American Family Physician's annual collection of the top 20 research studies for primary care addresses a variety of topics with the potential to change practice. A group of primary care clinicians with expertise in evidence-based practice rated these studies highly based on their clinical relevance, validity, and reported outcomes. Known as POEMs, for patient-oriented evidence that matters, the studies are organized by topic and summarized with a clinical question, bottom-line answer, and brief discussion. This collection includes the top POEMs consistent with the principles of the Choosing Wisely campaign.
See related AFP article, "Top POEMs of 2019 Consistent with the Principles of the Choosing Wisely Campaign."
Clinical question
In patients with an acute musculoskeletal injury that does not warrant parenteral analgesia, does adding ibuprofen and codeine to acetaminophen (paracetamol) increase pain relief?
Bottom line
In patients with moderate pain at rest due to an acute sprain, fracture, or contusion, adding ibuprofen and codeine to treatment with acetaminophen does not produce additional pain relief at 1 or 2 hours after a single dose. (LOE = 2b)
Reference
Gong J, Colligan M, Kirkpatrick C, Jones P. Oral paracetamol versus combination oral analgesics for acute musculoskeletal injuries. Ann Emerg Med 2019;74(4):521-529.
Study design: Randomized controlled trial (double-blinded)
Funding source: Foundation
Allocation: Concealed
Setting: Emergency department
Synopsis
These researchers recruited 119 patients who presented to an emergency department with an acute musculoskeletal injury (sprain, fracture, or contusion) with a pain score of at least 4 of a possible 10 (average 5.8) who were deemed by the treating physician not to require parenteral analgesia. The patients were randomly assigned to receive a single dose of 1000 mg acetaminophen, either with placebo or in combination with 400 mg ibuprofen and 60 mg codeine. Allocation to treatment was concealed from the clinician and researchers. Using intention-to-treat analysis, pain reduction at rest was similar in both groups after 1 hour: a decrease of 1.6 in the acetaminophen/placebo group and 2.0 in the combination analgesic group (difference = .4; P = .26). After 2 hours, pain reduction at rest was still similar with both treatments. Reduction in pain with activity was statistically greater with combination analgesia (3.0 vs 1.9; P = .04), but this difference, on average, did not meet the threshold for clinical relevance (a difference of 1.3). Almost 3 times as many (relative risk = 2.8) patients who received the combination reported a minor adverse effect than did patients who received acetaminophen alone. The study had a 90% power to find a difference at 60 minutes if one existed.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Are platelet-rich plasma injections beneficial in the nonoperative treatment of rotator cuff disease in adults?
Bottom line
This review found no evidence that supports any additional benefit of platelet-rich plasma (PRP) injections compared with various control interventions, including saline placebo, in the nonoperative treatment of rotator cuff disease in adults. Exercise therapy was shown to be superior to PRP injections in improving outcomes in the included studies. (LOE = 1a)
Reference
Hurley ET, Hannon CP, Pauzenberger L, Fat DL, Moran CJ, Mullett H. Nonoperative treatment of rotator cuff disease with platelet-rich plasma: A systematic review of randomized controlled trials. Arthroscopy 2019;35(5):1584-1591.
Study design: Systematic review
Funding source: Unknown/not stated
Setting: Various (meta-analysis)
Synopsis
The benefits of PRP injections in the management of various tendinopathies are inconsistent. These investigators searched MEDLINE, Embase, and the Cochrane Library for English-language-only randomized controlled trials that compared PRP injections with control treatments in adults with chronic rotator cuff disease. Two reviewers used a standard evaluation tool to independently analyze individual articles for inclusion criteria and risk of bias. Discrepancies were resolved after consensus discussion with a third reviewer. Five studies (N = 214) met inclusion criteria, with 1 considered at low risk, 3 at moderate risk, and 1 at high risk of bias. Various controls included corticosteroid injection, dry needling, saline solution injection, and formal exercise therapy. At follow-up (range = 6 to 12 months), no differences occurred between active treatment with PRP injections and any of the control interventions for pain scores, disability measures, or range of motion. In the 2 studies in which PRP injections alone were used in the treatment group, the control group receiving exercise therapy had superior clinical outcomes.
David C. Slawson, MD
Professor and Vice Chair of Family Medicine for Education and Scholarship
Atrium Health
Professor of Family Medicine, UNC Chapel Hill
Charlotte, NC
Clinical question
Is the injection of platelet-rich plasma beneficial in the management of symptomatic knee osteoarthritis in adults?
Bottom line
Compared with either normal saline or hyaluronic acid, multiple intra-articular injections of platelet-rich plasma (PRP) significantly improved knee function but did not reduce perceived pain or improve patient satisfaction in adults with symptomatic knee osteoarthritis. (LOE = 1a-)
Reference
Khoshbin A, Leroux T, Wasserstein D, et al. The efficacy of platelet-rich plasma in the treatment of symptomatic knee osteoarthritis: a systematic review with quantitative synthesis. Arthroscopy 2013;29(12):2037-2048.
Study design: Meta-analysis (other)
Funding source: Industry + govt
Setting: Various (meta-analysis)
Synopsis
These investigators thoroughly searched MEDLINE, EMBASE, the Cochrane Register, and bibliographies of retrieved citations for studies that evaluated the efficacy of intra-articular PRP in the treatment of knee osteoarthritis in adults. They included only high-quality randomized controlled trials and cohort studies of patients 18 years or older, with a minimum of 24 weeks of follow-up. No language restrictions were applied. Outcome assessments occurred using previously validated scoring tools for function, pain, and patient satisfaction. Two individuals independently critiqued potential articles for inclusion and methodologic quality. Discrepancies were resolved by consensus discussion with a third reviewer. Six trials (N = 577 patients) met the inclusion criteria, including 4 randomized trials and 2 comparative control groups. PRP preparation techniques varied among studies, but all 6 included multiple injections. Five of the studies used hyaluronic acid as the control injection and one used normal saline. After 24 weeks, functional scores improved significantly in the PRP groups compared with the control groups, but there were no significant differences between treatment groups and control groups with respect to pain scores and patient satisfaction. The PRP groups had a significantly higher incidence of adverse events compared with the control groups, though all the complications were nonsevere and self-limited. A formal analysis found minimal heterogeneity in the results among the various trials. No formal evaluation was performed to assess for publication bias.
David C. Slawson, MD
Professor and Vice Chair of Family Medicine for Education and Scholarship
Atrium Health
Professor of Family Medicine, UNC Chapel Hill
Charlotte, NC
Clinical question
Does knowledge of point-of-care C-reactive protein help physicians avoid antibiotics without sacrificing benefit in patients with an exacerbation of chronic obstructive pulmonary disease?
Bottom line
C-reactive protein (CRP) guidance regarding the likelihood that antibiotics will be helpful for patients with acute exacerbation of chronic obstructive pulmonary disease (COPD) safely reduced antibiotic use (number needed to treat [NNT] = 5). Physicians were advised that antibiotics are unlikely to be helpful if CRP is less than 20 mg/L, that antibiotics may be helpful if CRP is 20 to 40 mg/L, especially in the presence of purulent sputum, and that antibiotics are likely to be helpful if CRP is greater than 40 mg/L. (LOE = 1b-)
Reference
Butler CC, Gillespie D, White P, et al. C-reactive protein testing to guide antibiotic prescribing for COPD exacerbations. N Engl J Med 2019;381(12):111-120.
Study design: Randomized controlled trial (nonblinded)
Funding source: Government
Allocation: Uncertain
Setting: Outpatient (primary care)
Synopsis
CRP is an inflammatory biomarker elevated in patients with pneumonia and bacterial rhinosinusitis, and is recommended by UK guidelines to help physicians avoid antibiotics in patients with acute lower respiratory tract infection. These authors wondered if the use of CRP would also be effective in patients with an acute exacerbation of COPD. The researchers recruited 653 patients, 40 years and older, with documented COPD who were experiencing an exacerbation. The patients were randomized to usual care or care guided by the results of a point-of-care CRP test. The guidance provided was that antibiotics are unlikely to be helpful if CRP is less than 20 mg/L, that they may be helpful if CRP is 20 to 40 mg/L (especially if the patient also has purulent sputum), and that they are likely to be beneficial if CRP is greater than 40 mg/L. They were also told that the decision should be guided by all patient factors, not just CRP. All patients met at least one of the Anthonisen criteria (increased dyspnea, increased sputum volume, and increased sputum purulence). The mean age of patients was 68 years, 52% were men, and most had GOLD stage 2 or 3 severity of their COPD. Patients were telephoned at 1 and 2 weeks, and were seen in person at 4 weeks; data on antibiotic use were available for 83%. The primary outcome was antibiotic use, which occurred significantly less often with CRP-guided care (57% vs 77%; P < .05; number needed to treat = 5). In addition, at 2 weeks patients in the CRP-guided group had greater improvement in their COPD severity score. Overall, the distribution of CRP was as follows: 76% less than 20 mg/L, 12% 20 to 40 mg/L, and 12% greater than 40 mg/L. There were also no differences among groups in other prescriptions, follow-up visits or hospitalizations in the next 6 months, or the likelihood of pneumonia. The effect of CRP guidance was greater in patients who had more of the Anthonisen criteria, and was only statistically significant for those with at least 2 of the criteria.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Does the use of antibiotics in patients hospitalized with an asthma exacerbation result in better outcomes?
Bottom line
Antibiotics don't benefit patients hospitalized with an asthma exacerbation. Evidence from this study suggests worse outcomes with antibiotic use, including a longer hospital stay, higher costs, and greater risk of diarrhea. Moreover, antibiotic use was not associated with a decreased risk of treatment failure. Although findings were not unexpected, it is surprising that almost half the patients in this study were receiving antibiotics, despite current guidelines. (LOE = 2b)
Reference
Stefan MS, Shieh M, Spitzer KA, et al. Association of antibiotic treatment with outcomes in patients hospitalized for an asthma exacerbation treated with systemic corticosteroids. JAMA Intern Med 2019;Jan 28. doi: 10.1001/jamainternmed.2018.5394. [Epub ahead of print]
Study design: Cohort (retrospective)
Funding source: Unknown/not stated
Setting: Inpatient (any location) with outpatient follow-up
Synopsis
Although current evidence does not support the use of antibiotics for the management of acute asthma exacerbations, patients continue to receive them. Using an administrative database that includes 543 nonteaching hospitals in the United States, these investigators identified patients who were hospitalized during a 2-year period with a principal diagnosis of asthma being treated with systemic corticosteroids. Those with a potential indication for antibiotic treatment, such as a secondary diagnosis of chronic obstructive pulmonary disease, were excluded, leaving a total of 19,811 patients. The median age of the cohort was 46 years and 73% were women. The exposure of interest was the use of antibiotics within the first 2 days of hospitalization for a minimum of 2 days, which was noted in 44% of the overall cohort. The comparison was the subset of patients who either did not receive any antibiotics or did not receive antibiotics within the first 2 days. The primary outcome was hospital length of stay. Patients treated with early antibiotics were older and more likely to be white. They were also more likely to have comorbidities, including congestive heart failure, diabetes, and renal failure. The most common antibiotic class used was macrolides. In a propensity-matched cohort of 13,666 patients, exposure to early antibiotics was associated with a longer length of stay (4 days vs 3 days) and higher hospitalization costs ($4776 vs $3641). The risks of treatment failure and 30-day mortality were similar in the 2 groups. In a sensitivity analysis that compared patients who received early antibiotics with those who received no antibiotics at all, the risk of diarrhea was 2.6 times higher. However, in patients treated only with macrolides compared with those treated with other antibiotics, the incidence of antibiotic-related diarrhea was lower (1.1% vs 2.0%; P < .001).
Nita Shrikant Kulkarni, MD
Assistant Professor in Hospital Medicine
Northwestern University
Chicago, IL
Clinical question
What are the predictors and outcomes associated with excess duration of antibiotic treatment for patients hospitalized with pneumonia?
Bottom line
Patients hospitalized with pneumonia commonly receive treatment with antibiotics for longer than is recommended. Often this occurs because of excessive prescribing upon discharge. Although a longer duration of antibiotics is not associated with improved mortality or readmission rates, it is associated with a greater number of patient-reported adverse events. (LOE = 2b)
Reference
Vaughn VM, Flanders SA, Snyder A, et al. Excess antibiotic treatment duration and adverse events in patients hospitalized with pneumonia. Ann Intern Med 2019 Jul 9. doi: 10.7326/M18-3640. [Epub ahead of print]
Study design: Cohort (retrospective)
Funding source: Government
Setting: Inpatient (ward only)
Synopsis
Using data obtained from 43 hospitals in the Michigan Hospital Medicine Safety Consortium, these investigators identified 6481 patients hospitalized on general medicine wards with either community-acquired pneumonia (CAP) or the now defunct category of health care–associated pneumonia (HCAP). Patients were excluded if they required care in an intensive care unit, were being treated for an additional infection not related to pneumonia, were severely immunocompromised, or were admitted under comfort care. The median age of the cohort was 70 years and 57% were classified as having severe pneumonia. Overall, most patients improved quickly, with 87% clinically stable or discharged by day 5. The primary outcome was the rate of excess antibiotic treatment duration. The expected treatment duration for CAP was 5 days, with longer courses expected if time to clinical stability was longer. The expected treatment duration for patients with HCAP or pneumonia due to Staphylococcus aureus or Pseudomonas aeruginosa was 7 days. An impressive 68% of the cohort received antibiotics for longer than the shortest effective duration. The median duration was 8 days for CAP and 9 days for HCAP. Overall, there were 2526 excess days of antibiotic treatment per 1000 patients hospitalized with pneumonia. Antibiotics prescribed at discharge accounted for 93% of the excess days, commonly prescribed for an additional 5, 7, or 10 days. Although excess treatment was not associated with decreased mortality, readmission rates, or emergency department visits, it was associated with an increased number of patient-reported adverse events. Factors associated with excess treatment included patients with longer hospital stays, those who received diagnostic testing with either a sputum culture or a nonculture test, those who had received high-risk antibiotics in the past 90 days, and those who did not have antibiotic treatment duration documented in the discharge summary. Such documentation only occurred 32% of the time, but was associated with a lower rate of excess antibiotics (2.3 days vs 2.9 days).
Nita Shrikant Kulkarni, MD
Assistant Professor in Hospital Medicine
Northwestern University
Chicago, IL
Clinical question
Are shorter courses of antibiotic therapy as effective for community-acquired pneumonia as longer courses?
Bottom line
This methodology in this study was a bit lacking. I would have preferred that the authors not include any nonrandomized trials, and they should have done a subgroup analysis for higher quality trials using the more standard Cochrane Collaboration Risk of Bias tool. That said, the results consistently favored shorter courses of antibiotics for patients with community-acquired pneumonia (CAP) with regard to clinical cure, mortality, and adverse effects. (LOE = 1a-)
Reference
Tansarli GS, Mylonakis E. Systematic review and meta-analysis of the efficacy of short-course antibiotic treatments for community-acquired pneumonia in adults. Antimicrob Agents Chemother 2018;62(9):e00635-18.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Self-funded or unfunded
Setting: Various (meta-analysis)
Synopsis
The authors searched PubMed, EMBASE, clinicaltrials.gov, and the Cochrane Central Register of Controlled Trials to find trials that randomized patients to shorter (6 or fewer days) or longer (7 or more days) durations of antibiotic therapy for CAP. The reporting of this meta-analysis is not great, with a very short methods section and a nonstandard quality assessment. The authors included 2 nonrandomized trials, but the other 19 included studies were randomized. Approximately half were not double-blinded, and most did not report allocation concealment. Some studies compared different durations of the same antibiotic; others compared, for example, a 5-day course of azithromycin with a longer course of levofloxacin. For the outcome of clinical cure, there was no difference between groups for 5 studies (n = 1600 patients) that used the same antibiotic (relative risk [RR] 1.0; 95% CI 0.97 - 1.02) and 13 studies (n = 1695 patients) that used different antibiotics (RR 1.0; 0.96 - 1.04). The same was true for 2 studies (n = 774 patients) that compared a single dose of azithromycin with a longer course of the same drug. For mortality, 7 studies (n = 2028 patients) found a lower risk of death for courses of 3 to 5 days versus 7 or more days (2.4% vs 4.6%; RR 0.56; 0.35 - 0.90). Serious adverse events were also less likely with short courses of antibiotics (RR 0.73; 0.55 - 0.97). There was no difference in the likelihood of relapse, but the confidence interval was broad due to the relatively small number of studies that reported this outcome (RR 0.67; 0.30 - 1.46).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Does the addition of screening ultrasound add benefit or harm to screening mammography alone?
Bottom line
Adding ultrasound to screening mammography in women younger than 50 years at low, intermediate, or high breast cancer risk is not associated with an increase in breast cancer detection. It is associated, however, with increased unnecessary biopsy recommendations and results in more frequent follow-up. (LOE = 2b)
Reference
Lee JM, Arao RF, Sprague BL, et al. Performance of screening ultrasonography as an adjunct to screening mammography in women across the spectrum of breast cancer risk. JAMA Intern Med 2019;179(5):658-667.
Study design: Cohort (retrospective)
Funding source: Government
Setting: Outpatient (any)
Synopsis
These researchers compared the results from 6081 women who were screened for breast cancer with mammography and ultrasound with 30,062 screening mammograms from 15,176 women drawn from 13 years of data from 2 breast cancer surveillance registries in the United States. Most (74.3%) of the ultrasound screens were performed in women with dense breasts, and, as compared with the mammography-alone group, were more likely to be at higher risk of breast cancer or to be younger than 50 years. Despite these differences, the cancer detection rate was similar across groups (5.4 vs 5.5 per 1000 screens), as was the development of cancer between screenings (interval cancer rate). However, the rate of unnecessary biopsies was more than twice as high for the combination screening (52.0 vs 22.2 per 1000 screens), as were calls for rescreening at shorter-than-normally-recommended intervals (relative risk = 3.10; 95% CI 2.6 - 3.7).
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Do marine n-3 fatty acids (also called omega-3 fatty acids) reduce the risk of cardiovascular events or cancer in patients without known vascular disease or cancer?
Bottom line
This is the second large well-designed study (the first one included patients with diabetes, this one largely included patients without diabetes) that found no benefit to supplementation with marine n-3 fatty acids in a dosage of 1 gram daily for the primary prevention of cardiovascular disease or cancer. This study featured more than 20 comparisons, so the small reduction found in myocardial infarctions may be due to chance alone. (LOE = 1b)
Reference
Manson JE, Cook NR, Lee IM, et al, for the VITAL Research Group. Marine n-3 fatty acids and prevention of cardiovascular disease and cancer. N Engl J Med 2019;380(1):23-32.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Population-based
Synopsis
A recent POEM reported that a UK trial of n-3 fatty acids in 15,000 patients with diabetes found no evidence of benefit as primary prevention. The current study is the first adequately powered US trial of n-3 (omega-3) fatty acids for primary prevention. The researchers recruited a total of 25,781 men 50 years and older and women 55 years and older who had no history of cardiovascular disease or cancer. The groups were balanced at the start of the study, with a mean age of 67 years, 51% women, and 14% with diabetes. The investigators used a 3-month placebo run-in period to exclude patients who were noncompliant, which was approximately one-third of those initially recruited. This could overestimate the potential benefit seen in clinical practice. In this arm of the trial, participants were randomly assigned to 1 g n-3 fatty acids or placebo once daily. The median follow-up was 5.3 years. There was no reduction in the likelihood of cancer diagnosis or of the primary composite cardiovascular outcome (cardiovascular death, nonfatal myocardial infarction [MI], or ischemic stroke). There was no difference between groups regarding cardiovascular or all-cause mortality. There was a reduction in total MIs (hazard ratio 0.72; 95% CI 0.59 - 0.90), although the absolute reduction was small (145 vs 200 total MIs with >12,000 patients in each group). It is important not to make too much of this finding, as there were 24 comparisons made in the primary analysis, so this one could be caused by chance alone. There was no difference between groups with regard to adverse events. A subgroup analysis found a modest but statistically significant reduction in the primary composite outcome for persons who ate fewer than 1.5 servings of fish per week, but not in those who ate more.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
What percentage of visits to an emergency department for high blood pressure are prompted by home or pharmacy blood pressure readings?
Bottom line
Home or pharmacy blood pressure monitoring prompts half of all emergency department (ED) visits for high blood pressure. Many of these patients were "symptomatic" with nonspecific complaints. With recommendations for more at-home or other non-office monitoring, patients need to be given specific guidance on what to do when they have an elevated reading; for example, "Call your doctor rather than go to the emergency department." (LOE = 2b)
Reference
Atzema CL, Wong A, Masood S, et al. The characteristics and outcomes of patients who make an emergency department visit for hypertension after use of a home or pharmacy blood pressure device. Ann Emerg Med 2018;72(5):534-543.
Study design: Cohort (retrospective)
Funding source: Government
Setting: Emergency department
Synopsis
This retrospective study included all adults who made an ED visit to 1 of 5 hospitals in Ontario, Canada, and who received a primary diagnosis of hypertension. The median blood pressure upon presentation was 182/97 mm Hg among the 1508 patients. Home blood pressure readings prompted the ED visit in 40.9% of these patients. Another 8.3% were prompted by a pharmacy blood pressure reading, and the rest of the visits were prompted either by a physician referral or were not documented. Many patients reported general symptoms, including headache (38%), dizziness (30%), and chest pain (16%). The average blood pressure at home was 190/100 mm Hg though it dropped to 171/88 mm Hg by the second ED measurement; in pharmacy-identified high blood pressure, the average reported blood pressure was 185/100 mm Hg , dropping eventually to 169/94 mm Hg. Despite lack of evidence of benefit and recommendations to the contrary, 41% of patients received antihypertensive treatment in the ED. Only 3.1% of patients with home- or pharmacy-prompted visits were admitted as compared with 11.9% of physician referrals and 11.0% of patients with undocumented reasons for visiting. Over the next year, 11% of patients initially discharged from the ED made another ED visit for high blood pressure.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
In patients with atrial fibrillation and coronary artery disease, is rivaroxaban plus an antiplatelet agent superior to rivaroxaban alone?
Bottom line
For patients with stable coronary artery disease and atrial fibrillation, monotherapy with rivaroxaban is preferable to rivaroxaban plus a platelet inhibitor one year after percutaneous coronary intervention (PCI) or bypass surgery. Not only were there fewer deaths, there were fewer hemorrhagic strokes and major bleeding events. The findings are consistent with the results of a previous large Danish cohort study (Circulation 2019;139(6):775-786). (LOE = 1b-)
Reference
Yasuda S, Kaikita K, Akao M, et al, for the AFIRE Investigators. Antithrombotic therapy for atrial fibrillation with stable coronary disease. N Engl J Med 2019;381(12):1103-1113.
Study design: Randomized controlled trial (nonblinded)
Funding source: Industry
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
Many patients who receive dual antiplatelet therapy also have an indication for anticoagulation, such as atrial fibrillation. Current guidelines recommend a brief period of triple therapy with dual antiplatelet therapy plus an anticoagulant for the periprocedural period, followed by 1 to 12 months of an anticoagulant plus a P2Y12 inhibitor, followed by oral anticoagulant alone. This study compared monotherapy with the direct oral anticoagulant rivaroxaban with rivaroxaban plus an antiplatelet agent in 2215 adults with atrial fibrillation who had undergone PCI or bypass surgery at least 12 months previously, or had stenosis of at least 50% in one coronary vessel. All patients also had a score of at least 1 on the older CHADS2 scale, which puts them at moderate to high risk for stroke. The dosage of rivaroxaban was 15 mg once daily (10 mg once daily if creatinine clearance was 15 mL to 49 mL per minute) and 70% of the patients received aspirin as their antiplatelet agent. Groups were balanced at the start of the study with a mean age of 74 years, 79% were men, and 71% having undergone a PCI, most receiving a drug-eluting stent. Atrial fibrillation was paroxysmal in 53% of patients; persistent or permanent in 47%. This was a noninferiority trial with a primary endpoint that was a broad composite of stroke, systemic embolism, myocardial infarction, unstable angina requiring revascularization, or death; analysis was by modified intention to treat of all patients who underwent randomization. The trial was terminated after a median follow-up of 23 months because of an excess of deaths in the dual therapy group. The primary endpoint was significantly less likely in the rivaroxaban monotherapy group (4.14% vs 5.75% per year; P < .001; number needed to treat [NNT] = 61 per year). Regarding individual outcomes, death (1.85 vs 3.37%; P < .05; NNT = 66 per year), hemorrhagic stroke (0.18% vs 0.60%; P < .05; NNT = 312), and major bleeding (1.62% vs 2.76%; P = .01; NNT = 88 per year) were all less likely with rivaroxaban monotherapy. The benefit was consistent regardless of initial CHA2DS2-VASc score for stroke risk or HAS-BLED score for bleeding risk. Although the subgroup analyses should be considered exploratory, the benefit was greater for men than for women, for patients undergoing PCI rather than coronary artery bypass grafting, and for patients 75 years or older. This study was funded by the Japanese Research Foundation, but under a contract with Bayer Yakuhin.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Is it necessary to monitor for hematologic or hepatic transaminase abnormalities in patients who take oral terbinafine or griseofulvin?
Bottom line
Although baseline testing is still recommended, this study supports the abandonment of subsequent testing to monitor otherwise healthy patients during therapy with griseofulvin or terbinafine given the rarity of abnormalities. (LOE = 2b)
Reference
Stolmeier DA, Stratman HB, McIntee TJ, Stratman EJ. Utility of laboratory test result monitoring in patients taking oral terbinafine or griseofulvin for dermatophyte infections. JAMA Dermatol 2018; 154(12):1409-1416.
Study design: Cohort (retrospective)
Funding source: Foundation
Setting: Outpatient (any)
Synopsis
The package insert for terbinafine recommends a baseline measurement of alanine aminotransferase (ALT) and aspartate aminotransferase (AST), and a white blood cell count in patients with immunodeficiency who will be taking the medication for more than 6 weeks. However, physicians commonly order a baseline complete blood count (CBC) on all patients, and many obtain follow-up ALT and AST after some duration of treatment, often 1 month. This was a retrospective review of 4985 adults and children with a mean age of 43 years who had been given 4309 courses of terbinafine and 793 courses of griseofulvin. Patients were included if they were given the medication to treat an episode of dermatophyte infection, and excluded if they had a diagnosis of anemia, myelodysplastic syndrome, leukemia, or alcohol abuse. Laboratory results were divided into baseline results obtained between 90 days before and 7 days after starting the medication, and monitoring results obtained 7 days or more after the initiation of therapy. Of patients given terbinafine, 61% had at least one baseline test, 42% at least one monitoring test, and 32% had both tests. Testing was less frequent for patients taking griseofulvin (21%, 20%, and 9%, respectively). Approximately 1 in 3 patients given terbinafine and approximately 1 in 7 patients given griseofulvin had a baseline CBC. For both drugs, the rates of abnormal AST or ALT either at baseline or during monitoring was approximately 3%, of which less than 0.2% were grade 2 or worse elevations (> 3 times the upper limit of normal). The pattern was similar for griseofulvin. Only one patient developed hepatotoxicity and had the medication discontinued as a result.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
What treatments of urinary incontinence in women can cure or improve symptoms?
Bottom line
Broadly speaking, behavioral therapies—including bladder training, biofeedback, mindfulness-based stress reduction, yoga, and weight loss—are generally more effective than pharmacologic treatment for urinary incontinence in women and should be our first options. Topical and oral estrogens don't work. Fancier interventions, such as neuromodulation and botulinum injections, work better than no treatment. (LOE = 1a-)
Reference
Balk EM, Rofeberg VN, Adam GP, Kimmel HJ, Trikalinos TA, Jeppson PC. Pharmacologic and nonpharmacologic treatments for urinary incontinence in women. A systematic review and network meta-analysis of clinical outcomes. Ann Intern Med 2019;170(7):465-479.
Study design: Meta-analysis (other)
Funding source: Government
Setting: Various (meta-analysis)
Synopsis
These researchers searched 6 databases, including the Cochrane Central Register of Controlled Trials, to identify 84 randomized and nonrandomized controlled trials that evaluated the effectiveness of pharmacologic and nonpharmacologic treatments of urge, stress, or mixed urinary incontinence in women. Two researchers independently selected articles for inclusion. They used network meta-analysis, which is a means of evaluating treatments that were not directly compared. Most studies had low or moderate risk of bias. Although it's a useful method to get an overall picture of relative benefit, a network meta-analysis is complex, paints results with a broad brush, and requires a number of assumptions that can inflate the problems always present in statistical analysis. All interventions, except hormone treatment and injections designed to provide periurethral bulking, are effective at decreasing at least some symptoms. Behavioral therapy is more effective than pharmacologic treatments for producing cure or improvement in women with stress or urge incontinence. Neuromodulation (surgical implantation of a device to produce nerve stimulation) is more effective than no treatment to produce cure, improvement, and patient satisfaction in women with stress or urge incontinence. Botulinum toxin injection is also more effective than no treatment for urge incontinence.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Are patients who are prescribed tramadol instead of another opioid less likely to continue long-term use?
Bottom line
Given its touted lower risk of abuse, tramadol (Ultram, Ivodol) is perceived as being less likely to result in long-term use. Not so. Patients who receive tramadol following surgery are as likely as patients who receive other short-acting opioids to have prolonged use. Additionally, even though the overall subset is small, patients who take tramadol are more likely to receive chronic opioid treatment 6 months after surgery. (LOE = 2b)
Reference
Thiels CA, Habermann EB, Hooten WM, Jeffery MM. Chronic use of tramadol after acute pain episode: cohort study. BMJ 2019;365:l1849.
Study design: Cohort (retrospective)
Funding source: Foundation
Setting: Population-based
Synopsis
This study drew on administrative prescribing data of 444,764 patients who were given a short-acting opioid following surgery, Of these patients, 357,884 filled a discharge prescription for an opioid, including hydrocodone (53%), oxycodone (37.5%), and tramadol (4%). The prescription data indicated that 7% of these patients had prolonged use (defined as receiving at least one more prescription in the 90 to 180 days after surgery), and 0.5% had long-term use (defined as receiving an opioid for at least 3 months after the first 180 days following surgery). Patients who were prescribed tramadol were at least as likely as patients given other opioids to have prolonged use and were more likely (0.7% vs 0.5%) to require long-term opioid use (risk ratio 1.41; 95% CI 1.08 - 1.75; P = .013).
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Is the probiotic Lacobacillus rhamnosus GG safe and effective for the treatment of acute gastroenteritis in young children?
Bottom line
Treatment with the probiotic Lactobacillus rhamnosus GG does not result in faster symptomatic improvement or less moderate/severe diarrhea in young children with acute gastroenteritis. (LOE = 1b)
Reference
Schnadower D, Tarr PI, Casper TC, et al. Lactobacillus rhamnosus GG versus placebo for acute gastroenteritis in children. N Engl J Med 2018;379(21):2002-2014.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Emergency department
Synopsis
Although some previous studies have shown a benefit to probiotics for the treatment of acute gastroenteritis in children, they have generally been small, funded by industry, or had methodologic flaws. This study was a large, well-designed randomized trial that attempted to more definitively answer this question. Children aged 3 months to 4 years with at least 3 watery stools per day for less than 7 days were identified in 10 university pediatric emergency departments. Children were excluded if they or their caregivers were potentially immunocompromised, or if the children were taking chronic steroids, had been premature, or had chronic gastrointestinal disease. Children were also excluded if they had any biliary tract disease, hematochezia, or allergies to the study medication or to any antibiotics used to treat invasive L. rhamnosus infection. They were then randomized to receive the probiotic or placebo once daily for 5 days, stratified by presentation within or later than 48 hours of the onset of illness, and with the first dose given in the emergency department. Of the 971 patients randomized, 943 completed the study, with a similar rate of loss to follow-up in both groups. The median age of participants was 1.4 years, 53% were male, and the median duration of diarrhea was 53 hours; groups were balanced at the start of the study. After 14 days, the likelihood of moderate to severe symptoms based on the modified Vesikari scale score was similar between groups (11.8% in the probiotic group and 12.6% in the placebo group). There were also no differences between groups regarding secondary symptom outcomes such as the duration or frequency of diarrhea. There was more wheezing in the probiotic group (5 vs 0; P = .03).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Is augmentation treatment effective for patients with treatment-resistant depression?
Bottom line
The available evidence that treatment-resistant depression—depression unresponsive to 2 different treatments of adequate dose and length—responds well to augmentation treatment (ie, adding psychotherapy, lithium, or aripiprazole [Abilify] to current treatment) is weak. The available evidence shows no benefit with lithium and small benefit with psychotherapy or aripiprazole. (LOE = 1a-)
Reference
Strawbridge R, Carter B, Marwood L, et al. Augmentation therapies for treatment-resistant depression: systematic review and meta-analysis. Br J Psychiatry 2019;214(1):42-51.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Government
Setting: Various (meta-analysis)
Synopsis
The authors searched 2 databases (but not the Cochrane Library) for randomized studies of augmentation treatment for patients who did not respond to at least 2 courses of treatment for major depressive disorder.Two authors selected studies for inclusion and independently extracted the data. Most of the 28 studies of 5461 patients had low to moderate risk of bias (ie, were of medium to high quality) and included both drug treatment and psychological therapies. Instead of comparing directly across treatments (that is, the benefit in one group vs the other), the authors compared the before-after change in results within each group. In 3 low-quality studies, psychological treatment showed a moderate benefit. In 4 studies of aripiprazole, there was a small likelihood of benefit after short-term treatment (effect size = 1.33; 95% CI 1.23 - 1.44) as compared with placebo. Lithium produced an effect size similar to placebo.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
In patients with acute pain, does a higher dose of ibuprofen produce greater pain relief?
Bottom line
Higher doses of ibuprofen for acute pain relief offer no more benefit at 60 minutes than a single 400-mg dose. The same has been shown for chronic treatment of osteoarthritis--an anti-inflammatory dose is not needed. Furthermore, another study showed equivalence between 200-mg and 400-mg dose. The same has been shown for chronic treatment of osteoarthritis--an anti-inflammatory dose is not needed. Furthermore, another study showed equivalence between 200-mg and 400-mg doses of ibuprofen. (LOE = 2b)
Reference
Motov S, Masoudi A, Drapkin J, et al. Comparison of oral ibuprofen at three single-dose regimens for treating acute pain in the emergency department: a randomized controlled trial. Ann Emerg Med2019;74(4):530-537.
Study design: Randomized controlled trial (double-blinded)
Funding source: Foundation
Allocation: Concealed
Setting: Emergency department
Synopsis
These authors enrolled 225 adults who presented to a single emergency department with an acute painful condition (~75% with musculoskeletal pain). The average pain score was between 6 and 7 on a scale of 1 to 10, with higher scores indicating higher pain. Using concealed allocation, patients were randomly assigned to receive a single dose of ibuprofen, either 400 mg, 600 mg, or 800 mg. Using intention-to-treat analysis, pain scores after 60 minutes dropped to 4.36 to 4.50 in all 3 groups. The study had 80% power to find a difference of at least 1.3 points, if it existed, among the groups.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
What signs and symptoms are most useful for excluding the diagnosis of pneumonia in community-dwelling adults with an acute respiratory infection?
Bottom line
Community-dwelling adults who present to a primary care office with acute respiratory infection symptoms but normal vital signs and normal findings on a pulmonary examination have only a 0.4% likelihood of community-acquired pneumonia (CAP). (LOE = 1a)
Reference
Marchello CS, Ebell MH, Dale AP, Harvill ET, Shen Y, Whalen CC. Signs and symptoms that rule out community-acquired pneumonia in outpatient adults: A systematic review and meta-analysis. J Am Board Fam Med 2019;32(2):234-247.
Study design: Systematic review
Funding source: Self-funded or unfunded
Setting: Various (meta-analysis)
Synopsis
Identifying signs and symptoms that reliably rule out CAP may help reduce the overuse of radiography and/or laboratory testing. These investigators systematically searched MEDLINE and reference lists of pertinent articles for studies that used a clinical decision rule to diagnose CAP in the outpatient setting. Eligible criteria included the use of either a chest x-ray or computed tomography as the reference standard for either all enrolled patients or a random/systematic sample of the enrolled patients. In addition, only studies that recruited adults or adolescents in an outpatient setting, including the emergency department, were included. Two individuals independently reviewed potential studies for inclusion criteria and methodologic quality using standard criteria. The resolution of any disagreements occurred after consensus discussion with a third reviewer. A total of 12 studies met inclusion criteria, of which 6 were performed in an emergency department setting and 6 in a primary care setting. Sample sizes ranged from 246 to 2820 patients. Six studies were found to be at low risk of bias; the remaining 6 were at moderate risk of bias. The combination of normal vital signs (temperature, respiratory rate, and heart rate) plus normal findings on the pulmonary examination reliably excluded CAP (sensitivity = 0.96; 95% CI 0.92 - 0.28; negative likelihood ratio = 0.10; 0.07 - 0.13).
David C. Slawson, MD
Professor and Vice Chair of Family Medicine for Education and Scholarship
Atrium Health
Professor of Family Medicine, UNC Chapel Hill
Charlotte, NC
Clinical question
Does low-dose aspirin prevent cardiovascular events and cardiovascular-related death in otherwise healthy older persons?
Bottom line
Low-dose aspirin does not reduce the likelihood that otherwise healthy older patients will experience a major cardiovascular event during nearly 5 years of follow-up. (LOE = 1b)
Reference
McNeil JJ, Wolfe R, Woods RL, et al, for the ASPREE Investigator Group. Effect of aspirin on cardiovascular events and bleeding in the healthy elderly. N Engl J Med 2018;379(16):1509-1518.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Population-based
Synopsis
The Aspirin in Reducing Events in the Elderly (ASPREE) trial randomized 19,114 community-dwelling adults to receive either 100 mg of enteric-coated aspirin or placebo. The study was conducted in the United States and Australia, with patients recruited between 2010 and 2014. Participants were 70 years or older (65 years or older if black or Hispanic in the United States, because of their shorter average lifespan), had no serious comorbidity that would be expected to limit their life expectancy to less than 5 years, and no known cardiovascular (CV) or cerebrovascular disease, dementia, high bleeding risk, or contraindication to aspirin. The study included a 1-month placebo run-in period to ensure at least 80% adherence to the study medication. During the run-in period, 4049 patients were excluded, 61% because they failed adherence. Included patients were contacted every 3 months to further encourage adherence and to gather interim data. Outcomes were adjudicated by a committee masked to treatment assignment. The median age of participants was 74 years, 56% were women, and 8.7% were non-white. Most of the patients were recruited in Australia (87%), 74% had hypertension, 65% had hyperlipidemia, and only 11% had diabetes. Participants were followed up for a median of 4.8 years, and only 2.2% withdrew or were lost to follow-up. A separate report found no significant reduction in the composite of death, dementia, and disability. The current report looks at the composite outcome of fatal coronary heart disease, nonfatal myocardial infarction, stroke, and hospitalization for heart failure. This is a broad composite, so it is important to look at individual components of the outcome. In this case, there was no difference between groups regarding the composite or any of the individual components. Major hemorrhage was more common in the aspirin group (8.6 vs 6.2 events per 1000 person-years; hazard ratio 1.38; 95% CI 1.18 - 1.62; number needed to treat to harm = 417 per year).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Does aspirin improve disability-free survival in an otherwise healthy older person?
Bottom line
This landmark study found that in a contemporary population where risk factors such as hyperlipidemia and hypertension are more likely to be addressed, aspirin did not provide a benefit in terms of mortality, dementia, or disability in a largely white group of older patients. (LOE = 1b)
Reference
McNeil JJ, Woods RL, Nelson MR, et al, for the ASPREE Investigator Group. Effect of aspirin on disability-free survival in the healthy elderly. N Engl J Med 2018;379(16):1499-1508.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Population-based
Synopsis
These are the initial results of the landmark Aspirin in Reducing Events in the Elderly (ASPREE) trial. Two other reports describe the effect of aspirin on all-cause mortality and on cardiovascular disease. The authors randomized 19,114 community-dwelling adults to receive either 100 mg of enteric-coated aspirin or placebo. The study was conducted in the United States and Australia, with patients recruited between 2010 and 2014. Participants were 70 years or older (65 years or older if black or Hispanic in the United States, because of their shorter average lifespan), had no serious comorbidity that would be expected to limit their life expectancy to less than 5 years, and no known cardiovascular (CV) or cerebrovascular disease, dementia, high bleeding risk, or contraindication to aspirin. The study included a 1-month placebo run-in period to ensure at least 80% adherence to the study medication. During the run-in period, 4049 patients were excluded, 61% because they failed adherence. Included patients were contacted every 3 months to further encourage adherence and to gather interim data. Outcomes were adjudicated by a committee masked to treatment assignment. The median age of participants was 74 years, 56% were women, and 8.7% were non-white. Most of the patients were recruited in Australia (87%), 74% had hypertension, 65% had hyperlipidemia, and only 11% had diabetes. Participants were followed up for a median of 4.8 years, and only 2.2% withdrew or were lost to follow-up. The primary outcome was a composite of death, dementia, or physical disability, which occurred in 921 persons who received aspirin and 914 who received placebo (hazard ratio [HR] 1.10; 95% CI 0.92 - 1.11). All-cause mortality was slightly higher in the aspirin group (12.7 vs 11.1 events per 1000 person-years; HR 1.14, 1.01 - 1.29; number needed to treat to harm [NNTH] = 625 per year). Major hemorrhage was more common in the aspirin group (8.6 vs 6.2 events per 1000 person-years; hazard ratio 1.38; 95% CI 1.18 - 1.62; number needed to treat to harm = 417 per year).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
How long do colds last in children?
Bottom line
Most respiratory illnesses in children are mild and don't result in seeking medical care or missing school, but they can last as long as 3 weeks. (LOE = 1b-)
Reference
Hay AD, Anderson E, Ingle S, Beck C, Hollingworth W. Respiratory tract infections in children in the community: prospective online inception cohort study. Ann Fam Med 2019;17(1):14-22.
Study design: Cohort (prospective)
Funding source: Government
Setting: Population-based
Synopsis
These researchers from the United Kingdom enrolled a cohort of 485 generally healthy children aged between 3 months and 15 years from February to July 2016. They recruited patients from the enrollment lists of practices located within 10 miles of Bristol. They excluded immunocompromised children and those with terminal illnesses. If the child had a respiratory illness, the researchers asked the parents to start the cohort once the symptoms resolved. Each week, the researchers sent an e-mail or text message asking whether the child had specific respiratory tract symptoms, such as rhinorrhea, earache, sore throat, or cough. If the child had any of these symptoms, the researchers asked the parent to provide daily online updates. They also asked the parents about missed school, medication use, consultations, and so forth. Based on the patient lists from the participating practices, participating children tended to be 2 years younger and less socioeconomically deprived than the nonparticipating children. During 5 months of follow up, parents reported 346 new respiratory infections in 259 children—53% of children had at least one new respiratory infection; the attack rate was 71% and afflicted kids averaged 1.3 infections during this short interval. The researchers report on 197 children's first illness. The median duration of illness was 9 days, but this is a skewed distribution with a long tail: It took 23 days for 90% of the children to recover. Approximately half the parents reported symptoms associated with lower respiratory infections and the presence of these symptoms was associated with longer duration. Sniffles tended to last, while earaches were of the shortest duration. Dry cough was most severe for the first 10 days, but also persisted for about 3 weeks. The parents had primary care consultations (office or telephone encounters) only 8% of the time, and only 9% of infections caused a missed day of school. As one might expect, the severity of symptoms was worse among those kids. Since a healthy chunk of the study period occurred in spring and summer, these numbers might be worse during fall and winter months.
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
Clinical question
Can strep throat in children and adults be treated with 5 days of oral penicillin?
Bottom line
Five days of 800 mg penicillin 4 times a day produced results not worse than (i.e., noninferior to) 10 days of 1000 mg penicillin 3 times a day, with shorter symptom duration. This is not the first study to show similar benefits with a shorter duration of oral Augmentin/cephalosporin or amoxicillin. (LOE = 2b)
Reference
Skoog Stahlgren GS, Tyrstrup M, Edlund C, et al. Penicillin V four times daily for five days versus three times daily for 10 days in patients with pharyngotonsillitis caused by group A streptococci: randomised controlled, open label, non-inferiority study. BMJ 2019;367:l5337.
Study design: Randomized controlled trial (single-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (primary care)
Synopsis
The investigators enrolled 317 adults and 105 children (6 years and older) from 17 primary care centers in Sweden. Eligible patients had to have at least 3 Centor criteria and a positive rapid antigen test result for group A streptococcus. Using concealed allocation, patients were randomly assigned to receive either 800 mg penicillin 4 times a day for 5 days or 1000 mg penicillin 3 times a day for 10 days. This was an open-label trial, meaning that both patients and their clinicians were aware of the treatment the patient received. However, the researchers who analyzed the data were masked to treatment until the results were assembled. Clinical cure, defined as complete recovery without major residual symptoms or clinical findings, was assessed 5 days to 7 days after the completion of treatment; that is, on day 10 to day 12 in the 5-day treatment group and on day 15 to day 17 in the 10-day treatment group. This timing of assessment could possibly favor better results with longer treatment. In the per-protocol analysis (patients who completed treatment), cure rates at 5-7 days post treatment were 89.6% in the 5-day group and 93.3% in the 10-day group. Results were similar when using intention-to-treat analysis. Patients receiving the higher daily dose/shorter duration had quicker symptom resolution. Bacterial cure rates were higher with 10 days of treatment but there was no difference in complication rates or new episodes of tonsillitis at the 3-month follow-up. Adverse events such as diarrhea, nausea, or vaginitis were more likely and lasted longer in the 10-day group. The study had a power of 85% to detect a greater than 10% difference between treatments if one existed.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Are statins effective in patients older than 75 years?
Bottom line
Statins are effective in preventing major coronary events in patients older than 75 years, but this effect is exclusive to those with established vascular disease (ie, effective for secondary, not primary, prevention). This is consistent with the results from the ALLHAT trial, which also showed no benefit to primary prevention, and additionally showed a trend to harm in those older than 75 years. (LOE = 1a)
Reference
Cholesterol Treatment Trialists' Collaboration, Armitage J, Baigent C, et al. Efficacy and safety of statin therapy in older people: a meta-analysis of individual participant data from 28 randomised controlled trials. Lancet 2019;393(10170):407-415.
Study design: Meta-analysis (other)
Funding source: Government
Setting: Various (meta-analysis)
Synopsis
The members of the Cholesterol Treatment Trialists' Collaboration have pooled the patient level data from randomized trials of statins that evaluated low-density lipoprotein (LDL) concentrations and included at least 1000 patients who were followed up for at least 2 years. These data from 28 studies and more than 186,000 total patients included more than14,000 patients 75 years or older (~ 8% of the total pool). The researchers followed up these patients for a median of 5 years. They used an intention-to-treat analysis to evaluate the main outcomes of major coronary events (nonfatal myocardial infarction or coronary death), coronary revascularization, stroke, cancers, and all-cause mortality. Among patients older than 75 years, 2.6% of statin-treated patients experienced a major coronary event each year compared with 3% of control patients (number needed to treat = 250 per year). The authors report there was a barely significant 13% relative reduction in cardiovascular events for every mmol/L reduction in LDL concentration (relative risk 0.82; 95% CI 0.70 - 0.96). However, the rate of revascularization and stroke was not significantly decreased with statins. The authors also found no effect of statins on incident cancers or cancer mortality. Finally, although the data were limited, patients without vascular disease at the time of enrollment experienced no significant reduction in events.
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
Clinical question
Is fully automated blood pressure measurement more accurate than manual sphygmomanometry?
Bottom line
There are 2 takeaways and a recommendation from this analysis of in-office automated blood pressure measurement. The takeaways: (1) Automated measurement aligns better with ambulatory blood pressure monitoring, the best predictor of cardiovascular events, than manual measurement; and (2) manual readings are an average 13.4 to 14.5 mm Hg (systolic) higher than daytime ambulatory or automated readings in patients with hypertension. The recommendation: Since the recent guidelines from the American College of Cardiology/American Heart Association are based on automated readings, follow them only if you switch from the squeeze bulb to the machine.
Reference
Roerecke M, Kaczorowski J, Myers MG. Comparing automated office blood pressure readings with other methods of blood pressure measurement for identifying patients with possible hypertension. A systematic review and meta-analysis. JAMA Intern Med 2019;179(3):351-362.
Study design: Meta-analysis (other)
Funding source: Self-funded or unfunded
Setting: Various (meta-analysis)
Synopsis: The authors searched 3 databases, including the Cochrane Central Register of Controlled Trials, to identify studies that compared automated office blood pressure readings with standard or research-based manual measurement or ambulatory automated recording during awake hours (the latter used as the reference standard). The authors also searched reference lists of identified articles. They included papers in any language, 2 authors independently selected articles for inclusion, and a single investigator extracted data. Automated measurement had be performed without anyone activating the machine and used 3 to 5 readings separated by 1-minute to 2-minute intervals. In 31 studies of 9279 participants, the pooled mean differences between routine measurement and awake ambulatory measurements were 13.4 mm Hg systolic and 5.9 mm Hg diastolic. There was no difference between ambulatory and automated blood pressure. The difference between manual and automated blood pressures was 14.5 mm Hg systolic in patients with hypertension. There was a great deal of heterogeneity among studies for all outcomes that could not be explained by any of the variables available to the researchers. There was no evidence of publication bias.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
What is the yield of a screening program based on fecal immunochemical testing every 2 years for 10 years?
Bottom line
Over a 10-year period, the rates of detection of colorectal cancer (CRC) and advanced adenomas using fecal immunochemical testing (FIT) are similar to those seen in studies of screening colonoscopy. This is reassuring, but it does not prove that FIT reduces morbidity and mortality due to CRC as effectively as colonoscopy. Modeling concludes that a FIT-based screening program will result in half as many colonoscopies as a program based on colonoscopy, a significant reduction in cost, burden, and harm of screening. (LOE = 2b)
Reference
Zorzi M, Hassan C, Capodaglio G, et al. Long-term performance of colorectal cancer screening programmes based on the faecal immunochemical test. Gut 2018;67(12):2124-2130.
Study design: Cohort (prospective)
Funding source: Government
Setting: Population-based
Synopsis
The 2 most widely recommended strategies for CRC screening are FIT and colonoscopy. Several trials are currently underway to compare these approaches, with cancer-specific mortality as the primary outcome. Until then, we have to rely on observational studies and modeling to understand the benefit of each approach. Although colonoscopy is more sensitive than FIT, especially for the detection of advanced adenomas, what matters is performance over a long-term screening program, not one-time accuracy. This study reports the results of 5 rounds of biennial FIT in a screening population aged 50 to 69 years in the Veneto region of northern Italy. Not surprisingly, the rate of detection of CRC was highest in the first round of screening when prevalent lesions were detected (3.3/1000 persons), declining in subsequent rounds and stabilizing after the third round (~1/1000 persons). Between rounds 3 and 6, the CRC detection rate declined slightly from 0.95 to 0.84 per 1000. A similar pattern was seen for advanced adenomas, declining from 15.9 per 1000 persons to approximately 10 per 1000 persons in subsequent rounds. Over the 10-year study period, the cumulative rate of positive FIT results was 25% for men and 17.6% for women. The cumulative rate for advanced adenoma was 60 per 1000 persons, and for CRC was 8.5 per 1000 persons. These rates are similar to those seen in studies of colonoscopy in both Italy and the United States.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Does a lack of early symptom improvement in patients treated for depression predict treatment failure?
Bottom line
Don't be in a hurry to change treatment in patients with severe depression who do not respond to treatment within the first 2 weeks. Early response to treatment predicts eventual response or remission, but a lack of early response does not predict treatment failure. Approximately one third of patients who do not show an early response will respond by 6 weeks. No individual symptom response predicts eventual improvement. (LOE = 1a)
Reference
De Vries YA, Roest AM, Bos EH, Burgerhof JGM, van Loo HM, de Jonge P. Predicting antidepressant response by monitoring early improvement of individual symptoms of depression: individual patient data meta-analysis. Br J Psychiatry 2019; 214(1):4-10.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Unknown/not stated
Setting: Various (meta-analysis)
Synopsis: These researchers used individual patient data derived from 30 studies of the treatment of severe major depressive disorder with a second-generation antidepressant. Overall, they had data on 2184 patients who received a placebo and 6058 who received an antidepressant. By 6 weeks of treatment, approximately 50% of treated patients had responded, with 32% achieving remission of symptoms. By 12 weeks, the rate was up to approximately 68% response with 49% achieving remission. Patients with early improvement—by 2 weeks—were likely to respond by 6 weeks, but almost 33% of patients without early improvement responded by 6 weeks and 43% of them responded by 12 weeks. No individual symptom response predicted eventual response or remission.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Are the modern fecal immunochemical tests for occult blood in the stool less accurate in patients who are taking aspirin, an anticoagulant, or a nonsteroidal anti-inflammatory drug?
Bottom line
The use of aspirin, nonsteroidal anti-inflammatory drugs (NSAIDs), and oral anticoagulants have no clinically important effects on the positive predictive value of the fecal immunochemical test (FIT) in a screening population. (LOE = 1a)
Reference
Nieuwenburg SA, Vuik FE, Kruip MJ, Kuipers EJ, Spaander MC. Effect of anticoagulants and NSAIDs on accuracy of faecal immunochemical tests (FITs) in colorectal cancer screening: a systematic review and meta-analysis. Gut 2019;68(5):866-872
Study design: Systematic review
Funding source: Self-funded or unfunded
Setting: Various (meta-analysis)
Synopsis
Previous research has shown that approximately 6.2% of screening FIT results are positive, with a positive predictive value (PPV) of 35% to 55%. These authors wondered whether the use of something that can make you bleed would reduce the PPV by making benign lesions more likely to bleed (leading to more false-positive results), or whether it would increase the PPV by making precancerous and malignant lesions more likely to bleed? They did a careful search of the literature and identified 8 studies with 2022 patients that compared the accuracy of the FIT in a screening population of patients taking an oral anticoagulant or an NSAID with those not taking either drug. Of the 8 studies, all but one were done in Europe (the eighth was in Hong Kong), most used a FIT cutoff of 15 to 20 mcg Hb/g. Six studies included aspirin users, 4 included oral anticoagulant users, and 1 included NSAID users (3 of the studies included more than one medication). The authors found no consistent effect on PPV among users and nonusers of oral anticoagulants and NSAIDs, and in general the differences were small. For example, the pooled PPV for advanced neoplasia was 38.2% in aspirin/NSAID users and 39.4% in nonusers. The PPVs for colorectal cancer in oral anticoagulant users and nonusers were 5.7% and 6.2%, respectively. And the PPVs for advanced neoplasia in oral anticoagulant users and nonusers were 37.6% and 40.3%, respectively.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Levels of Evidence definitions from Essential Evidence Plus
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.