The website may be down at times on Saturday, November 30, and Sunday, December 1, for maintenance.
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 2018 Consistent with the Principles of the Choosing Wisely Campaign."
Clinical question
In patients with symptoms of shoulder impingement syndrome is subacromial decompression surgery more effective than sham arthroscopy or exercise therapy to decrease pain and improve function?
Bottom line
Despite being one of the most common orthopedic surgeries performed, subacromial decompression is not significantly better than physical therapy to treat patients with pain and limited function due to shoulder impingement. This study is backed up by a meta-analysis that found the same results (doi:10.3109/09638288.2014.907364). Get out the stretchy bands or hand weights: another meta-analysis (doi:10.1136/bjsports-2016-096515) demonstrated the benefit of shoulder exercises over other physical therapy modalities. (LOE = 1a)
Reference
Paavola M, Malmivaara A, Taimela S, et al, for the Finnish Shoulder Impingement Arthroscopy Controlled Trial (FIMPACT) Investigators. Subacromial decompression versus diagnostic arthroscopy for shoulder impingement: randomised, placebo surgery controlled clinical trial. BMJ2018;362:k2860.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (specialty)
Synopsis
These Finnish researchers enrolled 210 adults aged 35 to 65 years with a clinical presentation of shoulder impingement syndrome who, by magnetic resonance imaging, had no evidence of rotator cuff tear and who had not responded to 3 months of conventional treatment. The patients were first randomized to receive surgery or physical therapy using concealed allocation; patients tapped for surgery underwent diagnostic arthroscopy to rule out tears or other pathology and then, in the operating room, were randomized again to receive arthroscopic subacromial decompression or (to keep everyone unaware of treatment assignment) kept in the operating theater but without further intervention for the length of time of a typical decompression. After 2 years, patients in all 3 groups had a large decrease in reported pain, from approximately 75 to between 20 and 30 on a 100-point visual analog scale. Decompression was statistically better than exercise therapy, but the result would not be clinically relevant (a difference of at least 15 points) and was no different than diagnostic arthroscopy. There was also no difference in pain or function scores at earlier time points. The researchers did not attempt to stratify patients by degree of joint narrowing or by the presence of osteoarthritis or other morphology, and targeted therapy aimed at specific changes may have found a difference in treatment outcomes.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Do patients with subacromial shoulder pain for at least 3 months who are treated surgically have better outcomes than those who are treated without surgery?
Bottom line
In patients with subacromial shoulder pain of at least 3 months duration who receive physical therapy, surgical decompression is no better than arthroscopy without decompression in improving pain or function… and neither is much better than no invasive intervention at all. (LOE = 1b)
Reference
Beard DJ, Rees JL, Cook JA, et al. Arthroscopic subacromial decompression for subacromial shoulder pain (CSAW): a multicentre, pragmatic, parallel group, placebo-controlled, three-group, randomised surgical trial. Lancet 2018;391(10118):329-338.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (specialty)
Synopsis
These authors randomized adults with subacromial pain of at least 3 months duration into 1 of 3 groups: arthroscopic decompression of the acromion (n = 106), arthroscopy without decompression (n = 103), or no additional treatment (n = 104). Before enrollment, all patients underwent physical therapy and had at least one corticosteroid injection. The authors excluded patients with complete rotator cuff tears. A healthy percentage of the patients (23%, 42%, and 12%, respectively) allocated to decompression, arthroscopy only, and no treatment did not receive their assigned treatment by 6 months because they were already better! Additionally, approximately 15% of the patients did not complete 12 months of follow up. After 6 months and 1 year, the patients treated with either surgical decompression or arthroscopy without decompression had improvements in pain and function (as measured by the Oxford Shoulder Score) compared with patients who received no treatment, but the differences were not clinically important. Additionally, there was no difference between the decompression and arthroscopy without decompression groups. Two patients in each group developed adhesive capsulitis. The authors don't report on surgical complications.
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
Clinical question
Is low-dose amitriptyline effective for the management of chronic lower back pain?
Bottom line
When compared with a similarly mouth-drying placebo (the anticholinergic benztropine), low-dose amitriptyline provides no greater improvement after 3 months and 6 months of treatment in patients with well-established low back pain. (LOE = 2b)
Reference
Urquhart D, Wluka A, van Tulder M, et al. Efficacy of low-dose amitriptyline for chronic low back pain. JAMA Intern Med 2018;178(11):1474-1481.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
These researchers recruited (through general advertising) 146 patients with a history of low back pain. The patients had an average of more than 14 years of back pain. Average pain scores at the start of the study were in the mild range (40 to 43 on a scale of 0 to 100) and disability scores were also low (7.5 to 8.2 on a scale of 0 to 23). The patients were randomized, using concealed allocation, to receive either amitriptyline 25 mg or benztropine 1 mg at bedtime. Benztropine causes similar side effects to amitriptyline, but is not known to be effective for low back pain. Analysis was by intention to treat, though there were significant dropouts in both groups (18% to 20%), most due to side effects. At both 3 months and 6 months, average pain and disability scores were decreased in both groups, but neither reduction was clinically significant. There was no significant difference between reports of pain at 3 months or 6 months. Disability scores were better at 3 months in the amitriptyline group but not at 6 months. There was also no difference in patients' global assessment or general health status. The study did not have the power to find a difference if one existed, though the overall improvement, even if different, is not likely to be clinically relevant.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
In patients with mild asthma, is as-needed use of an inhaled steroid plus a short-acting beta-agonist as effective as daily use of an inhaled steroid?
Bottom line
Increasingly, studies are showing that as-needed use of an inhaled steroid plus a short-acting beta-agonist (SABA) is as effective or very nearly as effective as the daily use of an inhaled steroid, with much lower cumulative steroid doses. The small benefit in terms of weeks with good symptom control with daily steroid use must be balanced against the cost and long-term effects. (LOE = 1b)
Reference
O'Byrne PM, FitzGerald JM, Bateman ED, et al. Combined budesonide-formoterol as needed in mild asthma. N Engl J Med 2018;378(20):1865-1876.
Study design: Randomized controlled trial (double-blinded)
Funding source: Industry
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
In this industry-sponsored trial, mild asthma was defined in a patient when it could be controlled by a daily low-dose steroid inhaler or leukotriene inhibitor, but not adequately controlled by intermittent SABAs. In this study, 3849 patients, 12 years and older, with mild asthma were randomized to receive 1 of 3 strategies: (1) placebo inhaler twice daily plus terbutaline 0.4 mg inhaler, as needed; (2) placebo inhaler twice daily plus budesonide 160 mcg/formoterol 4.5 mcg, as needed; and (3) budesonide 160 mcg twice daily plus terbutaline 0.4 mg inhaler, as needed. Patients were enrolled, given a 2- to 4-week run-in period of as-needed terbutaline only, and were followed up for one year. The mean age of patients was 40 years, 20% had a severe exacerbation in the year prior to enrollment, the mean pre-treatment FEV1 was 84%, and 87.4% completed the trial. Patients used a diary to record symptoms. Those in the budesonide maintenance group had a higher percentage of weeks with good control than the other 2 groups that used only as-needed inhalers (44% vs 34% for prn budesonide/formoterol vs 31% for prn terbutaline; P < .05). Over one year, this amounts to 5 additional weeks of good control for patients using a maintenance steroid inhaler compared with those using budesonide/formoterol as needed only. Both the budesonide maintenance and the budesonide/formoterol groups had significantly lower rates of severe exacerbations than the terbutaline only group (0.09 vs 0.07 vs 0.20, respectively). For moderate to severe exacerbations, the rates were also significantly lower for the regimens containing budesonide than for as-needed terbutaline only (0.14 vs 0.36). The overall steroid dose given to patients in the as-needed budesonide/formoterol group was approximately one-fifth that of those using maintenance steroid therapy.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Is as-needed use of budesonide plus formoterol similarly effective to daily maintenance with budesonide plus as-needed terbutaline in patients with mild asthma?
Bottom line
As-needed use of budesonide plus formoterol is as effective as the daily use of maintenance budesonide plus as-needed terbutaline at preventing severe exacerbations, and results in a much lower cumulative steroid dose. (LOE = 1b-)
Reference
Bateman ED, Reddel HK, O'Byrne PM, et al. As-needed budesonide-formoterol versus maintenance budesonide in mild asthma. N Engl J Med 2018;378(20):1877-1887.
Study design: Randomized controlled trial (double-blinded)
Funding source: Industry
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
This industry-sponsored noninferiority trial included 4215 patients, of whom 4176 had data available for analysis. Their mean age was 41 years, and approximately 50% controlled their asthma by using a daily inhaled glucocorticoid during the previous year; the other half had uncontrolled asthma using a short-acting beta-agonist (SABA) alone. Fully 22% had a severe exacerbation during the previous year, defined as the need for at least 3 days of systemic steroids, hospitalization, or an emergency department visit, which seems high for "mild asthma." As in the similar SYGMA 1 trial published in the same issue of this journal, mild asthma was defined as asthma that is uncontrolled using only a SABA as needed, or well controlled using a low-dose steroid inhaler. After a run-in period during which patients used only an as-needed SABA (terbutaline 0.5 mg), the patients were randomized to receive: (1) placebo inhaler twice daily plus budesonide 200 mcg/formeterol 6 mcg, as needed; or (2) budesonide 200 mcg twice daily plus as-needed use of terbutaline 0.5 mg. This trial was initially designed as a superiority trial to show that one of the interventions was better than the other. However, a lower-than-expected rate of exacerbations and a higher-than-expected rate of adherence to the daily inhaled steroid hurt the power, so the authors moved the goalposts midgame and declared it a noninferiority trial. Noninferiority was defined as no more than a 20% increase in the number of severe exacerbations. Although it's not good research practice to change goals midstream, in some ways this is a more interesting research question, and the noninferiority margin the authors chose seems clinically reasonable. After one year, there was indeed no difference between the groups regarding the likelihood of a severe exacerbation (0.11 for as-needed use and 0.12 for daily steroid inhaler per patient per year). There was also no differences between groups regarding the time to a first severe exacerbation or regarding different types of severe exacerbations (ie, hospitalization or emergency department visit). Adverse events between groups were similar. Not surprisingly, patients in the as-needed inhaler group had only about one-fourth the total steroid dose during the study period.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Are broad-spectrum antibiotics the preferred treatment in children with acute respiratory tract infections?
Bottom line
Broad-spectrum antibiotics are no more effective than narrow-spectrum antibiotics for treating acute respiratory tract infections in infants and children, and adverse events are significantly more common in children treated with broad-spectrum antibiotics. (LOE = 2b-)
Reference
Gerber JS, Ross RK, Bryan M, et al. Association of broad- vs narrow-spectrum antibiotics with treatment failure, adverse events, and quality of life in children with acute respiratory tract infections. JAMA 2017;318(23):2325-2336.
Study design: Cohort (retrospective)
Funding source: Government
Setting: Population-based
Synopsis
These investigators collected data both retrospectively and prospectively from a network of pediatric primary care practices on outcomes of infants and children, aged 6 months to 12 years, who met international standards for the diagnosis of acute respiratory tract infection, including otitis media, group A streptococcal pharyngitis, and sinusitis. Exclusion criteria included not receiving a prescription for an oral antibiotic, antibiotic use in the last 30 days, and being younger than 3 years with a diagnosis of group A streptococcal pharyngitis. The children who were prescribed broad-spectrum antibiotics, including amoxicillin-clavulanate, cephalosporins, and macrolides, were defined as exposed; children who were prescribed narrow-spectrum antibiotics, including penicillin and amoxicillin, were defined as unexposed. The authors do not specifically state whether the individuals who assessed outcomes remained masked to group assignments. Of the 30,159 children in the retrospective cohort that met inclusion criteria with complete data, 4307 (14%) were prescribed broad-spectrum antibiotics. Broad-spectrum antibiotic use was not significantly associated with a lower rate of treatment failure compared with narrow-spectrum antibiotics (3.4% vs 3.1%, respectively). Similarly broad-spectrum antibiotics were not associated with a clinically significant difference in quality of life scores compared with narrow-spectrum antibiotics. However, broad-spectrum antibiotics were significantly associated with a higher risk of reported adverse events compared with narrow-spectrum antibiotics (3.7% vs 2.7%, respectively, as documented by clinicians, and 35.6% vs 25.1%, respectively, as documented by the parents and/or patients). Adverse events included diarrhea, candidiasis, rash, other unspecified allergic reactions, and vomiting.
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
In patients with chronic rhinosinusitis, does the addition of budesonide to a saline irrigation solution result in further improvement in symptoms?
Bottom line
This study showed that patients with chronic rhinosinusitis who continue to use a saline nasal wash (NeilMed) will often experience an improvement in symptoms that can be clinically meaningful, but the addition of the corticosteroid budesonide has yet to be shown to provide extra benefit. (LOE = 2b)
Reference
Tait S, Kallogjeri D, Suko J, Kukuljan S, Schneider J, Piccirillo JF. Effect of budesonide added to large-volume, low-pressure saline sinus irrigation for chronic rhinosinusitis. A randomized clinical trial. JAMA Otolaryngol Head Neck Surg 2018;144(7):605-612.
Study design: Randomized controlled trial (double-blinded)
Funding source: Self-funded or unfunded
Allocation: Uncertain
Setting: Outpatient (specialty)
Synopsis
These researchers recruited 80 patients with chronic rhinosinusitis (2 or more symptoms, including mucopurulent drainage, nasal obstruction, facial pain, and decreased sense of smell at least 12 weeks) to be randomized, allocation concealment unknown, to receive treatment using a large-volume saline sinus irrigation with either placebo or budesonide 1 mg once daily for 30 days. The patients, average age 51 years, had a Sino-Nasal Outcome Test (SNOT-22) score of 44.1 out of a possible 110. A significant number of patients dropped out (23%), leaving 61 to be evaluated. The average decrease in scores was 20.7 points in the treated group and 13.6 points in the control group, which was not statistically significant. More participants in the treated group (79%) received a clinically important benefit of at least a 9-point improvement than in the saline-only group (59%; not statistically different). This small study, with significant dropouts, did not have the power to find a difference if one exists. I'm a little grumpy that the authors did not give specific data to judge the degree of benefit beyond a 9-point improvement for the responders.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Is the new definition of hypertension valuable to patients?
Bottom line
After the publication of guidelines from the American College of Cardiology and the American Heart Association that redefined the definition of high blood pressure to be higher than 130/80 mm Hg, placing an additional 31 million American adults under the hypertension tent, the authors of this commentary stopped to ask, "What makes a disease a disease?" They point out that using a new definition of 130/80 mm Hg will now identify many people (80%) as hypertensive, but they will not need treatment. Those patients who qualify for treatment only under the new definition will be exposed to the risk of medication side effects though they are at low risk of cardiovascular disease and do not stand to benefit from treatment. Two US groups—the American College of Physicians and the American Academy of Family Physicians—have stated that they'll keep their existing definition of 140/90 mm Hg, thank you very much, consistent with most of the rest of the world. (LOE = 5)
Reference
Bell KJ, Doust J, Glasziou P. Incremental benefits and harms of the 2017 American College of Cardiology/American Heart Association high blood pressure guideline. JAMA Intern Med 2018 doi:10.1001/jamainternmed.2018.0310 . [Epub ahead of print]
Study design: Other
Funding source: Self-funded or unfunded
Setting: Not applicable
Synopsis
What makes a disease a disease? Can mere mortals change disease definitions? To paraphrase Humpty Dumpty, a "disease" means just what we choose it to mean, neither more nor less. In 2017, two US cardiology societies decided to change the definition of hypertension, lowering the threshold to 130/80 mm Hg, largely based on the results of a single study (SPRINT) that did not measure blood pressure in the typical way. To evaluate the risk of overdiagnosis occurring as a result, the authors of this commentary used a checklist for redefining disease (JAMA Intern Med 2017;177(7):1020-1025). Here it is: (1) What is the difference between definitions? The new definition elevates the status of pre-hypertension (> 80/130 mm Hg) to hypertension and suggests drug treatment in patients at high risk. (2) How will the new definition change the prevalence of the disease? It adds 31 million more US adults to the roles and qualifies 4.2 million for treatment. (3) Why change the definition? The authors presume it's because of the results of the SPRINT trial. (4) How well does the new definition predict clinically important outcomes? Observational data suggest cardiovascular mortality doubles with every 20-mm Hg increase in systolic blood pressure, starting at 115 mm Hg. (5) How reproducible is the measurement of blood pressure? Systolic blood pressure can easily vary 10 mm Hg over repeated measures in the same person, So, not very good. (6) What are the additional benefits of including more people? It all depends on the patients' baseline risk. One estimate suggests 80% of people with newly diagnosed hypertension will not benefit (Circulation 2018;137(2):109-118). (7) What harms can come from the new definition? Fear and anxiety over having a disease; inability to get health insurance due to a pre-existing condition; risks of treatment, (8) What are the net benefits and harms to the new definition? For patients at low risk, there is no benefit to the change and some increased risk. At patients at high risk, it may be beneficial. In patients older than 80 years, those with diabetes, and those with renal disease, the benefits and risks are offsetting.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
How often does an elevated troponin level signify acute myocardial infarction?
Bottom line
Unnecessary cardiac troponin testing for patients with a low likelihood of myocardial infarction (MI) will mislead more than enlighten. Approximately 1 in 8 adults presenting to an emergency department for any reason will have an elevated serum troponin level if they are unfortunate enough to be tested, though only 1.6% will have evidence of MI. Even in patients for whom MI is suspected, troponin will be elevated without evidence of MI up to 84% of the time in the United States. Only when patients have a triad of chest pain, ischemic changes noted on cardiography, and a history of ischemic heart disease will an elevated serum troponin level consistently identify MI. (LOE = 2b)
Reference
Shah AS, Sandoval Y, Noaman A, et al. Patient selection for high sensitivity cardiac troponin testing and diagnosis of myocardial infarction: prospective cohort study. BMJ 2017;359:j4788.
Study design: Cohort (retrospective)
Funding source: Foundation
Setting: Emergency department
Synopsis
To conduct this study, investigators enrolled consecutive patients drawn from 3 different groups. The first group consisted of every adult presenting to a United Kingdom emergency department for any reason for whom blood was drawn (n = 1054). This blood was used to determine their cardiac troponin level. The second and third groups comprised patients presenting to all emergency departments in the United Kingdom (n = 5815) and one emergency department in the United States (n = 1631) in whom high sensitivity troponin testing was requested by the attending clinician. To determine whether an MI occurred, 2 physicians independently reviewed all clinical information, including noninvasive and invasive investigations and outcomes occurring within 30 days of presentation. In the first group of unselected patients, cardiac troponin was elevated in approximately 1 in 8 patients (13.7%), though the prevalence of MI was only 1.6%. In patients specifically selected for troponin testing (presumably because they were symptomatic), the prevalence of MI was 14.5% in the United Kingdom but only 4.2% in the United States. For these patients the positive predictive value of an elevated troponin level was 59.7% in the United Kingdom but only 16.4% in the United States. The presence of all three indicators of MI—chest pain, ischemia on electrocardiography, and a history of ischemic heart disease—dramatically increased the positive predictive value of the test. A negative troponin test result in any of the populations effectively ruled out MI (negative predictive value 100%).
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
In women with postmenopausal vaginal symptoms, is local estrogen treatment better than vaginal lubricant or placebo?
Bottom line
Time for a rethink the idea that vaginal atrophy due to diminished estrogen is the cause of vaginal symptoms associated with menopause. Vaginal estradiol (Vagifem) is no more effective than a nonprescription vaginal lubricant (Replens Long-Lasting Vaginal Moisturizer) or placebo lubricant in the treatment of women with painful intercourse, vaginal dryness, or other symptoms usually associated with menopause. (LOE = 1b)
Reference
Mitchell CM, Reed SD, Diem S, et al. Efficacy of vaginal estradiol or vaginal moisturizer vs placebo for treating postmenopausal vulvovaginal symptoms. A randomized clinical trial. JAMA Intern Med2018;178(5):681-690.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
This study evaluated the effect of a vaginal estradiol 10 mcg, a nonprescription vaginal lubricant, and matching placebo tablet and lubricant in 302 postmenopausal women with moderate to severe symptoms of vulvovaginal itching, pain, dryness, irritation, or pain with penetration. Estradiol was used daily for 2 weeks, then twice weekly; vaginal moisturizer was used every 3 days for 12 weeks. Results were analyzed using a modified intention-to-treat analysis, including women who returned for evaluation at 1 month and 3 months after beginning treatment, regardless of their continued use of treatment, At the beginning of the study women were asked to identify their most bothersome symptom, which was pain with vaginal penetration (60%) or vaginal dryness (21%) for most women; the main outcome measured was the effect of treatment on this outcome. Treatment with estradiol, lubricant, or placebo produced similar results, decreasing the average score (on a scale of 0 to 3) from 2.4 ("moderate" severity) at the beginning of treatment to 1.0 ("mild"). Half the women in each group, including the placebo group, had a clinically significant drop in scores with treatment. Sexual function, as measured by the Female Sexual Function Index, improved in all groups, with nearly half the women in all groups improving from "frequently" or "always" distressed about their sex life to "rarely" or "never" distressed about their sex life.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Do long-acting insulin analogs, such as glargine (Lantus) or detemir (Levemir), reduce the risk of clinically significant hypoglycemia compared with NPH insulin?
Bottom line
This study found that, compared with expensive long-acting insulin analogs costing 2 to 10 times as much, human neutral protamine Hagedorn (NPH) insulin results in a similar number, if not fewer, episodes of severe hypoglycemia that result in emergency department visits and hospitalizations. NPH insulin also improves glycemic control as well, if not better, than insulin analogs. In a previous report (Singh SR, et al. CMAJ 2009;180(4):385-96), overall quality of life was also similar with NPH insulin or insulin analogs. Compared with long-acting insulin analogs, NPH insulin is as Safe if not safer, equally Tolerated, equally or more Effective, and at a much lower Price (STEP). One in 4 adults with diabetes either stop or cut back significantly on their insulin because they can't afford it. (LOE = 2b-)
Reference
Lipska KJ, Parker MM, Moffet HH, Huang ES, Karter AJ. Association of initiation of basal insulin analogs vs neutral protamine Hagedorn insulin with hypoglycemia-related emergency department visits or hospital admissions and with glycemic control in patients with type 2 diabetes. JAMA2018;320(1):53-62.
Study design: Cohort (retrospective)
Funding source: Government
Setting: Population-based
Synopsis
Marketing efforts have convinced most clinicians that long-acting insulin analogs, such as glargine and detemir, reduce the risk of hypoglycemia and are thus safer than traditional NPH insulin. These investigators analyzed data from 2006 and 2015 from multiple patient and prescription registries with the Kaiser Permanente of Northern California. Outcomes of interest included pharmacy use; laboratory results; and outpatient, emergency department, and hospitalization diagnoses of diabetes and related complications. The inception cohort consisted of 25,489 adults, 19 years or older, with type 2 diabetes who were initiating basal insulin therapy without any insulin prescription fills during the prior 12 months. Results were analyzed after controlling for multiple potential confounders, including demographics, index year, clinician specialty, comorbidity index, chronic kidney and/or liver disease, visual impairment, history of depression, glycemic control, history of severe hypoglycemia episodes requiring third-party intervention, and medication nonadherence. The risk of a subsequent severe hypoglycemic episode resulting in an emergency department visit or hospital admission was nonsignificantly lower in patients who initiated NPH insulin at baseline compared with those initiating insulin analogs (8.8 vs 11.9 events per 1000 person-years, respectively). In addition, glycemic control was significantly more improved in patients using NPH insulin versus insulin analogs (difference in HbA1C -0.22%; 95% CI -0.09% to -0.37%).
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
Which is the better oral pain reliever for children with postoperative pain: ibuprofen or morphine?
Bottom line
This carefully designed and adequately powered study found no difference in pain reduction between ibuproften and oral morphine in children with postoperative pain. Adverse effects, however, were much more likely with morphine. (LOE = 1b)
Reference
Poonai N, Datoo N, Ali S, et al. Oral morphine versus ibuprofen administered at home for postoperative orthopedic pain in children: a randomized controlled trial. CMAJ 2017;189(40):E1252- E1258.
Study design: Randomized controlled trial (double-blinded)
Funding source: Foundation
Allocation: Concealed
Setting: Inpatient (any location) with outpatient follow-up
Synopsis
As concern increases about overly broad uses of opiates, it is good to see studies that evaluate their effectiveness in different populations. This study identified 154 children, aged 5 years to 17 years, who underwent an outpatient orthopedic surgical procedure (most commonly hardware removal, open reduction and internal fixation of a fracture, or arthroscopy). The patients were randomized to receive up to 8 doses, given 6 hours apart, of either 0.5 mg/kg morphine or 10 mg/kg ibuprofen. Pain was assessed by the patient immediately before and 30 minutes after each dose of medication using the well validated 10-point FACES scale; a difference of at least 1 point on this scale is considered to be clinically meaningful. Medication was given using a "double-dummy" design, so each patient simulataneously got one active medication and one placebo. Groups were similar at baseline and analysis was by intention to treat. Both medications reduced pain by approximately 1 point at each dose, with no difference between groups in efficacy. There was a significantly greater risk of adverse events in the morphine group, primarily nausea, vomiting, drowsiness, and dizziness (number needed to treat to harm for any adverse event = 3).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Does the addition of oil to the bath decrease the symptoms of children with eczema?
Bottom line
Wow. According to a cited review in this paper, bath additives account for one-third of the costs of treating eczema in the United Kingdom. This study—conceived of and conducted by the James Lind Alliance, a group aimed at addressing uncertainties of routine clinical practice—found no improvement with the addition of oil to the bath of a child with eczema. (LOE = 1b)
Reference
Santer M, Ridd MJ, Francis NA, et al. Emollient bath additives for the treatment of childhood eczema (BATHE): multicentre pragmatic parallel group randomised controlled trial of clinical and cost effectiveness. BMJ 2018;361:k1332.
Study design: Randomized controlled trial (nonblinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (primary care)
Synopsis
The investigator enrolled 483 children between the ages of 1 year and 11 years with atopic dermatitis, recruited from 96 general practices in the United Kingdom. The children were randomized, using concealed allocation, to a usual care group or a group that, in addition to usual care, were instructed to add an oil-based emollient of their choice to their bath water. Oilatum was used by 45%, 26% used Aveeno bath oil, 4.5% used Balneum bath oil, and 30% used another type of bath product. Most (74%) carers reported using the bath oil at every bath. Interestingly, two-thirds of children in each group were bathed at least every other day (14% were bathed at least daily). Using the 7-question patient-oriented eczema measure (0 - 7: mild; 8 - 16: moderate; 17 - 28: severe), scores improved equally in both groups, from a baseline of 9.5 to 10.1 to an average weekly score of 7.5 to 8.4 over 16 weeks of treatment. Scores were minimally better (-2.27; a clinically important difference is at least 3 points) with bath oil for children who bathed 5 or more times a week.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
What is a better predictor of mortality: ambulatory or office-based measurement of blood pressure?
Bottom line
This study supports the guidelines recommending that treatment decisions be based on ambulatory blood pressure (BP) measurements rather than in-office BP results. The difference between the 2 measurements in this cohort was 19/11 mm Hg, which is enough to change the decision to prescribe a medication at all, or to add a second or third medication. (LOE = 2b)
Reference
Banegas JR, Ruilope LM, de la Sierra A, et al. Relationship between clinic and ambulatory blood-pressure measurements and mortality. N Engl J Med 2018;378:1509-1520.
Study design: Cohort (prospective)
Funding source: Government
Setting: Population-based
Synopsis
How we measure things matters: For example, nonfasting lipid levels are a better predictor of mortality than fasting lipid levels. Recent guidelines for hypertension, including from the U.S. Preventive Services Task Force, have emphasized the need to confirm elevated BPs in most patients using some form of ambulatory BP monitoring. This study used data from a large Spanish hypertension registry to look at the association between clinic BPs, ambulatory blood BPs, and mortality. The registry includes adults with an indication for ambulatory BP monitoring, such as suspected white coat hypertension, borderline or labile hypertension, or hypertension refractory to treatment. The registry supplies data on clinic BPs, measured by automated devices after 5 minutes of seated rest, and 24-hour ambulatory BP measurements. These data were linked to national vital statistics databases to determine cardiovascular and all-cause mortality. The analysis was adjusted for comorbidities, age, sex, tobacco use, and body mass index. The mean age of patients was 58 years, 58% were male, and only 11% had a diagnosis of cardiovascular disease. During a median 4.7 year follow-up, there were a total of 3808 deaths including 1295 cardiovascular deaths. The mean ambulatory BP was 129/76, compared with 148/87 in the clinic. Recall, the clinic BPs were measured by an automated device after 5 minutes of rest, yet they were still far higher than the ambulatory measurements. In the fully adjusted model that adjusted for clinic BPs, the hazard ratio for all-cause mortality was 1.58 (95% CI 1.56 - 1.60) for the ambulatory systolic BP versus 1.02 (1.00 - 1.04) for the clinic systolic BP adjusted for ambulatory BP. A similar pattern was seen for diastolic BPs. The inflection point for an increase in both cardiovascular and all-cause mortality is at a systolic BP of 140 to 160. Mortality was not increased in patients with controlled hypertension, but was increased in those with both white-coat and masked (normal in clinic, abnormal at home) hypertension.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Is a single office blood pressure measurement reliable to assess hypertension?
Bottom line
Don't rely on a single blood pressure measurement. The first blood pressure reading taken during an office visit will be substantially different than subsequent readings in almost half of typical patients, and if relied upon will result in 1 in 8 patients being falsely labeled as hypertensive. (LOE = 1b)
Reference
Burkhard T, Mayr M, Winterhalder C, Leonardi L, Eckstein J, Vischer AS. Reliability of single office blood pressure measurements. Heart 2018;104(14):1173-1179.
Study design: Cross-sectional
Funding source: Self-funded or unfunded
Setting: Outpatient (primary care)
Synopsis
The authors enrolled 1000 consecutive patients who presented for internal medicine, obstetric, or gynecologic care, though only 802 patients completed the study. Each patient, after 5 minutes of rest, had 4 consecutive blood pressure measurements, 2 minutes apart, by 1 of 13 trained physicians. The first systolic blood pressure was more than 10 mm Hg higher than the mean of subsequent measurements in 23.9% of patients; in total, 45.9% of patients had a systolic difference of more than 5 mm Hg. Similarly, diastolic blood pressures were more than 10 mm Hg different in 4.4% of patients; in total, 21.6% of patients had a difference of more than 5 mm Hg. More important, hypertension would have been diagnosed in error in 1 in 8 patients (12%) if only the first measurement had been obtained, and 2% of patients would have had their hypertension undiagnosed by the single measurement.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Alex Shreiber, PharmD
Clinical question
In patients with high blood pressure, does a second reading show lower results?
Bottom line
If you're not rechecking high blood pressures, you should. In fact, set your electronic health record to remind you to do it. In this large study, when reminded, clinicians rechecked elevated blood pressures 83% of the time, finding a median drop in blood pressure of 8 mm Hg during the same visit. That drop is equivalent to a typical reduction in blood pressure with pharmacologic treatment over time, and resulted in one-third fewer patients being labeled with high blood pressure at that visit. (LOE = 2b)
Reference
Einstadter D, Bolen SD, Misak JE, Bar-Shain DS, Cebul RD. Association of repeated measurements with blood pressure control in primary care. JAMA Intern Med 2018 doi:10.1001/jamainternmed.2018.0315. [Epub ahead of print]
Study design: Cohort (prospective)
Funding source: Government
Setting: Outpatient (primary care)
Synopsis
This study was conducted in primary care offices of a large US health system. The electronic health record was set to remind clinicians to recheck the blood pressure of a patient when a value greater than 140/90 mm Hg was documented. The reminder worked: Clinicians rechecked high blood pressures 83% of the time. The authors then evaluated the effect of this simple intervention on 38,260 patients, average age 61 years, 39% of whom had a high initial reading. With repeated measurement, the median drop in blood pressure was 8 mm Hg (interquartile range 2 mm Hg - 17 mm Hg) and 36% of patients no longer had a blood pressure of 140/90 mm Hg or higher.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
At what systolic blood pressure should we begin treatment for the most benefit?
Bottom line
Beginning antihypertensive treatment when the systolic blood pressure (SBP) is greater than 140 mm Hg delays death and prevents major cardiovascular events in some people without pre-existing heart disease; in patients with existing heart disease it prevents further events, but does not extend life. These results may appear to conflict with those from SPRINT trial, which found benefit with lowering SBP to below 120 mm Hg. However, the SPRINT investigators measured blood pressure using automated devices which give readings 10 mm Hg to 20 mm Hg lower than typical office measurements. So, the goal of less than 120 mm Hg in the SPRINT study is likely to be very similar to the goal of less than 140 mm Hg in this study. (LOE = 1a)
Reference
Brunstrom M, Carlberg B. Association of blood pressure lowering with mortality and cardiovascular disease across blood pressure levels. A systematic review and meta-analysis. JAMA Intern Med2018;178(1):28-36.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Foundation
Setting: Various (meta-analysis)
Synopsis
The authors followed PRISMA guidelines to search 3 databases, including Cochrane CENTRAL, as well as reference lists of identified studies to identify all randomized trials with at least 1000 patient-years of follow-up that compared drug treatment with placebo or compared blood pressure targets against one another. Two researchers independently extracted the data and assessed the quality of the research (more than two-thirds of the studies had a low risk of bias). They identified 74 studies enrolling 306,273 patients (60.1% men, average age 63.6 years). In patients without pre-existing heart disease (ie, primary prevention), lowering SBP that was initially greater than 140 mm Hg decreased the risk of death (relative risk [RR] = .93, 95% CI -.88 to 1.0 if SBP > 160 mm Hg; RR = 0.87, .75 to 1.00 if SBP 140 - 159 mm Hg) and major cardiovascular events (RR = .78, .7 to .87 if > 160 mm Hg; RR = .88, .8 - .96 if 140 - 159 mm Hg). Treating SBP that was initially less than 140 mm Hg did not affect morbidity or mortality. In patients with previous coronary heart disease and a mean SBP of 138 mm Hg, treatment reduced the risk for further major cardiovascular events (RR = .9; .84 to .97), but did not extend life. There was a high degree of heterogeneity among these trial results, reducing our confidence in the results. There was some evidence of publication bias in studies that evaluated the effect on major cardiovascular events, meaning that studies failing to show a difference in outcomes were not published.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Is lower systolic blood pressure associated with better outcomes in elderly patients who take antihypertension medications?
Bottom line
In this small cohort study of patients older than 85 years, lower systolic blood pressure during treatment with antihypertensive medications is associated with higher death rates and greater cognitive decline. (LOE = 1b-)
Reference
Streit S, Poortvliet RKE, Gussekloo J. Lower blood pressure during antihypertensive treatment is associated with higher all-cause mortality and accelerated cognitive decline in the oldest-old data from the Leiden 85-plus Study. Age Ageing 2018;47(4):545-550.
Study design: Cohort (prospective)
Funding source: Government
Setting: Population-based
Synopsis
Are you tired of all the ping-ponging, guideline-based blood pressure targets? Unfortunately, this study won't improve your fatigue. These researchers assembled a cohort of 570 residents of Leiden in the Netherlands who turned 85 years of age between 1997 and 1999. They excluded people who died within 3 months of enrollment and those who had no blood pressure measurement at baseline. At baseline, and periodically over the course of 5 years of follow up, the researchers collected all kinds of information: sociodemographics, medical diagnoses, medications, mental status, grip strength (as a proxy for frailty), blood pressure, and so forth. They assessed the main outcome—death from any cause—by using municipal records. Slightly fewer than half of the residents (44%) took antihypertensive medications at baseline; these patients were more likely to have other cardiovascular disorders than those not taking antihypertensive medications (62% vs 36%). During the 5 years of follow-up, 263 (46%) participants died. For those taking antihypertensive medications, all-cause mortality was significantly higher with decreasing systolic blood pressure (hazard ratio 1.29 per 10 mmHg lower systolic blood pressure; 95% CI 1.15 - 1.46). For the residents who were not taking antihypertensive medications, there was no significant correlation between systolic blood pressure and all-cause mortality. Additionally, the patients taking antihypertensives had more rapid cognitive decline with lower systolic blood pressure. Although many explanations for the differences in treatment thresholds are given by the various guidelines, one is how we value clinical trial versus observational data: The guidelines that promulgate lower blood pressure targets are more likely to value observational data. The data from this study are subject to many of the biases inherent in cohort studies, but they should moderate the enthusiasm for lower blood pressure targets.
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
Clinical question
Are short courses of antibiotics as effective as longer courses for common outpatient infections?
Bottom line
Just about every time someone asks "Can I get away with a shorter course of antibiotics," the answer is "Yes, you can." Shorter courses reduce cost, and may reduce the likelihood of adverse events. (LOE = 1a)
Reference
Dawson-Hahn EE, Mickan S, Onakpoya I, et al. Short-course versus long-course oral antibiotic treatment for infections treated in outpatient settings: a review of systematic reviews. Fam Pract2017;34(5):511-519.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Government
Setting: Various (meta-analysis)
Synopsis
This is a relatively new kind of study: a systematic review of systematic reviews, also called a systematic overview. The authors searched 5 databases and identified 9 systematic reviews that compared the duration of antibiotic therapies for a common outpatient infection. The reviews included between 2 studies and 17 studies, with a total of between 395 and 5763 patients. The best studied conditions were urinary tract infection (UTI), sinusitis, and community-acquired pneumonia (CAP). The authors found that, in children, 5 to 7 days was as good as 10 days for strep pharyngitis; 3 days was as good as 5 days for CAP; more than 2 days was as good as 7 or more days for otitis media, and 2 to 4 days was as good as 7 to 14 days for UTI. In adults, 3 to 7 days was as good as 6 to 10 days for acute bacterial sinusitis, 3 days was as good as 5 or more days for uncomplicated UTI in nonpregnant women, and 7 to 14 days was as good as 14 to 42 days for acute pyelonephritis. The authors also found that 7 or fewer days was as good as more than 7 days for CAP, and 3 to 6 days was as effective as 7 to 14 days for UTI in older women. There was some evidence that shorter courses resulted in fewer adverse events when treating acute otitis media in children and acute sinusitis in adults.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Which treatments are safe and effective for cough associated with the common cold?
Bottom line
Suck it up: you have a cold, it'll get better. That seems to be the bottom line from this expert panel report, which found little evidence of benefit for most commonly used medications for the self-limited condition of the common cold. Ultimately, physicians must often act in the absence of good evidence, and it is reasonable to recommend safe options for the treatment of cough even if the optimal evidence is not available. These treatments include honey, dextromethorphan, and possibly zinc for patients with cough. (LOE = 1a-)
Reference
Malesker MA, Callahan-Lyon P, Ireland B, Irwin RS, CHEST Expert Cough Panel. Pharmacologic and nonpharmacologic treatment for acute cough associated with the common cold: CHEST Expert Panel Report. Chest 2017;152(5):1021-1037.
Study design: Systematic review
Funding source: Other
Setting: Other
Synopsis
Although billed as an "expert panel report," this was really a hybrid of a systematic review, an "umbrella review" of published systematic reviews, and a guideline. The authors did a careful literature search, identifying randomized controlled trials (RCTs), as well as previous systematic reviews, and updated the searches of the published systematic reviews. They included any randomized trial of any treatment for acute cough in patients with an upper respiratory tract infection or the common cold. Although the authors said they looked for studies of herbal supplements, and included a few of them, the Cochrane Review of pelargonium sidioides was not included.The studies were assessed for quality using the Cochrane Risk of Bias tool for RCTs and a similar tool for systematic reviews, and excluded any studies at high risk of bias. The expert panel reviewed the evidence, then made 4 key recommendations. They concluded that there was insufficient evidence to make a recommendation for acetylcysteine or carbocysteine, and they recommend against the use of over-the-counter cold medications available in the United States to treat cough. They also recommend against the use of nonsteroidal anti-inflammatory drugs, given their lack of proven efficacy and, of course, their potential harms. Honey gets some love: The authors conclude that for children older than 1 year and adolescents with cough, honey is probably better than placebo or diphenhydramine, but not any better than dextromethorphan (…which they just told us not to use). Zinc had mixed evidence, and some of the benefit was attributed to underlying zinc deficiency in some countries, as well as the difficulty in masking patients to the treatment. In the end, the panel did not recommend the use of zinc. They also found no good evidence supporting or refuting the benefits of over-the-counter antitussives, expectorants, mucolytics, antihistamines, or combinations of any of them. Finally, the panel recommends against using codeine-containing medications in children.
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
Which treatments for subacute cough are effective?
Bottom line
The available evidence for treating patients with subacute cough is limited and fails to demonstrate meaningful improvements. (LOE = 1a-)
Reference
Speich B, Thomer A, Aghlmandi S, Ewald H, Zeller A, Hemkens LG. Treatments for subacute cough in primary care: systematic review and meta-analyses of randomised clinical trials. Br J Gen Pract2018;68(675):e694-e702.
Study design: Systematic review
Funding source: Foundation
Setting: Outpatient (any)
Synopsis
Subacute cough—defined as cough that lasts no more than 8 weeks, is not accompanied by radiographic evidence of pneumonia, and resolves on its own—is fairly common, especially following respiratory infections. These authors systematically searched PubMed and the Cochrane Central Register of Clinical Trials to identify randomized trials published in English that evaluated various treatments in patients at least 16 years old with subacute cough. The studies could have included drug and nondrug treatments, but excluded Chinese or Asian herbal remedies. They also hand-searched reference lists of included studies, relevant systematic reviews, and clinical practice guidelines. Two authors independently assessed the inclusion of studies and the risk of bias for each study. They used a third author to resolve disagreements. They were able to find only 6 trials with between 30 and 276 patients (median = 96). The studies included montelukast, inhaled albuterol (also called salbutamol) plus ipratropium, gelatin, inhaled corticosteroids (fluticasone propionate, budesonide), and opioids. Five of the studies compared treatment with placebo, one with usual care. Overall, the reporting of the studies made assessing their risk of bias difficult. The studies used a variety of cough severity scores. Although the authors identified statistically significant improvement of cough scores with some interventions, none were clinically important. Additionally, some interventions provided short-term improvement but none were sustained beyond 2 weeks. Five studies reported on adverse effects of treatment, which were mostly mild and ranged from 0% to 40% for active treatment and 0% to 27% for placebo or usual care.
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
Clinical question
Are opioid medications preferable for improving pain-related function in adults with severe chronic back, hip, or knee pain?
Bottom line
Nonopioid medications were at least as effective as opioid medications for improving pain-related function over 12 months in adults with severe chronic back pain or knee or hip osteoarthritis pain. The evidence that opioids are NOT superior to nonopioid medications for both chronic and acute pain continues to mount. The tough job will be getting patients and their clinicians to believe the evidence. (LOE = 1b)
Reference
Krebs EE, Gravely A, Nugent S, et al. Effect of opioid vs nonopioid medications on pain-related function in patients with chronic back pain or hip or knee osteoarthritis pain. The SPACE randomized clinical trial. JAMA 2018;319(9):872-882.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Outpatient (primary care)
Synopsis
For decades both patients and clinicians have believed/assumed that opioids are superior for reducing pain and improving function in patients with severe chronic pain. These investigators identified adults with chronic back pain or hip or knee osteoarthritis pain that rated at least moderately severe on a standard pain rating scale and persisted every day for at least 6 months. Patients with severe depression or posttraumatic stress disorder symptoms were NOT excluded. Study participants (N = 240) randomly received assignment (concealed allocation) to either an opioid or nonopioid pain management group. Patients in the opioid group started taking immediate release (IR) oral opioids with escalation to sustained-released (SA) oral opioids and finally to transdermal fentanyl, if needed. Titration continued to a maximum daily dose of 100 morphine-equivalent milligrams. Patients in the nonopioid medication group started with acetaminophen and nonsteroidal anti-inflammatory drugs, with step-up as needed to adjuvant oral medications (eg, amitriptyline, gabapentin) and topical analgesics (eg, capsaicin, lidocaine), and finally to pregabalin, duloxetine, and/or tramadol, if needed. Medication adherence was monitored by urine drug testing and with regular checking of a state prescription monitoring program. Individuals who assessed outcomes remained masked to treatment group assignment. Follow-up rates ranged from 90% to 98% of patients at 12 months. Mean age was 58.3 years (range = 21 – 80 years) and 13% were women. Using intention-to-treat analyses, there was no significant group difference in pain-related function over 12 months based on standard rating scales. Overall, pain intensity was significantly better in the nonopioid group over 12 months. Drop-outs due to adverse medication-related symptoms were significantly higher in the opioid group than in the nonopioid group (19% vs 8%, respectively). No deaths or diversions were detected in either group. Tramadol was dispensed to 11% of patients in the nonopioid group over the 12 months of follow-up.
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
What oral analgesic combinations are effective for reducing the pain of an acute extremity injury in adults in the emergency department?
Bottom line
In adults presenting to the emergency department with acute extremity pain severe enough to warrant radiologic investigation, ibuprofen plus acetaminophen was equally effective in reducing pain intensity at 2 hours compared with 3 different opioid and acetaminophen combination analgesics. In a similar study (Friedman BW, et al. JAMA 2015;314(15):1572-80), naproxen alone was as effective as naproxen plus oxycodone/acetaminophen or naproxen plus cyclobenzaprine for reducing pain from acute musculoskeletal low back pain. It's time we stopped believing that opioids are superior to nonsteroidal anti-inflammatory drugs for acute pain control. We'd save a lot of lives. (LOE = 1b)
Reference
Chang AK, Bijur PE, Esses D, Barnaby DP, Baer J. Effect of a single dose of oral opioid and nonopioid analgesics on acute extremity pain in the emergency department. A randomized clinical trial. JAMA 2017;318(17):1661-1667.
Study design: Randomized controlled trial (double-blinded)
Funding source: Government
Allocation: Concealed
Setting: Emergency department
Synopsis
Opioid use for just 3 days can significantly increase the risk of opioid dependence. These investigators identified adults, aged 21 to 64 years, presenting to the emergency department for acute extremity pain, defined as pain originating distal to and including the shoulder joint in the upper extremities and distal to and including the hip joint in the lower extremities. Eligible patients (N = 411) included those with an injury severe enough to require radiologic imaging according to the judgment of the attending physician. After baseline pain measurement, patients randomly received (concealed allocation assignment) identical capsules containing either ibuprofen 400 mg plus acetaminophen 1000 mg; oxycodone 5 mg plus acetaminophen 325 mg; hydrocodone 5 mg plus acetaminophen 300 mg; or codeine 30 mg plus acetaminophen 300 mg. Patients masked to their treatment group assignment self-assessed pain intensity using a verbal numerical rating scale from 0 (no pain) to 10 (worse pain imaginable). The minimum clinically important difference was predefined as a mean pain scale score of 1.3. Complete follow-up occurred for 100% of patients at 2 hours. Using intention-to-treat analysis, pain intensity significantly declined by 3.5 to 4.4 points at 2 hours compared with baseline in all treatment groups, but was not significantly different among the 4 groups. Pain intensity was also similarly reduced in all treatment groups at 1 hour and there were no group differences in the use of rescue analgesia. Even with post hoc analysis, no statistical difference was present for those with 10/10 pain (severe) and those with acute fractures.
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
Are anticonvulsants an effective treatment for low back pain?
Bottom line
The use of anticonvulsants like gabapentin for painful conditions has increased greatly in recent years. This review finds good evidence that these drugs are not an effective treatment for low back pain with or without radiculopathy, and are associated with an increased risk for adverse events. (LOE = 1a)
Reference
Enke O, New HA, New CH, et al. Anticonvulsants in the treatment of low back pain and lumbar radicular pain: a systematic review and meta-analysis. CMAJ 2018;190(26):E786-E793.
Study design: Meta-analysis (randomized controlled trials)
Funding source: Self-funded or unfunded
Setting: Various (meta-analysis)
Synopsis
Particularly in this era of heightened awareness of the potential harms of opioids, anticonvulsants are often prescribed for the treatment of painful conditions. Although there is evidence of their effectiveness primarily for peripheral and diabetic neuropathy, they are increasingly prescribed for other conditions, including low back pain. This systematic review included a comprehensive search of the literature and the authors identified 9 randomized trials (3 of which were crossover studies) that compared topiramate, pregabalin, or gabapentin with placebo in patients with low back pain with or without radiculopathy. They excluded studies of pregnant women; pre-operative patients; and patients with mixed conditions, such as neck and back pain. The trials were assessed for risk of bias, and only 1 was at high risk. The studies used a range of pain scales, so the standardized mean difference in pain scores between treatment and placebo groups was the primary outcome. The 9 studies reported a total of 14 comparisons, with only 2 showing a statistically significant benefit. One was a small study of high-dose gabapentin (3600 mg/day) in 43 patients with lumbar radicular pain, and the other was a study of 96 patients with low back pain who were given 300 mg topiramate 300 each day. The other topiramate study found no benefit, as did none of the other studies of pregabalin or gabapentin. Where results could be pooled, there was essentially no difference between groups. Where results could be pooled, there was essentially no difference between groups. There was no difference in serious adverse events: 4 in the pregabalin group, 6 in the placebo group (though these were reported in only 2 studies with a total of 423 patients). Any adverse events, however, were significantly more common with active treatment (relative risk 1.4; 95% CI 1.2 - 1.7).
Mark H. Ebell, MD, MS
Professor
University of Georgia
Athens, GA
Clinical question
In older people without a history of cardiovascular disease, is statin treatment associated with better outcomes
Bottom line
In this retrospective study, statin treatment in patients 75 years or older without pre-existing cardiovascular disease (CVD) did not change the likelihood of developing CVD or reduce any-cause mortality. However, patients aged 75 to 84 years with diabetes benefitted from treatment. (LOE = 2b)
Reference
Ramos R, Comas-Cufi M, Marti-Lluch R, et al. Statins for primary prevention of cardiovascular events and mortality in old and very old adults with and without type 2 diabetes: retrospective cohort study. BMJ 2018;362:k3359.
Study design: Cohort (retrospective)
Funding source: Government
Setting: Population-based
Synopsis
This study enrolled 46,864 patients 75 years or older with no cardiovascular disease from a population database in Spain. The patients had an average age of 76 years (63% were women) and were followed up for an average of 5.6 years. Of these, 6550 patients began statin treatment in the 18 months before the start of the study. In participants without diabetes there was no difference in the onset of CVD (hazard ratio [HR] 0.94; 95% CI 0.86 - 1.04) or the rate of mortality due to any reason (HR 0.98; 0.91 - 1.05). In patients 85 years or older, there also was no reduction in the likelihood of CVD (HR 0.93; 0.82 - 1.06) or all-cause mortality (HR 0.97; 0.90 - 1.05), However, in patients with diabetes who were between the ages of 75 and 84 years, the likelihood of developing CVD was reduced (HR 0.76; 0.65 - 0.89). All-cause mortality was decreased over an average 5.6 years, with one additional person alive for every 16 persons treated with a statin (number needed to treat [NNT] = 15.63; 9.5 - 49.6). The difference was not significant for any patients 85 years or older.
Allen F. Shaughnessy, PharmD, MMedEd
Professor of Family Medicine
Tufts University
Boston, MA
Clinical question
Is low-dose aspirin effective for the primary prevention of cardiovascular disease in moderate-risk patients?
Bottom line
In this study, after 5 years of treatment, patients at moderate risk of heart disease who took low-dose aspirin did not show a decrease in coronary events and all-cause mortality, and had slightly more, albeit minor, gastrointestinal bleeding. If you are confused by all the aspirin-related folderol of late, join the club. Using aspirin for primary prevention of cardiovascular disease is not a one-size-fits-all proposition. We need to risk-stratify patients according to benefits and harms and engage in shared decision-making with each patient. (LOE = 1b)
Reference
Gaziano JM, Brotons C, Coppolecchia R, et al, for the ARRIVE Executive Committee. Use of aspirin to reduce risk of initial vascular events in patients at moderate risk of cardiovascular disease (ARRIVE): a randomised, double-blind, placebo-controlled trial. Lancet 2018;392(10152):1036-1046.
Study design: Randomized controlled trial (double-blinded)
Funding source: Industry
Allocation: Concealed
Setting: Outpatient (any)
Synopsis
Low-dose aspirin for secondary prevention and in the face of acute coronary events is pretty much a slam dunk. But despite of years of research, several meta-analyses, and numerous guidelines, its use for primary prevention still seems to rile people up. These researchers point out that most of the recommendations are largely for patients whose 10-year risk of a coronary event exceeds 20% and the role of aspirin in patients of intermediate risk is less clear. So, they conducted a double-blind randomized trial of 100 mg aspirin daily (n = 6270) or placebo (n = 6276) in patients at moderate risk of coronary artery disease. The study participants were men at least 55 years of age or women at least 60 years of age with a 10% to 20% 10-year risk based on age, sex, smoking status, blood pressure, lipid concentrations, and family history. They excluded patients with diabetes and those at high risk for bleeding complications. Using intention-to-treat analysis, after 5 years the rate of events (a composite of myocardial infarction, stroke, cardiovascular death, unstable angina, or transient ischemic attack) was similar between the treatment groups (4.3% vs 4.5%, respectively). The overall death rate was the same (2.6%) in each group. The aspirin-treated patients had more bleeding events (1% vs 0.5%), although very few had moderate or severe gastrointestinal bleeding. The graphs in the paper demonstrate nearly a linear relationship in outcomes over time, so the projected 10-year outcomes indicate that 9% of the placebo-treated patients would have had a coronary event. Recall last month another study that suggested aspirin's effect was potentially influenced by weight and sex (Rothwell et al. Lancet 2018;392(10145):387-399).
Henry C. Barry, MD, MS
Professor
Michigan State University
East Lansing, MI
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.