Search date January 27 Clare Bradley-Stevenson, Paddy O'Neill, and Tony Roberts................................................... ABSTRACT INTRODUCTION: In the UK, about 3% of children under 3 years of age visit their GPs each year with acute otitis media (AOM), and 97% of these receive antibiotics. In the USA, AOM is the most common reason for outpatient antibiotic treatment. Without antibiotics, AOM resolves within 2 hours in about 6% of children, and within 3 days in about 8% of children. METHODS AND OUTCOMES: We conducted a systematic review and aimed to answer the following clinical questions: What are the effects of treatments for AOM in children; and what are the effects of interventions to prevent recurrence? We searched: Medline, Embase, The Cochrane Library and other important databases up to January 27. (BMJ Clinical Evidence reviews are updated periodically; please check our website for the most up-to-date version of this review). We included harms alerts from relevant organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA). RESULTS: We found 19 systematic reviews, RCTs, or observational studies that met our inclusion criteria. We performed a GRADE evaluation of the quality of evidence for interventions. CONCLUSIONS: In this systematic review we present information relating to the effectiveness and safety of the following interventions: analgesics, antibiotics, delayed antibiotics, immediate antibiotics, long-term antibiotic prophylaxis, longer courses of antibiotics, myringotomy, pneumococcal vaccination, tympanostomy with ventilation tubes. QUESTIONS What are the effects of treatments for AOM in children?........................................... 3 What are the effects of interventions to prevent recurrence of AOM in children?......................... 8 INTERVENTIONS TREATMENTS FOR AOM IN CHILDREN Likely to be beneficial PREVENTING RECURRENCE OF AOM IN CHILDREN Trade off between benefits and harms Analgesics................................ 3 Antibiotic prophylaxis (long term).............. 8 Trade off between benefits and harms Antibiotics (reduce symptoms more quickly than placebo but increse adverse effects)................... Choice of antibiotic regimen................... 5 Immediate compared with delayed antibiotic treatment.......................................... 5 Longer courses of antibiotics (reduce treatment failure in the short term but not the long term).......... 7 Likely to be ineffective or harmful Myringotomy............................... 7 Unlikely to be beneficial Vaccination (pneumococcal)................... 9 Likely to be ineffective or harmful Tympanostomy (ventilation tubes)............. 1 Covered elsewhere in Clinical Evidence see chronic suppurative otitis media see otitis media with effusion To be covered in future updates Influenza vaccines Key points AOM is characterised by sudden onset of earache with a cloudy or bulging erythematous eardrum caused by middle-ear infection. Middle-ear effusion without signs of infection lasting more than 3 months suggests otitis media with effusion ( glue ear'), while chronic suppurative otitis media is characterised by continuing middle-ear inflammation and discharge through a perforated eardrum. These disorders are assessed in other reviews in BMJ Clinical Evidence. The most common pathogens in AOM in the USA and UK are Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarhalis. In the UK, about 3% of children under 3 years of age visit their GP each year with AOM, and 97% of these receive antibiotics. In the USA, AOM is the most common reason for outpatient antibiotic treatment. Without antibiotics, AOM resolves within 2 hours in about 6% of children, and within 3 days in about 8% of children. Analgesics and topical anaesthetics may reduce earache when given with antibiotics. Antibiotics may lead to more rapid reduction in symptoms of AOM, but increase the risk of adverse effects. Antibiotics seem to reduce pain at 2 7 days, and may prevent development of contralateral AOM, but increase the risks of vomiting, diarrhoea, and rashes compared with placebo. BMJ Publishing Group Ltd 27. All rights reserved...................... 1..................... Clinical Evidence 27;8:31
Immediate antibiotic use seems to be most beneficial in children younger than 2 years with bilateral AOM. We do not know whether any one antibiotic regimen should be used in preference to another. Longer courses of antibiotics reduce short-term treatment failure, but have no benefit over the longer term compared with shorter regimens. Immediate use of antibiotics may reduce some, but not all, symptoms of AOM, but increases the risk of vomiting, diarrhoea, and rashes compared with delayed treatment. Myringotomy appears to be less effective than antibiotics in reducing symptoms. Tympanostomy with ventilation tube insertion leads to short-term reduction in the number of episodes of AOM, but increases the risk of complications. We found limited evidence that there was only a short-term benefit from tympanostomy with ventilation tubes, with possibly increased risks of tympanosclerosis. Long-term antibiotic prophylaxis may reduce recurrence rates; however, the possibility of adverse effects and antibiotic resistance should be taken into account. We do not know whether any one regimen should be used in preference to another to prevent recurrent attacks. In children aged from 2 months to 7 years, large scale pneumococcal vaccination strategies are unlikely to be effective. DEFINITION INCIDENCE/ PREVALENCE Otitis media is an inflammation in the middle ear. Subcategories include AOM, recurrent AOM, and chronic suppurative otitis media. AOM is the presence of middle-ear effusion in conjunction with rapid onset of one or more signs or symptoms of inflammation of the middle ear. AOM presents with systemic and local signs, and has a rapid onset. The diagnosis is made on the basis of signs and symptoms, principally earache in the presence of a cloudy or bulging eardrum (and immobility of the eardrum if pneumatic otoscopy is performed). Erythema is a moderately useful sign for helping establish the diagnosis. If the eardrum has a normal colour, then risk of AOM is low. [1] Uncomplicated AOM is limited to the middle-ear cleft. [2] The persistence of an effusion beyond 3 months without signs of infection defines otitis media with effusion (also known as glue ear ; see review on otitis media with effusion), which can arise as a consequence of AOM, but can also occur independently. Chronic suppurative otitis media is characterised by continuing inflammation in the middle ear causing discharge (otorrhoea) through a perforated tympanic membrane (see review on chronic suppurative otitis media). This review deals only with AOM in children. AOM is common, and has a high morbidity and low mortality in otherwise healthy children. In the UK, about 3% of children under 3 years visit their general practitioner with AOM each year, and 97% receive antimicrobial treatment. [3] By 3 months, 1% of children have had an episode of AOM. It is the most common reason for outpatient antimicrobial treatment in the USA. [] AETIOLOGY/ The most common bacterial causes of AOM in the USA and UK are Streptococcus pneumoniae, RISK FACTORS Haemophilus influenzae, and Moraxella catarrhalis. [3] Similar pathogens are found in Colombia. [5] There is some evidence that the predominant causative pathogen in recurrent AOM is changing from Streptococcus pneumoniae to Haemophilus influenzae after the release and widespread use of pneumococcal conjugate vaccine. [6] The established risk factors for recurrent AOM which are capable of being modified are the use of pacifiers, and care in daycare centres. Probable risk factors are privation of mother s milk, presence of siblings, craniofacial abnormalities, passive smoking, and presence of adenoids. [7] PROGNOSIS Without antibiotic treatment, AOM symptoms improve in 2 hours in about 6% of children, and in about 8% of children the condition resolves in about 3 days. Suppurative complications occur in about.12% of children if antibiotics are withheld. [8] Serious complications are rare in otherwise healthy children but include hearing loss, mastoiditis, meningitis, and recurrent attacks. [3] The World Health Organization estimates that, in resource-poor countries, 51, children under the age of 5 years die from complications of otitis media each year. [9] AIMS OF To reduce the severity and duration of pain and other symptoms; to prevent complications; to INTERVENTION minimise adverse effects of treatment. OUTCOMES Pain control (in infants this can be assessed by surrogate measures such as parental observation of distress/crying, and analgesic use); incidence of complications such as deafness (usually divided into short- and long-term hearing loss), recurrent attacks of AOM, mastoiditis, and meningitis; resolution of otoscopic appearances; incidence of adverse effects of treatment. METHODS BMJ Clinical Evidence search and appraisal January 27. The following databases were used to identify studies for this systematic review: Medline 1966 to January 27, Embase 198 to January 27, and The Cochrane Database of Systematic Reviews and Cochrane Central Register of BMJ Publishing Group Ltd 27. All rights reserved............................................................ 2
Controlled Clinical Trials 26, Issue. Additional searches were carried out using these websites: NHS Centre for Reviews and Dissemination (CRD) for Database of Abstracts of Reviews of Effects (DARE) and Health Technology Assessment (HTA), Turning Research into Practice (TRIP), and National Institute for Health and Clinical Excellence (NICE). Abstracts of the studies retrieved from the initial search were assessed by an information specialist. Selected studies were then sent to the author for additional assessment, using pre-determined criteria to identify relevant studies. Study design criteria for inclusion in this review were: published systematic reviews and RCTs in any language; criteria for RCTs was at least single blinded unless blinding was impossible, and containing 2 or more individuals of whom 8% or more were followed up. There was no minimum length of follow up required to include studies. We excluded all studies described as open, open label, or not blinded unless blinding was impossible. In addition, we use a regular surveillance protocol to capture harms alerts from organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA), which are added to the reviews as required. We have performed a GRADE evaluation of the quality of evidence for interventions included in this review (see table, p 13 ). QUESTION What are the effects of treatments for AOM in children? OPTION ANALGESICS................................................................. Pain Topical anaesthetic compared with placebo The effects after 1 3 minutes of topical anaesthetic drops in children taking paracetamol are unclear compared with placebo (low-quality evidence). Oral analgesics compared with placebo Oral ibuprofen or paracetamol reduce pain after 8 hours compared with placebo in children taking antibiotics (moderate-quality evidence). Benefits: Topical anaesthetics: We found one systematic review, (search date May 26, RCTs) which assessed the effect of topical analgesia for AOM [1] and identified one RCT (5 children aged 5 to 19 years) that compared topical anaesthetic drops with placebo. [11] All children were also given paracetamol. The RCT found that topical anaesthetics increased the proportion of children reporting 25% earache reduction at 3 minutes after instillation of ear drops (number of children reporting 25% pain reduction 26/27 [96%] with anaesthetic v 19/27 [7%] with placebo; P less than.2, RR 1.37, 95% CI 1.6 to 1.77). Differences in 25% pain reduction at 1 and 2 minutes after eardrop instillation or 5% difference in pain reduction were not significant (number of children reporting 25% pain reduction at 1 minutes after instillation: 13/27 [8%] with anaesthetic v 11/27 [1%] with placebo, RR 1.18, 95% CI.65 to 2.15; number of children reporting 25% pain reduction at 2 minutes after instillation: 21/27 [78%] with anaesthetic v 17/27 [63%] with placebo; RR 1.2, 95% CI.87 to 1.76; 5% pain reduction at 1 minutes after instillation: 9/27 [33%] with anaesthetic v /27 [15%] with placebo RR = 2.25, 95% CI.79 to 6.3; 5% pain reduction at 2 minutes after instillation: 13/27 [8%] with anaesthetic v 12/27 [%] with placebo, RR 1.8, 95% CI.61 to 1.93; 5% pain reduction at 3 minutes after instillation: 21/27 [78%] with anaesthetic v 15/27 [56%] with placebo; RR 1. 95% CI.95 to 2.7). Oral analgesics: We found one RCT (219 children, aged 1 6 years with otoscopically diagnosed AOM and receiving antibiotic treatment with cefaclor for 7 days) which compared the effect of treatment with ibuprofen or paracetamol three times daily versus placebo for 8 hours. [12] It found that ibuprofen significantly reduced the incidence of earache after 2 days compared with placebo as assessed by parental observation (AR 5/71 [7%] with ibuprofen v 19/75 [25%] with placebo; RR.28, 95% CI.11 to.71; NNT 5, 95% CI 3 to 15), as did paracetamol (AR 7/73 [1%] with paracetamol v 19/75 [25%] with placebo; RR.38, 95% CI.17 to.85; NNT 6, 95% CI 3 to 28; see comment below). Harms: Topical anaesthetics: The systematic review gave no information on adverse effects. [1] Oral analgesics: The RCT found that 11 children experienced mild nausea or vomiting or abdominal pain (5/71 [7%] taking ibuprofen v 3/73 [%] taking paracetamol v 3/75 [%] taking placebo; no further details reported). None were withdrawn from treatment. [12] Comment: Oral analgesics: The evidence from this RCT is limited because the assessment of the child's pain relief was based on parental observation using a scale of or 1. [12] The paracetamol versus BMJ Publishing Group Ltd 27. All rights reserved.... 3
placebo result has been recalculated by BMJ Clinical Evidence from data in the original publication, and corrects the stated conclusions of the RCT. OPTION ANTIBIOTICS VERSUS PLACEBO................................................ Pain Compared with placebo Antibiotics may reduce pain and other symptoms of AOM after 2 days compared with placebo (very low-quality evidence). Antibiotics increase the risk of vomiting, diarrhoea, or rashes compared with placebo. Benefits: We found five systematic reviews. [2] [13] [1] [15] [16] The first systematic review (search date 1992) identified four RCTs (535 children, aged months to 18 years), which compared antibiotics versus placebo in children receiving analgesics or other symptomatic relief. [13] AOM was defined as bulging or opacification of the tympanic membrane with or without erythema, accompanied by at least one of the following signs: fever, earache, irritability, otorrhoea, lethargy, anorexia, vomiting, diarrhoea, and absent or poor mobility of the tympanic membrane. It found a significant increase in the rate of resolution of symptoms with a range of antibiotics (cephalosporins, erythromycin, penicillins, co-trimoxazole) after 7 days of treatment compared with placebo (ARI 13.7%, 95% CI 8.2% to 19.2%; NNT 7, 95% CI 5 to 12). The second systematic review (search date 1997, 71 children, aged less than 2 years) identified four RCTs comparing antibiotics (penicillins, sulphonamide, amoxicillin/clavulanic acid [co-amoxiclav]) versus placebo alone or versus placebo plus myringotomy. [1] Three RCTs based diagnosis of AOM on otoscopic appearance of the tympanic membrane and clinical signs of acute infection, and one RCT based diagnosis on otoscopy findings alone. The review found no significant difference between antibiotics and placebo in symptomatic improvement within 7 days (OR 1.31, 95% CI.83 to 2.8). The third systematic review (search date 1999, 5 RCTs, 1518 children, aged weeks to 18 years) compared the effects of antibiotics (ampicillin, amoxicillin) versus placebo or observation. [2] AOM was defined as the presence of middle ear effusion in conjunction with rapid onset of one or more signs or symptoms of inflammation of the middle ear, and was categorised as uncomplicated AOM when limited to the middle-ear cleft. Clinical failure was defined as the presence of pain, fever, middle ear effusion, clinical signs of otitis media, or suppurative complications such as mastoiditis. The review found that antibiotics (ampicillin, amoxicillin) significantly reduced clinical failure rate within 2 7 days compared with placebo or observational treatment (ARR 12.3%, 95% CI 2.8% to 21.8%; NNT 8, 95% CI 5 to 36). The fourth systematic review (search date 23, 8 RCTs, 2287 children, aged 6 months to 15 years) compared early use of antibiotics (erythromycin, penicillins, sulphonamides) versus placebo. [15] AOM was defined as acute earache with at least one abnormal eardrum, otoscopic middle-ear effusion, and general signs and symptoms. Pain was assessed using parental report/score card/diary or clinician assessment at days. The review found that antibiotics significantly reduced the proportion of children still in pain 2 7 days after presentation compared with placebo (182/1169 [16%] with antibiotics v 28/1118 [22%] with placebo; ARR 7%, 95% CI 3.% to 9.8%; RR.3, 95% CI.19 to.; NNT 15, 95% CI 11 to 3). In addition, it found that significantly fewer children experienced contralateral AOM with antibiotics (35/329 [11%] with antibiotics v 56/337 [17%] with placebo; ARR 6%, 95% CI 1.% to 11%; RR.65, 95% CI.5 to.9). The review found no significant difference between groups in the rate of subsequent recurrence of AOM (187/86 [21.6%] with antibiotics v 175/8 [21.8%] with placebo; RR.99, 95% CI.83 to 1.19), abnormal tympanometry at 1 month (85/23 [36%] with antibiotics v 91/238 [38%] with placebo; RR.9, 95% CI.7 to 1.19), or abnormal tympanometry at 3 months (38/182 [21%] with antibiotics v 9/188 [27%] with placebo; RR.8, 95% CI.55 to 1.16). Four RCTs (717 children) reported pain outcomes (parental report of pain or symptom diary) 2 hours after presentation. All four RCTs found no significant difference in pain outcomes between antibiotics and placebo (RR 1.2, 95% CI.85 to 1.22). The review found only one reported case of mastoiditis in the included RCTs, which occurred in a penicillin-treated group. [15] The fifth systematic review was the only systematic review to meta-analyse at patient level (163 children ages 6 months to 12 years with AOM; 6 RCTs, 3 of which were included in the fourth systematic review) [16] compared antibiotics versus placebo ( RCTs, 1,15 children) or immediate versus delayed antibiotics (2 RCTs, 538 children) with the aim of identifying subgroups of children who would benefit more than others from antibiotics. [16] The review defined the primary outcome as an extended course of AOM, consisting of pain, fever, or both at 3 7 days. It found that, at follow-up of 3 7 days in children younger than 2 years with bilateral AOM, antibiotics significantly reduced the proportion of children who had pain, fever, or both compared with placebo (2/126 [3%] with antibiotics v 7/13 [55%] with placebo; RR.6, 95% CI.62 to.8). The differences for children under 2 years with unilateral AOM and in children over 2 years with unilateral or bilateral AOM were not significant. BMJ Publishing Group Ltd 27. All rights reserved....
Harms: Two systematic reviews gave no information on adverse events. [13] [1] The third systematic review found that adverse effects, primarily gastrointestinal, were more common in children taking cefixime than in children taking amoxicillin or ampicillin (5 RCTs, total number of participants not reported; ARI 8.%, 95% CI 3.8% to 13.1%; NNH 12, 95% CI 8 to 27), and were more common in children taking amoxicillin/clavulanate (original formulation) than in those taking azithromycin (3 RCTs, total number of participants not reported; ARI 18.%, 95% CI 8.% to 28.%; NNH 6, 95% CI to 13). [2] The fourth systematic review found that antibiotics significantly increased the risk of vomiting, diarrhoea, or rashes compared with placebo (689 children, aged 6 months to 15 years; AR 57/35 [17%] with antibiotics v 38/353 [11%] with placebo; RR 1.55, 95% CI 1.11 to 2.16; NNH 17, 95% CI 9 to 152). [15] The fifth systematic review did not directly compare adverse effects between antibiotics and placebo or between immediate and delayed antibiotics. [16] Comment: The first systematic review [13] excluded two placebo-controlled trials that were included in the fourth review [15] because they included myringotomy as part of the treatment.this may have biased the results in favour of antibiotic treatment, and may explain the higher absolute-risk reduction quoted in the first review. [13] The third systematic review commented on the difficulty in performing meta-analyses because of the varying criteria between studies for defining AOM and outcome measures. [2] The results of systematic reviews that compare antibiotics versus placebo may vary because of differences in entry criteria and outcome measures. Clinical guide: One quasi-randomised trial from Sweden conducted in 195 comparing the effects of antibiotics versus placebo found no cases of mastoiditis in the penicillin-treated group, whereas 17% of the control group developed mastoiditis. [17] Therefore, in populations in which the incidence of complicating mastoiditis is high, antibiotic treatment would be advised. OPTION CHOICE OF ANTIBIOTIC REGIMEN............................................... Different antibiotics compared with each other Different antibiotics may be as effective as each other at reducing signs and symptoms of AOM after 7 days (very low-quality evidence). Antibiotics increase the risk of vomiting, diarrhoea, and rashes compared with placebo, but rates may vary between different types of antibiotics. Benefits: We found two systematic reviews. [2] [13] One systematic review (search date 1992, 33 RCTs, 5 children, aged months to 18 years) compared a range of antibiotics (cephalosporins, erythromycin, penicillins, trimethoprim sulfamethoxazole [co-trimoxazole]). [13] AOM was defined as bulging or opacification of the tympanic membrane with or without erythema, accompanied by at least one sign (fever, earache, irritability, otorrhoea, lethargy, anorexia, vomiting, diarrhoea, poor or absent mobility of the tympanic membrane). Treatment success was defined as absence of all presenting signs and symptoms of AOM at the evaluation point closest to 77 days after start of treatment. The systematic review found no significant differences between different antibiotics in rate of treatment success at 7 days, or of middle-ear effusion at 3 days. The second systematic review found no significant difference between penicillin and ampicillin or amoxicillin in clinical failure rates within 7 days (search date 1999, 3 RCTs, 91 children, aged weeks to 18 years; clinical failure rate difference: +.5%, 95% CI.8% to +1.7%). [2] The review also found no significant difference in clinical failure rates within 3 7 days between cefaclor and ampicillin or amoxicillin ( RCTs, 56 children, aged weeks to 18 years; clinical failure rate difference 5.%, 95% CI 5.2% to +.%). Clinical failure was defined as the presence of pain, fever, middle ear effusion, clinical signs of otitis media, or suppurative complications such as mastoiditis. Harms: See harms of antibiotics, p. Comment: Many RCTs have studied a variety of antibiotic regimens for the treatment of otitis media, but there is heterogeneity in participants, treatment regimens, controls, and outcome measures. A metaanalysis is needed to provide a valid summary that would be useful to clinicians. OPTION IMMEDIATE VERSUS DELAYED ANTIBIOTIC TREATMENT............................ BMJ Publishing Group Ltd 27. All rights reserved.... 5
Immediate antibiotics compared with delayed antibiotics Immediate antibiotics may reduce pain and other symptoms of AOM compared with delayed antibiotics or watchful waiting after 2 hours, but not after 7 days (moderatequality evidence). Immediate antibiotics are associated with higher rates of diarrhoea and rashes compared with a delayed antibiotic strategy, and may increase the development of bacterial multidrug resistance. Benefits: We found no systematic review. We found three RCTs, one of which was reported in two articles. [6] [18] [19] [2] The first RCT compared immediate versus delayed antibiotic (amoxicillin or erythromycin) use. [18] AOM was defined as acute earache and otoscopic evidence of acute inflammation of the eardrum, such as dullness or cloudiness with erythema, bulging, or perforation. Immediate antibiotic treatment was defined as a prescription given to parents at the initial consultation. Delayed antibiotic treatment was defined as follows: parents were asked to wait 72 hours after seeing the doctor before using the prescription, and to use the prescription only if the child still had substantial earache or fever, or was not starting to get better. Earache was assessed from daily diary of symptoms and perceived severity of pain scores (1 = no pain to 1 = extreme pain). The RCT found that, after the first 2 hours of illness, immediate antibiotic use significantly reduced the duration of earache (mean difference.1 days, 95% CI.5 days to.8 days), duration of ear discharge (mean difference.66 days, 95% CI.19 days to.13 days), number of disturbed nights (mean difference.72 days, 95% CI.3 days to.13 days), number of days crying (mean difference.69 days, 95% CI.31 days to.8 days), and the number of teaspoons of paracetamol used (mean difference.52 teaspoons daily, 95% CI.26 to.79 teaspoons daily) compared with delayed antibiotic use. The RCT found no significant difference between immediate and delayed antibotics in mean daily pain score (mean difference.16, 95% CI.2 to +.11), number of daily episodes of distress (mean difference.12, 95% CI.3 to +.11), or days of absence from school (mean difference.18 days, 95% CI.76 days to +.1 days). A follow-up study in the same group of children found no significant difference between immediate and delayed antibiotics in episodes of earache or poor scores on function scales at 3 months and at 1 year (earache at 3 months: OR.89 95% CI.8 to 1.65; earache at 1 year: OR 1.3, 95% CI.6 to 1.78; function scale scores at 3 months OR 1.16%, 95% CI.61 to 2.22 ; 1 year OR 1.12, 95% CI.57 to 2.19; absolute numbers not reported). [19] The second RCT (223 children aged 6 months to 12 years) compared immediate antibiotics versus watchful waiting. [2] Children with symptoms of ear infection and otoscopic evidence of AOM including middle ear effusion who had non-severe infection (as judged by a severity score devised for the trial) received immediate antibiotics (amoxicillin) plus analgesics or analgesics alone. At 12 and 3 days' follow-up, parent satisfaction was similar for both groups (parent satisfaction score at 12 days' follow-up. with immediate antibiotics v. with watchful waiting; at 3 days' follow-up.6 for both groups; reported as not significant). Immediate antibiotics significantly reduced the number of analgesic doses received by children, reduced treatment failures, and resolved symptoms more quickly during the first 1 days compared with watchful waiting (analgesic doses 3. with immediate antibiotics v 7.7 with watchful waiting; P less than.1; treatment failure for children under 2 years /65 [5%] v 12/5 [2%], P less than.1; treatment failure for children over 2 years 1/ [2%] v 9/5 [18%], P less than.1; symptoms scores P =., absolute numbers not reported, results presented graphically). Prescriptions for antibiotics were reduced by 73% in the watchful waiting group compared to the immediate antibiotic group. Of the children in the watchful waiting group, 73/111 [ 66%] completed the study without needing antibiotics. Immediate antibiotic treatment resulted in eradication of S pneumoniae carriage in most children, but S pneumoniae strains cultured from children in the antibiotic group at day 12 were more likely to be multidrug-resistant than strains from children in the watchful waiting group.the third RCT (283 children aged 6 months to 12 years) compared standard prescription (SP) to a wait and see prescription (WASP) for children diagnosed with AOM. [6] The diagnosis was made at the discretion of the clinician, using evidence-based diagnostic criteria. All children received a written prescription for an antibiotic, chosen and dosed by the clinician. Those randomised to the WASP group were given written and verbal instructions not to fill the antibiotic prescription unless your child is not better or worse in 8 hours (2 days) after today s visit, while the SP group, who were was advised to fill the antibiotic prescription and give the antibiotic to your child after today s visit. All children received ibuprofen suspension and otic analgesic drops. The outcome measures were the filling of the antibiotic prescription and the clinical course of the AOM. At to 6 days follow-up, the RCT found that more parents in the WASP group did not fill the prescription compared with the SP group (number of parents not filling the antibiotic prescription 82/132 [62%] v 17/133 [13%]; P less than.1). Within the WASP group there was a significant association for filling the prescription because of fever and earache (number of children with fever 27/5 [5%] who filled the prescription v 15/82 [18%] who did not fill the prescription, RR 2.95, 95% CI 1.75 to.99, P less than.1; earache 2/ [8%] v 3/82 [52%]; RR 1.62, 95% CI 1.26 to 2.3, P less than.1). At 11 to 1 days follow-up there was no significant difference BMJ Publishing Group Ltd 27. All rights reserved.... 6
Harms: Comment: in the frequency of subsequent fever, earache, or unscheduled visits for medical care (number of children with fever /12 [32%] with WASP v 38/133 [31%], P =.37; earache 83/12 [67%] with WASP v 75/123 [61%] with SP, P =.7; unscheduled visits for medical care 18/12 [15%] with WASP v 1/123 [11%], P =.51). The first RCT [18] found that immediate treatment significantly increased diarrhoea compared with delayed treatment (AR 25/135 [19%] with immediate v 1/15 [9%] with delayed; RR 1.9, 95% CI 1.8 to 3.66; NNH 11, 95% CI 5 to 125), but it had no significant effect on rash (AR 6/133 [5%] with immediate v 8/19 [5%] with delayed; RR.8, 95% CI.3 to 2.36). The second RCT reported more antibiotic-related adverse effects in the antibiotic group compared with the watchful waiting group (no further details on adverse effects or absolute numbers reported). [2] In the third RCT, at to 6 days follow-up, diarrhoea was reported significantly more in the SP group compared with the WASP group (number of children with diarrhoea 1/132 [8%] with WASP v 31/133 [23%] with SP; P less than.1). [6] There was no significant difference of vomiting between groups (15/132 [11%] with WASP v 15/133 [11%] with SP, P =.56). None. Otitis media in children (acute) OPTION LONGER VERSUS SHORTER COURSES OF ANTIBIOTICS............................ Longer courses of antibiotics compared with shorter courses Longer (8 day) courses of antibiotics reduce symptoms and prevent relapse or re-infection at 89 days compared with 5-day courses, but are no more effective than shorter courses after 2 2 days (high-quality evidence). Benefits: We found one systematic review [21] and two subsequent RCTs. [22] [23] The systematic review (search date 1998, 3 RCTs, 8215 children, aged weeks to 18 years with AOM) found that, at early evaluation (89 days), longer courses (8 days) of antibiotics reduced treatment failure (lack of clinical resolution, relapse, or re-infection) compared with shorter courses (5 days) of antibiotics (5 RCTs, 152 children; treatment failure at 89 days: AR 112/76 [15%] with longer courses v AR 165/778 [22%] with shorter courses; RR.71, 95% CI.58 to.89). [21] However, by 2 3 days there were no significant differences between treatment groups (9 RCTS, 2115 children; treatment failure at 2 3 days: AR 211/131 [21%] with shorter courses v 199/18 [19%] with longer courses; RR 1.16, 95% CI.98 to 1.37). The first subsequent RCT (385 younger children with newly-diagnosed AOM, mean age 13.3 months, age range. 3. months) compared amoxicillin/clavulanate for 1 days (3 times daily) versus 5 days (3 times daily) followed by 5 days of placebo. [22] Clinical success or failure was assessed at 12 days and again at 28 2 days after starting treatment. Intention-to-treat analysis found that the 1-day regimen significantly increased clinical success on days 12 compared with the 5-day regimen (AR 158/186 [85%] for 1 days v 11/192 [73%] for 5 days; RR 1.16, 95% CI 1. to 1.28; NNT 8, 95% CI 5 to 3). However, by days 28 2 there was no significant difference in clinical success between the two groups (AR 18/185 [58%] for 1 days v 12/19 [5%] for 5 days; RR 1.9, 95% CI.91 to 1.3). The second subsequent RCT compared cefpodoxime/proxetil twice daily at 8 mg/kg daily for 1 days versus cefpodoxime/proxetil for 5 days followed by 5 days of placebo. It found significantly higher success rates in the 1-day than in the 5-day treatment group after 12 days (AR 199/222 [9%] for 1 days v 18/226 [8%] for 5 days; RR 1.13, 95% CI 1. to 1.22; NNT 1, 95% CI 6 to 3), but found no significant difference after 28 2 days (AR 19/222 [67%] for 1 days v 11/226 [62%] for 5 days; RR 1.8, 95% CI.9 to 1.23). [23] Harms: Comment: The systematic review and the two subsequent RCTs found similar rates of diarrhoea and/or [21] [22] [23] vomiting and rash between longer and shorter courses of antibiotics. None. OPTION MYRINGOTOMY............................................................... Compared with no myringotomy Myringotomy may be no more effective than no myringotomy at reducing the symptoms of AOM after 1 7 days (low-quality evidence). Compared with antibiotics Myringotomy may be less effective at reducing symptoms of AOM compared with antibiotics after 12 hours to 11 days (low-quality evidence). BMJ Publishing Group Ltd 27. All rights reserved.... 7
Myringotomy may be less likely than antibiotics to cause diarrhoea. Otitis media in children (acute) Benefits: We found no systematic review but found three RCTs. [2] [25] [26] The first RCT (15 infants, aged 3 months to 1 year with AOM) compared three treatments: antibiotic only (amoxicillin/clavulanic acid [co-amoxiclav]), myringotomy plus placebo, and myringotomy plus antibiotic (co-amoxiclav). [2] AOM was defined as the presence of middle-ear effusion and bulging (with or without redness of the tympanic membrane) associated with recent irritability or fever. The RCT found that those in the myringotomy-plus-placebo group had higher rates of persistent ear infection at 91 days compared with those receiving antibiotic alone (21/3 [7%] with myringotomy plus placebo v 2/3 [7%] with antibiotic alone; NNH 3, 95% CI 2 to 12) and lower rates of otoscopic recovery (7/32 [23%] with myringotomy plus placebo v 18/32 [6%] with antibiotic alone; NNH 3, 95% CI 2 to 5). The second RCT (171 children, aged 22 years with AOM) compared no treatment, myringotomy only, amoxicillin (25 mg 3 times daily for 7 days) only, and amoxicillin plus myringotomy. [25] Diffuse redness or bulging of the eardrum, or both, was taken as decisive. The RCT found no significant difference in pain between myringotomy only, amoxicillin only, and no treatment at 2 hours or after 7 days (pain at 2 hours: 26/36 [72%] with myringotomy v 3/7 [72%] with amoxicillin v 29/ [72%] with no treatment; pain after 7 days: 31/35 [89%] with myringotomy v 3/6 [93%] with amoxicillin v 3/38 [9%] with no treatment; no further details reported).the third RCT (536 infants and children, aged between 7 months and 12 years with severe AOM or recurrent AOM) compared treatment with amoxicillin ( mg/kg per day in 3 divided doses for 1 days) only, amoxicillin plus myringotomy, or myringotomy plus placebo. [26] AOM was diagnosed on the basis of fever, earache, or irritability with redness and/or bulging of the eardrum. An episode of AOM was classified as severe or non-severe according to the child's temperature and an earache score. The RCT found significantly higher rates of initial treatment failure (no resolution of symptoms within 12 hours) for severe episodes of AOM in children aged 22 years treated with myringotomy plus placebo compared with amoxicillin only (23% with placebo plus myringotomy v % with amoxicillin only; P =.6). Harms: The first RCT reported that, in those children who had no diarrhoea at the start of the study, 7/6 (12%) receiving antibiotic (co-amoxiclav) had three or more loose or watery bowel movements a day, as compared with /3 (%) children in the myringotomy plus placebo group (P =.5). [2] The second RCT reported similar rates of adverse effects between the two groups. [25] The third RCT gave no information on adverse effects. [26] Comment: Two RCTs provided results in the form of children or as individual ears as the unit measured. [2] [25] Because randomisation was based on children, the figures reported here exclude those results based on individual ears. QUESTION What are the effects of interventions to prevent recurrence of AOM in children? OPTION ANTIBIOTIC PROPHYLAXIS (LONG TERM)......................................... Incidence of AOM Compared with placebo Prophylactic antibiotics reduce the incidence of AOM compared with placebo or no treatment in children at risk of otitis media (high-quality evidence). Antibiotics are associated with adverse effects and the development of bacterial drug resistance. Benefits: Harms: We found one systematic review (search date 26, 16 RCTs, 183 children) that compared antibiotics versus placebo or no treatment for the prevention of AOM, AOM with perforation, or chronic suppurative otitis media. [7] All RCTs enrolled children at increased risk of AOM, and in seven RCTs the children were prone to otitis media. The review found that long-term antibiotics significantly reduced the proportion of children with any episode of AOM or chronic suppurative otitis media and the number of episodes of AOM or chronic suppurative otitis media (number of children with otitis media 253/78 [3%] with antibiotic v 331/61[5%] with control; RR =.62, 95% CI.52 to.75, number of episodes of otitis media 36 with antibiotics v 752 with control; incidence rate ratio (IRR).8, 95% CI.37 to.62). Antibiotics prevented 1.5 episodes of AOM for every 12 months of treatment per child (NNT 5, 95% CI to 6). The review found that antibiotics were not associated with a significant increase in adverse effects (adverse effects 1/5 [2%] with antibiotics v 3/39 [1%]; RR 1.99, 95% CI.25 to 15.89). [7] BMJ Publishing Group Ltd 27. All rights reserved.... 8
Comment: Otitis media in children (acute) Clinical guide: We found insufficient evidence on which antibiotic to use and for how long, and how many episodes of AOM to justify starting preventive treatment. OPTION VACCINATION (PNEUMOCOCCAL)............................................... Incidence of AOM Compared with placebo or control vaccine Pneumococcal vaccine (8 to 1 valent, or 7 to 9 valent) may be no more effective than placebo at reducing the incidence of AOM in children aged from 2 months to 7 years (very low-quality evidence). The adverse effects associated with pneumococcal vaccination are unclear. Note We found no clinically important results about the effects of the 23 valent pneumococcal vaccine that is currently available. Benefits: We found one systematic review (search date 23). [27] It included RCTs in children up to 12 years of age, with a follow-up of at least 6 months after vaccination. The review identified 12 RCTs (6,57 children aged from 2 months to 7 years) comparing pneumococcal vaccination versus placebo or a control vaccination. Eight RCTs (628 children) used 8 to 1 valent pneumococcal polysaccharide vaccine (PPV), three RCTs (39,79 children) used 7 to 9 valent pneumococcal conjugate vaccine (PCV), and one RCT (383 children) used both PPV and PCV. No RCTs directly examined the effects of the currently available 23 valent PPV (see comment below). Five RCTs randomised children with previous or current AOM, six RCTs studied healthy children, and one RCT included both healthy children as well as those with a history of AOM. The studies varied in the length of follow up-reported, and because of this the review stated that risk ratios would not provide a valid estimate of effect size. The review therefore calculated rate ratios (see comment below). When assessing effects in all children, the review found no significant difference in the number of AOM episodes per person month between PPV and control (overall rate ratio:.9, 95% CI.81 to 1.). [27] This analysis excluded 1 RCT that had poor methods. When analysing healthy children aged under 2 months or over 2 months separately, the review also found no significant difference in the number of AOM episodes per person month between PPV and placebo or control (children under 2 months: rate ratio.9, 95% CI.81 to 1.2, children aged over 2 months: rate ratio.7, 95% CI.7 to 1.17). [27] In two RCTs assessing children deemed 'at increased risk for AOM becuase they were attending day care or had previous AOM, the review also found no significant difference between PCV and control in episodes of AOM (rate ratio 1.9, 95% CI.69 to 1.71). [27] These results should be interpreted with caution as the review reported a variety of analyses on various subgroups of age, previous disease status, and study, some of which were of borderline significance or not significant depending on which subgroups were analysed and which statistical adjustments were made. Harms: The review gave no information on adverse effects. [27] Comment: The review defined the rate ratio as follows: (total AOM episodes in pneumococcal vaccination group/number of children in pneumococcal vaccination group follow up time in months)/ (total AOM episodes in control group/number of children in control group follow up time in months). [27] It should be noted that a 23 valent PPV is currently available, rather than the 8 valent PPV used in the RCTs in the review. The review found no RCTs comparing the 23 valent PPV versus control in AOM. The review noted that trials evaluating the impact of PCV on AOM in specific highrisk groups are ongoing around the world, and that their results may alter advice about the use of vaccines for prevention of AOM. [27] Clinical guide: The review concluded that, based on currently available results, a large-scale use of pneumococcal vaccination for this specific indication is not yet recommended. [27] It suggested that, before widespread use of pneumococcal vaccines could be recommended, quantification of the stress of vaccination to children, and analysis of the cost effectiveness of implementation, would be needed. It also noted that pneumococcal vaccination might lead to pneumococcal shift and that, therefore, any overall reduction in clinical episodes of AOM may not be permanent. BMJ Publishing Group Ltd 27. All rights reserved.... 9
OPTION TYMPANOSTOMY (VENTILATION TUBES)......................................... Incidence of AOM Compared with no surgery or myringotomy alone Tympanostomy plus insertion of drainage tubes may reduce the incidence of AOM after 6 months, but not after 18 months, compared with no surgery or with myringoplasty alone (low-quality evidence). Tympanostomy plus drainage tubes may increase the risk of tympanosclerosis and hearing impairment. Benefits: We found no systematic review but found one RCT. [28] The RCT ( children, aged 9 months to 7 years with bilateral recurrent AOM of equal severity in each ear despite more than 3 months of antibiotic prophylaxis) compared tympanostomy with ventilation tube insertion (tympanostomy tube insertion) into a randomly selected ear versus the contralateral ear receiving either no surgery or myringotomy alone. Recurrent AOM was defined as the recurrent presence (more than episodes) of earache with red and bulging tympanic membranes. The RCT found that tympanostomy tube insertion significantly reduced the mean number of episodes of AOM during the first 6 months after treatment compared with myringotomy alone or no surgery (actual mean difference:.6% with tympanostomy v 1.8% with control; difference in mean number of episodes:.2, 95% CI 2.2 to.9). However, the RCT found that tympanostomy tube insertion did not significantly reduce the mean number of episodes of AOM during the subsequent 18 months (actual mean difference:.8% with tympanostomy v.8% with control; difference in mean number of episodes %, 95% CI.3 to +.3). Harms: Comment: The RCT reported a non-significant trend (P =.3) towards more recurrent infections and worse hearing in ears that had received tympanostomy tubes, which became apparent after tube extrusion. [28] Anatomical abnormalities (tympanosclerosis, atrophy, or retraction and chronic perforation), although not thought to be clinically significant, were more common in the ears receiving tympanostomy tubes. There was significantly more tympanosclerosis in ears that received tympanostomy tubes than in those that received myringotomy alone (35/61 [57.%] with tympanostomy tubes v 5/26 [19.2%] with myringotomy alone; P =.) or no surgery (35/61 [57%] with tympanostomy tubes v 2/27 [7%] with no surgery; P less than or equal to.1). At the 2 year evaluation the hearing was poorer in ears with anatomical abnormalities. The RCT included some children with otitis media with effusion, although the results concerning benefits presented here refer only to those children in the study with recurrent AOM. It was not possible from the data available to differentiate the evidence on harms into children with recurrent AOM compared with otitis media with effusion. Medical treatment and antibiotic prophylaxis were allowed whenever indicated. It was not possible from the data presented to tell whether the different groups differed in the amount of medical treatment and prophylactic antibiotics. GLOSSARY Mastoiditis The presence of infection in mastoid cavity. Myringotomy The surgical creation of a perforation in tympanic membrane. Tympanostomy The surgical creation of a perforation in tympanic membrane for the purpose of inserting a ventilation tube. High-quality evidence Further research is very unlikely to change our confidence in the estimate of effect Low-quality evidence Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Moderate-quality evidence Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Very low-quality evidence Any estimate of effect is very uncertain. SUBSTANTIVE CHANGES Analgesics for AOM in children One systematic review added, which found that topical anaesthetics increased the proportion of children reporting 25% earache reduction 3 minutes after ear-drop instillation. [1] Categorisation: Likely to be beneficial Antibiotic prophylaxis for preventing recurrence of AOM in children One systematic review found that longterm antibiotics reduced otitis media episodes. [7] Categorisation unchanged: Trade-off between benefits and harms Antibiotics versus placebo for AOM in children One systematic review added, which found that antibiotics were most beneficial in children under 2 years of age with bilateral AOM compared with older children or children with unilateral AOM. [16] Categorisation: Trade-off between benefits and harms. BMJ Publishing Group Ltd 27. All rights reserved.... 1
Immediate or delayed antibiotics for AOM in children Two RCTs added which found that immediate antibiotic use reduced symptoms faster, and that the reasons for immediate antibiotic use were fever and earache. Categorisation: Trade-off between benefits and harms REFERENCES 1. Rothman R, Owens T, Simel DL. Does this child have acute otitis media? JAMA 23;29:16336.[PubMed] 2. Marcy M, Takata G, Shekelle P, et al. Management of Acute Otitis Media. Evidence Report/Technology Assessment No. 15. (Prepared by the Southern California Evidence Based Practice Centre under contract No. 29-97-1.) AHRQ Publication No. 1-E1. Rockville, MD: Agency for Healthcare Research and Quality, May 21. Search date 1999; primary sources Medline, Cochrane Library, Health STAR, International Pharmaceutical Abstracts, CINAL, BIOSS, and EM- BASE. 3. Froom J, Culpepper L, Jacobs M, et al. Antimicrobials for acute otitis media? A review from the International Primary Care Network. BMJ 1997;315:982.[PubMed]. Del Mar C, Glasziou P, Hayem M. Are antibiotics indicated as initial treatment for children with acute otitis media? A meta-analysis. BMJ 1997;31:1526529. Search date 199; primary sources Medline and Current Contents.[PubMed] 5. Berman S. Otitis media in developing countries. Pediatrics 1995;96:12631.[PubMed] 6. Spiro DM, Tay KY, Arnold DH, et al. Wait-and-see prescription for the treatment of acute otitis media: a randomized controlled trial. JAMA 26;296:123521. 7. Leach AJ, Morris PS. Antibiotics for the prevention of acute and chronic suppurative otitis media in children. In:The Cochrane Library, Issue 2, 26. Chichester, UK: John Wiley & Sons, Ltd. 8. Rosenfeld RM. Natural history of untreated otitis media. Laryngoscope 23;113:165657.[PubMed] 9. World Health Organization. World Development Report 1993: Investing in Health. Oxford: Oxford University Press, 1993:215 222. 1. Foxlee R, Johansson A, Wejfalk J, et al. Topical analgesia for acute otitis media. In: The Cochrane Library, Issue 2, 26. Chichester, UK: John Wiley & Sons, Ltd. 11. Hoberman A, Paradise JL, Reynolds EA, et al. Efficacy of Auralgan for treating ear pain in children with acute otitis media.[see comment]. Arch Pediatr Adolescent Medicine 1997;151:675 678. 12. Bertin L, Pons G, d'athis P, et al. A randomized double blind multicentre controlled trial of ibuprofen versus acetaminophen and placebo for symptoms of acute otitis media in children. Fundam Clin Pharmacol 1996;1:387 392.[PubMed] 13. Rosenfeld RM, Vertrees JE, Carr J, et al. Clinical efficacy of antimicrobial drugs for acute otitis media: meta-analysis of 5 children from thirty-three randomised trials. J Pediatr 199;12:355 367. Search date 1992; primary sources Medline and Current Contents.[PubMed] 1. Damoiseaux RA, van Balen FAM, Hoes AW, et al. Antibiotic treatment of acute otitis media in children under two years of age: evidence based? Br J Gen Pract 1998;8:186186. Search date 1997; primary sources Medline, Embase, and hand searched references.[pubmed] 15. Glasziou PP, Del Mar CB, Sanders SL, Hayem M. Antibiotics for acute otitis media in children. In: The Cochrane Library, Issue 2, 26. Chichester, UK: John Wiley & Sons, Ltd. Search date 23; primary sources Medline, Current Contents, reference lists, Cochrane Controlled Trials Register, and Index Medicus. Otitis media in children (acute) 16. Rovers MM, Glasziou P, Appelman CL, et al. Antibiotics for acute otitis media: a meta-analysis with individual patient data. Lancet 26/1/21;368:12935. 17. Rudberg, RD. Acute otitis media: comparative therapeutic results of sulphonamide and penicillin administered in various forms. Acta Otolaryngol 195;113(suppl):1 79. 18. Little P, Gould C, Williamson I, et al. Pragmatic randomised controlled trial of two prescribing strategies for childhood acute otitis media. BMJ 21;322:336 32.[PubMed] 19. Little P, Moore M, Warner G, et al. Longer term outcomes from a randomised trial of prescribing strategies in otitis media.[see comment]. Br J Gen Pract 26;56:17682. 2. McCormick DP, Chonmaitree T, Pittman C, et al. Nonsevere acute otitis media: A clinical trial comparing outcomes of watchful waiting versus immediate antibiotic treatment. Pediatrics 25;115:15565. 21. Kozyrskyj AL, Hildes-Ripstein GE, Longstaffe SEA, et al. Short course antibiotics for acute otitis media. In: The Cochrane Library, Issue 2, 26. Chichester, UK: John Wiley & Sons, Ltd. Search date 1998; primary sources Medline, Embase, Science Citation Index, Current Contents, hand searches of reference lists, and personal contacts. 22. Cohen R, Levy C, Boucherat M, et al. A multicenter randomized, double blind trial of 5 versus 1 days of antibiotic therapy for acute otitis media in young children. J Pediatr 1998;133:63 639.[PubMed] 23. Cohen R, Levy C, Boucherat M, et al. Five vs. ten days of antibiotic therapy for acute otitis media in young children. Pediatr Infect Dis J 2;19:58 63.[PubMed] 2. Engelhard D, Cohen D, Strauss N, et al. Randomised study of myringotomy, amoxicillin/clavulanate, or both for acute otitis media in infants. Lancet 1989;2:113.[PubMed] 25. van Buchem FL, Dunk JH, van t Hof MA. Therapy of acute otitis media: myringotomy, antibiotics, or neither? A double blind study in children. Lancet 1981;318:883 887. 26. Kaleida PH, Casselbrant ML, Rockette HE, et al. Amoxicillin or myringotomy or both for acute otitis media: results of a randomised clinical trial. Pediatrics 1991;87:66 7.[PubMed] 27. Straetemans M, Sanders EAM, Veenhoven RH, et al. Pneumococcal vaccines for preventing otitis media. In: The Cochrane Library, Issue 2, 26. Chichester, UK: John Wiley & Sons, Ltd. Search date 23. Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 2, 23), Cochrane Acute Respiratory Infection Group specialised register (3th June 23), MEDLINE (January 1966 to June 23), EMBASE, (January 199 to June 23) and reference lists of all studies and review articles retrieved. Authors also contacted two vaccine manufacturers and first or corresponding authors of some of the included studies. 28. Le CT, Freeman DW, Fireman BH. Evaluation of ventilating tubes and myringotomy in the treatment of recurrent or persistent otitis media. Pediatr Infect Dis J 1991;1:21.[PubMed] Paddy O'Neill General Practitioner Norton Medical Centre Stockton on Tees UK Tony Roberts Clinical Effectiveness Advisor and Honorary Research Fellow South Tees Hospitals Trust North Tees Primary Care Trust, and The University of Durham Newcastle UK Clare Bradley Stevenson Paediatric Registrar Freeman Hospital Newcastle UK Competing interests: PON, TR, and CBS declare that they have no competing interests. BMJ Publishing Group Ltd 27. All rights reserved.... 11