Do women who are screened for ovarian cancer have better health outcomes than women who are not screened?
In this study (with methods that are biased in favor of the intervention), women who are screened for ovarian cancer are unlikely to experience any mortality benefit. (LOE = 1b-)(www.essentialevidenceplus.com)
Jacobs IJ, Menon U, Ryan A, et al. Ovarian cancer screening and mortality in the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS): a randomised controlled trial. Lancet 2016;387(10022):945-956.
Study design: Randomized controlled trial (nonblinded)
Funding source: Government
In this British study, the researchers randomized more than 200,000 women between the ages of 50 years and 74 years to annual screening with pelvic ultrasound plus CA-125 testing, to ultrasound alone, or to no screening. Women were recruited through the National Health Service Trust registries. They used a modified intention-to-treat analysis to evaluate the rate of ovarian cancer mortality. The researchers followed the women up for a median of 11 years. At the end of that time, the rate of ovarian cancer detection was the same in each group (< 1%) and the ovarian cancer mortality rate was the same in each group (~ 0.3%). One could just stop there and conclude that screening is ineffective, as did the authors of the Prostate, Lung, Colorectal, Ovarian Cancer (PLCO) trial. Not these researchers. My guess is that they either had an agenda to push or were too embarrassed by spending 15 years and millions of dollars on a negative study. They went back and selectively excluded women who actually had oophorectomies or ovarian cancer and those who had exited the registry before randomization. Turns out these women were disproportionately distributed into the control group. Finally, after statistical modelling, they found that there might be a delayed benefit in ovarian cancer mortality that is not evident until after 8 years. So, here we have a study with methods that stack the deck in favor of interventions (modified intention to treat, lack of masking, postrandomization exclusions, and so forth) that finds no statistical differences in outcomes. But after complex modelling and other data manipulations they find a possible delayed screening benefit on ovarian cancer mortality. No amount of data mining can overcome these biases. Oh, by the way-the authors report total mortality on page 23 of an appendix: no difference.
Henry C. Barry, MD, MS
Michigan State University
East Lansing, MI
How effective is screening for lung cancer?
Screening current or former smokers for lung cancer using chest radiography has no effect on lung cancer mortality or all-cause mortality. Screening with low-dose computed tomography reduces lung cancer mortality and all-cause mortality. However, the harms of screening are significant. (LOE = 1a)(www.essentialevidenceplus.com)
Usman Ali M, Miller J, Peirson L, et al. Screening for lung cancer: A systematic review and meta-analysis. Prev Med 2016;89:301-314.
Study design: Systematic review
These authors searched several databases and data from a Cochrane Review to identify studies that evaluated various lung cancer screening modalities. The authors only included randomized trials to evaluate the potential benefits of screening, but they included any study design to evaluate the harms. Two members of the team independently evaluated articles for inclusion and also evaluated the methodologic quality of the included studies (ie, risk of bias). They resolved any disagreements through discussion or consultation with a third member of the team. They ultimately included 34 studies with thousands of patients, 13 randomized trials that assessed the potential benefits of screening and 31 that assessed the harms. Two of the trials were determined to be at high risk of bias, 4 were at low risk of bias, and the rest were "uncertain." The length of follow-up ranged from 1 year to 20 years. Most of the studies included only current or former smokers who were generally older than 40 years. Six of the trials also included only men. Seven trials evaluated chest radiographs alone or in combination with sputum cytology. These studies found no reduction in lung cancer mortality or all-cause mortality compared with usual care. Three relatively small, low-quality studies evaluated annual low-dose computed tomography (LDCT) with less than 10 years of follow up. Compared with usual care, these studies resulted in no reduction in lung cancer mortality or all-cause mortality. One large study, the National Lung Screening Trial, compared 3 annual LDCT screens with chest radiographs and found a 0.33% absolute reduction in lung cancer mortality (number needed to screen [NNS] to prevent one death from lung cancer = 308; 95% CI 201 - 787). Additionally, this study found an absolute all-cause mortality reduction of 0.46% (NNS = 219; 115 - 5556). A few of the studies also reported that screened patients were slightly more likely to quit smoking or decrease tobacco consumption. Many of the harms of screening (eg, overdiagnosis, death, major complications of treatment, and false positive results and their consequences) were inconsistently reported. However, when reported, the rates were not inconsequential. For example, in one study, 23 of 778 patients died after undergoing invasive procedures after screening. The authors were unable to identify any studies reporting anxiety or the consequences of incidental findings.
In patients with a family history of colorectal cancer who do not develop early-onset colorectal cancer, what is the subsequent risk of developing the disease?
People who have one first-degree relative with colorectal cancer but have not developed the cancer by age 55 years have only a small increase in the risk of colorectal cancer compared with the general population. The risk increases, however, with the number of first-degree relatives with the disease. (LOE = 2b)(www.essentialevidenceplus.com)
Schoen RE, Razzak A, Yu KJ, et al. Incidence and mortality of colorectal cancer in individuals with a family history of colorectal cancer. Gastroenterology 2015;149(6):1438-1445.
Study design: Cohort (prospective)
The authors used data from a randomized trial of cancer screening, and included the 144,768 enrolled patients who were 55 to 74 years of age with no history of colorectal cancer. Of these, approximately 10% had a family history of the disease. There was an even split between the sexes and most patients (89%) were were white. All patients underwent sigmoidoscopy at the start of the study and again at either 3 years or 5 years, and then were sent a mailed questionnaire once a year, for a total of 13 years of follow-up. During screening and follow-up, 1.44% of patients had colorectal cancer identified. The risk of cancer was slightly higher in patients with a family history of colorectal cancer (16.5 per 10,000 patients/year vs 12.8 per 100,000 patients/year). The risk increased slightly with the number of first-degree relatives with cancer. The risk did not increase based on first-degree relative's age at diagnosis. Deaths due to colorectal cancer were slightly higher in patients with a first-degree relative (4.19 per 10,000 patients/year vs 3.2 per 100,000 patients/year).
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Does the American College of Cardiology/American Heart Association Pooled Cohort Risk equation accurately predict cardiovascular risk in a typical cohort of patients?
This risk calculator substantially overestimates the actual occurrence of atherosclerotic cardiovascular disease (ASCVD) events over 5 years in a typical multiethnic population. The risk calculator estimates 10-year risk, but the authors extrapolated the equation to 5 years and compared the results since the risk increase seems to be linear. As has been pointed out before, this frequently used risk calculator substantially overestimates the percentage of patients who will benefit from lipid-lowering treatment. (LOE = 1b)(www.essentialevidenceplus.com)
Rana JS, Tabada GH, Solomon MD, et al. Accuracy of the atherosclerotic cardiovascular risk equation in a large contemporary, multiethnic population. J Am Coll Cardiol 2016;6(18):2118-2130.
Study design: Decision rule (validation)
Funding source: Foundation
These US investigators identified a target population of 307,591 patients, 40 years to 75 years old, from a large integrated health care delivery system, for consideration of cholesterol-lowering therapy in 2008. The follow up lasted through 2013. The authors excluded people with known ASCVD, diabetes mellitus, low-density lipoprotein cholesterol levels of less than 70 mg/dL (1.8 mmol/L) or greater than 190 mg/dL (4.9 mmol/L), and those receiving lipid-lowering therapy. They also excluded from the results people with incomplete 5-year follow-up. Though the majority of the patients were white, 7.2% were black, 17.2 were Asian/Pacific Islander, and 6.1% were Hispanic. The average patient age was 54.8 years, 61.6% were women, 28% were obese, and 33% were treated with medication for hypertension. Using the Pooled Cohort Risk Equation, 31% would have qualified for lipid-lowering treatment (ie, a 10-year risk of at least 7.5%). Over 5 years, the actual incidence of ASCVD increased linearly each year but did not approach the rates calculated by the equation for the same 5-year period. In a separate analysis, risk calculations for patients with diabetes similarly overestimated risk. These results line up with the results from other studies that have attempting to validate the risk equation.
Can daily administration of cranberry capsules reduce the risk of bacteriuria plus pyuria and subsequent urinary tract infections among elderly women residing in nursing homes?
This study found no significant benefit to the daily administration of cranberry capsules to elderly women residing in nursing homes. In particular, there was no difference in the risk of bacteriuria plus pyuria, symptomatic urinary tract infections (UTIs), hospitalizations, or all-cause mortality. (LOE = 1b)(www.essentialevidenceplus.com)
Juthani-Mehta M, Van Ness PH, Bianco L, et al. Effect of cranberry capsules on bacteriuria plus pyuria among older women in nursing homes. A randomized clinical trial. JAMA 2016;316(18):1879-1887.
Study design: Randomized controlled trial (double-blinded)
Funding source: Industry + govt
Setting: Nursing home/extended care facility
Bacteriuria plus pyuria is very common among women residing in nursing homes. These investigators identified women, 65 years or older, currently residing in a nursing home with or without baseline bacteriuria plus pyuria, determined using standard diagnostic criteria, and no other clinically significant renal or urologic disease. Eligible patients (N = 185) randomly received (concealed allocation assignment) either 2 cranberry capsules or 2 matched placebo capsules daily. Clean-catch urine samples were obtained every 2 months for 12 months and assessed for bacteriuria plus pyuria. Secondary outcomes included symptomatic UTIs, hospitalizations, and all-cause mortality. Individuals masked to treatment group assignment assessed all reported outcomes. Complete follow-up occurred for 80.1% of patients at 12 months. Using intention-to-treat analysis, no significant differences occurred in the prevalence of bacteriuria plus pyuria in the treatment group versus control group (29.1% vs 29.0%, respectively). In addition, no significant differences occurred in symptomatic UTIs, UTIs with multidrug resistant organisms, hospitalizations, or all-cause mortality. The study was 80% powered to detect a predetermined clinically significant difference between the treatment group and the control group.
David Slawson, MD
Director of Information Sciences
University of Virginia Health System
POEMs are provided by Essential Evidence Plus, a point-of-care clinical decision support system published by Wiley-Blackwell. For more information, please see http://www.essentialevidenceplus.com(www.essentialevidenceplus.com).
Share this page
This page will be removed from your Favorites Links. Are you sure?
Top POEMs of 2016 Consistent with the Principles of the Choosing Wisely Campaign