POEMs

				Using the search form

To view this article, including Web-exclusive content, in PDF form, click here.

To view only this article's Web-exclusive content (in PDF form), click here.

Combined ultrasound and mammogram for high-risk women increases cancer yield

Clinical question Is mammography plus ultrasound more beneficial for breast cancer detection in high-risk women than mammography alone?

Bottom line In this high-risk group of women, the combination of mammography plus ultrasound resulted in the detection of more cases of cancer (32 of 40) than mammography alone (20 of 40). Mammography plus ultrasound also resulted in significantly more recommendations for biopsy (84 vs 276), possibly resulting in both physical and emotional harm to those women with false positive screening test results. Because the prevalence of breast cancer is much lower in the general population of women aged 40 to 50 years than in this high-risk group (0.04% vs 1.5%, respectively), we need more patient-oriented evidence of benefit before routinely recommending both tests for all women in this age group. (Level of evidence = 1b)

Synopsis More than half of the study participants were women previously treated for breast cancer; the other half consisted of women with at least one or more family members with breast cancer. A total of 2,637 women underwent both mammography and ultrasound screening examinations in randomized order, with the interpreting radiologists for each separate exam masked to the results of the other. If either the mammogram or ultrasound results alone were concerning, another radiologist made an integrated interpretation of both studies together. Biopsy-proven cancer diagnosed within 365 days of the initial screening examination served as the gold reference standard. Forty (1.5%) participants had cancer, including six with ductal carcinoma in-situ. The test characteristics were as follows: Mammogram alone (sensitivity = 50%; specificity = 95.5%; positive predictive value [PPV] = 14.7%; negative predictive value [NPV] = 99.2%) and mammogram plus ultrasound (sensitivity = 77.5%; specificity = 89.4%; PPV = 10.1%; NPV = 99.6%). What this really means is that mammography alone will result in eight missed cancers (false negatives) for every 1,000 cases reported as normal. Likewise, mammography plus ultrasound will result in four false negatives. Mammography alone led to a recommendation for biopsy in 84 women, whereas mammogram plus ultrasound led to 276 biopsies. The author of an accompanying editorial references a recent article on the psychological impact of a false-positive screening diagnosis concluding that women who are recalled do not appear to be harmed by screening (J Clin Oncol. 2007;25[25]: 3823-3830). However, no mention is made of a previous study reporting a significant loss in the quality of life in women called back because of false positive mammogram results (Ann Intern Med. 1991;114[8]:657-661).

Berg WA, Blume JD, Cormack JB, et al; ACRIN 6666 Investigators. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA. 2008;299(18):2151-2163.

Accelerated care decreases length of stay in joint replacement surgery

Clinical question Does accelerated perioperative care decrease length of stay and affect quality of life in patients undergoing primary hip or knee arthroplasty?

Bottom line In this study, a program of aggressive symptom control and mobilization reduced hospital length of stay by 3 days compared with usual care. The rate of complications was low in both groups. (Level of evidence = 2b)

Synopsis Danish patients undergoing primary hip arthroplasty or knee replacement (total or unicompartmental) were randomly assigned to receive usual care (n = 45) or to intervention (n = 45). The intervention consisted of an orientation to the perioperative program, admission to a new unit on the day of surgery, and specific goals regarding information, control of pain and nausea, nutrition, mobilization, and elimination. And the patients could wear their own clothes during the whole stay. Although this was an unblinded study, the researchers used approaches to reduce potential biases resulting from this. They used explicit discharge criteria that were applied to both groups, they isolated the patients and staff to avoid contamination, they did not use surgeons who were part of the research team, and they used written questionnaires to assess quality of life. The average length of stay in the intervention group was 5 days compared with 8 days in the control group. Although the quality of life was statistically significantly improved in the accelerated care patients, the differences were small and of uncertain clinical meaning. In the control group, one patient died and one was readmitted for an additional 15 days. In the intervention group, no patients died, and two were readmitted (for 1 and 11 additional days).

Larsen K, Sørensen OG, Hansen TB, et al. Accelerated perioperative care and rehabilitation intervention for hip and knee replacement is effective: a randomized clinical trial involving 87 patients with 3 months of follow-up. Acta Orthop. 2008;79(2):149-159.

Wait-and-see approach to AOM treatment decreases antibiotic use

Clinical question Does a wait-and-see approach to the treatment of acute otitis media (AOM) decrease antibiotic use when a prescription is not given?

Bottom line A wait-and-see approach to acute otitis media—when a prescription for antibiotics is withheld for 2 or 3 days—produced less antibiotic use than giving parents a prescription to be filled should symptoms worsen or not resolve. Only approximately 1 in 5 children in the wait-and-see group were treated with antibiotics compared with almost half of the other group. Just as important, parents were equally satisfied with both approaches. (Level of evidence = 1b–)

Synopsis The authors conducting this study enrolled 232 children, aged 2 to 12 years, to receive delayed treatment of typical AOM. The study was conducted in a pediatric emergency department. The children were randomized (allocation concealment uncertain) to 1 of 2 groups. In one group, parents of patients were given a prescription and asked to not have it filled for 2 days; in the other group, patients were asked to return if their symptoms did not resolve in 2 or 3 days. All patients were given complimentary bottles of a topical anesthetic and an oral analgesic. Of the children for whom a prescription was given, 46% used a prescription antibiotic within 7 to 10 days of the initial visit, whereas 19% of the children in the wait-and-see group used the antibiotics (P < .01). More than 90% of parents in both groups were either very satisfied or extremely satisfied with their visit.

Chao JH, Kunkov S, Reyes LB, et al. Comparison of two approaches to observation therapy for acute otitis media in the emergency department. Pediatrics. 2008;121(5):e1352-e1356.

High random glucose level is a weak predictor of diabetes

Clinical question Does an elevated random glucose level identify patients who have diabetes?

Bottom line An elevated random plasma glucose level is somewhat useful for identifying patients who will subsequently be diagnosed as having glucose intolerance. It cannot make a diagnosis of diabetes in itself; although more than 60% of patients with a random glucose level as high as 140 mg/dL (7.7 mmol/L) will be confirmed to have glucose intolerance, only 1 in 4 will have diabetes. (Level of evidence = 1c)

Synopsis To answer the question of whether a random plasma glucose determination is a useful screen for diabetes in adults, the researchers enrolled 990 people without known diabetes recruited from a community. The study was called “Screening for Impaired Glucose Tolerance” and may have attracted a group of people at higher-than-normal risk for diabetes; 24% were subsequently given a diagnosis of glucose intolerance or frank diabetes (5% of the total group). The patients—54% black, 66% women, with an average age of 48 years, and a body mass index of 30.4 kg/m²—had a random plasma glucose determination and then, within 3 weeks, had a 75-g oral glucose tolerance test following an overnight fast. The specificity of a random glucose test for diabetes was 93% at 125 mg/dL (6.94 mmol/L), increasing only slightly with higher levels. However, given the low prevalence of diabetes in this population, the positive predictive value was only 22%. When setting the bar considerably lower—to identify any patient with any glucose intolerance—the specificity ranged from 85% at 110 mg/dL (6.10 mmol/L) to 98% at a cutoff of 140 mg/dL (7.77 mmol/L). Even given the high rate of glucose intolerance in this group, these specificities translate into only 47% to 59% of patients being correctly identified (positive predictive values = 47% at 110 mg/dL and 59% at 140 mg/dL).

Ziemer DC, Kolm P, Foster JK, et al. Random plasma glucose in serendipitous screening for glucose intolerance: screening for impaired glucose tolerance study 2. J Gen Intern Med. 2008;23(5):528-535.

Hypertension treatment is effective even in patients older than 80 years

Clinical question Does the treatment of hypertension in persons older than 80 years improve clinical outcomes?

Bottom line Treatment of hypertension in the very elderly reduces the risk of fatal stroke and death from any cause. Previous studies using high-dose diuretics and beta-blockers had not found a similar benefit, perhaps because of the adverse effects of high-dose diuretics and the lack of benefit of beta-blockers. (Level of evidence = 1b)

Synopsis Data regarding the benefit of treating hypertension in the very elderly are sparse and mixed. Although some studies have shown a reduced risk of stroke, there is also data suggesting an increase in all-cause mortality, especially with target systolic BPs below 140 mm Hg. In this study, 3,845 patients older than 80 years with a systolic BP between 160 and 199 mm Hg without medication were identified. The patients were a mix of those with systolic hypertension and systolic/diastolic hypertension. They were randomly assigned to receive sustained-release indapamide 1.5 mg daily or placebo. Patients with recent stroke, secondary or accelerated hypertension, heart failure, or renal impairment were excluded. If the target BP of 150/80 mm Hg was not achieved, perindopril (2 mg or 4 mg) or matching placebo could be added. Appoximately 25% of active treatment patients were receiving indapamide alone at the end of the study; the rest were receiving indapamide and perindopril. The mean duration of follow-up was 2.1 years, with a range of 0 to 6.5 years. Patients in the active treatment group had lower rates of fatal stroke (absolute risk reduction [ARR] 0.42%/year, P = .046, number needed to treat [NNT] = 240/year), all-cause mortality (ARR 1.2%, P = .02, NNT = 80/year), heart failure (ARR 0.95%, P < .001, NNT = 105), and any cardiovascular event (ARR 1.7%, P < .001, NNT = 59). There were fewer serious adverse events in the active treatment group as well.

Beckett NS, Peters R, Fletcher AE, et al; HYVET Study Group. Treatment of hypertension in patients 80 years of age or older. N Engl J Med. 2008;358(18):1887-1898.

Carotid bruits are associated with acute MI and CV death

Clinical question What is the clinical significance of carotid bruits?

Bottom line Patients with carotid bruits are at an increased risk of MI and death from cardiovascular (CV) causes. (Level of evidence = 1a–)

Synopsis These researchers searched MEDLINE and EMBASE to identify studies that prospectively assessed outcomes associated with carotid bruits. They specifically only looked for published studies. Two members of the team independently assessed the quality of the studies and two others independently extracted the data from the studies. They included 22 studies with more than 17,000 patients followed up for more than 60,000 patient-years. Most of the studies (14) evaluated patients with asymptomatic bruits. Among patients with bruits, the rate of MI was 3.69 per 100 patient-years (95% CI, 2.97-5.40) compared with 1.86 (0.24-3.48) in those without bruits. Among patients with bruits, the rate of cardiovascular death was 2.85 per 100 patient-years (2.16-3.54) compared with 1.11 (0.45-1.76) in those without bruits. In spite of the authors’ pronouncement, there are no data to suggest that aggressive evaluation, risk factor modification, or interventions based on the presence or absence of bruits will delay these events.

Pickett CA, Jackson JL, Hemann BA, Atwood JE. Carotid bruits as a prognostic indicator of cardiovascular death and myocardial infarction: a meta-analysis. Lancet. 2008;371(9624):1587-1594.