-up. “Clinical trials on prostate cancer screenings have limitations, so you have to do modeling, and some people think modeling takes a leap of faith,” Dr. Etzioni says.

Meanwhile, Dr. Brawley weighs in on criticism of all the studies. “Every study ever done is flawed,” he says. “What we all should be realizing is that we have to make a guess. While we know that screening finds cancer, we also know it finds cancer that doesn't need to be treated.”

He adds that the PLCO study was “sabotaged” by proponents of screening and that men participating in the study had to pass advertisements saying that screening saves lives, when no legitimate study reaching that conclusion had been published. “PLCO still has statistical power,” he says. “It's flawed, but not so flawed that you can't learn anything.”

Part 2 of this series, which will be published in the May 1 issue, will explore the drawbacks of screening and the impact of the guidelines.

Ancillary