Treatment of Mycoplasma Pneumonia: a Systematic Review

REVIEW ARTICLE

Treatment of Mycoplasma Pneumonia: A Systematic Review

AUTHORS: Eric Biondi, MD,a Russell McCulloh, MD,b Brian Alverson, MD,c Andrew Klein, BS,a Angela Dixon, BSN, MLS, abstract a d AHIP, and Shawn Ralston, MD BACKGROUND AND OBJECTIVE: Children with community-acquired aDepartment of Pediatrics, University of Rochester, Rochester, lower respiratory tract infection (CA-LRTI) commonly receive antibiotics New York; bDepartment of Pediatrics, Children’s Mercy Hospitals & Clinics, Kansas City, Missouri; cDepartment of Pediatrics, for Mycoplasma pneumoniae. The objective was to evaluate the effect Hasbro Children’s Hospital, Providence, Rhode Island; and of treating M. pneumoniae in children with CA-LRTI. dDepartment of Pediatrics, Children’s Hospital at Dartmouth– Hitchcock, Hanover, New Hampshire METHODS: PubMed, Cochrane Central Register of Controlled Trials, and bibliography review. A search was conducted by using Medical Subject KEY WORDS pediatric, pneumonia, macrolide, azithromycin, atypical Headings terms related to CA-LRTI and M. pneumoniae and was not pneumonia, mycoplasma restricted by language. Eligible studies included randomized controlled ABBREVIATIONS trials (RCTs) and observational studies of children #17 years old with CA-LRTI—community-acquired lower respiratory tract infection confirmed M. pneumoniae and a diagnosis of CA-LRTI; each must have CAP—community-acquired pneumonia CI—95% confidence interval also compared treatment regimens with and without spectrum of IDSA—Infectious Diseases Society of America activity against M. pneumoniae. Data extraction and quality assessment RCT—randomized controlled trial were completed independently by multiple reviewers before arriving — URTI upper respiratory tract infection at a consensus. Data were pooled using a random effects model. Dr Biondi conceptualized and designed the review, reviewed articles, ran the meta-analysis, and drafted the original RESULTS: Sixteen articles detailing 17 studies were included. The most manuscript; Dr McCulloh conceptualized and designed the commonly selected primary outcome was symptomatic improvement. review, reviewed articles, helped to run the meta-analysis, and Nine studies examined M. pneumoniae treatment in CA-LRTI secondary drafted the original manuscript; Dr Alverson participated in the design, helped interpret the included studies, and critically to M. pneumoniae, and 5 RCTs met criteria for meta-analysis. The reviewed the manuscript; Mr Klein participated in study design, suggested pooled risk difference of 0.12 (95% confidence interval, tabulated articles, and reviewed the manuscript; Ms Dixon 20.04 to 0.20) favoring treatment was not significantly different participated in study design, helped perform the literature and demonstrated significant heterogeneity. Limitations included review, and critically reviewed portions of the manuscript; Dr Ralston participated in study design, supervised Drs Biondi substantial bias and subjective outcomes within the individual and McCulloh in study interpretation and translated 1 study, and studies, difficulty interpreting testing modalities, and the inability to participated in the meta-analysis; and all authors approved the correct for mixed infections or timing of intervention. final manuscript as submitted. fi fi www.pediatrics.org/cgi/doi/10.1542/peds.2013-3729 CONCLUSIONS: We identi ed insuf cient evidence to support or refute treatment of M. pneumoniae in CA-LRTI. These data highlight the need doi:10.1542/peds.2013-3729 for well-designed, prospective RCTs assessing the effect of treating M. Accepted for publication Feb 21, 2014 pneumoniae in CA-LRTI. Pediatrics 2014;133:1081–1090 Address correspondence to Russell McCulloh, MD, Pediatric Infectious Diseases, Children’s Mercy Hospitals & Clinics, 2401 Gillham Rd, Kansas City, MO 64108. E-mail: [email protected] PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). Copyright © 2014 by the American Academy of Pediatrics FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose. FUNDING: No external funding. POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no conflicts of interest to disclose. COMPANION PAPER: A companion to this article can be found on page 1124, and online at www.pediatrics.org/cgi/doi/10.1542/peds. 2014-0871.

PEDIATRICS Volume 133, Number 6, June 2014 1081 Downloaded from www.aappublications.org/news by guest on October 2, 2021 Community-acquired pneumonia (CAP) bination therapy decreased length of ,18 years of age with CA-LRTI. The re- accounts for .150 000 pediatric hos- stay by 20% to 30% over b-lactam view was conducted in concordance pitalizations each year in the United monotherapy.2,4 Data on antibiotic treat- with the Preferred Reporting Items for States, and there are great disparities ment of pediatric CA-LRTI caused by Systematic Reviews and Meta-Analyses in patterns of care and outcomes.1 M. pneumoniae are generally consid- statement. Specifically, wide variations in antibi- ered inconclusive,3,4,10,14 yet macro- otic prescribing practices exist among lides remain the most commonly Outcome Types physicians, often because the causa- overprescribed antibiotic at pediatric The primary outcome was clinical im- tive agent is not identified.2 For this clinics in the United States: .6 million provement or cure at follow-up. Clinical reason, recent evidence-based pediat- annual doses for respiratory symp- improvement or cure could include reso- ric CAP practice guidelines published toms without a clear indication.15 lution of fever; resolution or improvement by the Infectious Diseases Society of A Cochrane review on the use of anti- in symptoms such as cough, congestion, America (IDSA) recommend that chil- biotics to treat CA-LRTI secondary to M. shortness of breath, fatigue, or chest pain; dren hospitalized with CAP be tested pneumoniae in children found in- or improvement or cure as defined by the for Mycoplasma pneumoniae.3 M. sufficient evidence to draw any specific authors of the individual studies. pneumoniae is a common cause of CAP conclusions about the efficacy of anti- and other community-acquired lower biotics for the condition.10 However, Literature Search respiratory tract infections (CA-LRTIs), this review omitted at least 3 RCTs With the assistance of a librarian (A.D.), particularly in school-age children and (roughly one-third of the total RCTs) for we performed a comprehensive search adolescents, but there are large gaps which there appeared to be applicable of PubMed (January 1966–August 2012) in our understanding of this disease.3,4 data on the topic,12,16,17 and data from with the Medical Subject Headings Prevalence estimates vary from 10% to observational studies were omitted. terms and keywords (Table 1). After 40% in pediatric CA-LRTI,5–9 and few The objective of our study was to pro- duplicates were removed, we used in- studies address treatment recommenda- vide a more comprehensive review of clusion and exclusion criteria to select tions.4,10 Additionally, the IDSA guidelines all available published literature on the studies based on their title and ab- use of antibiotics in children to treat target the use of macrolides in CA-LRTI as stract to include in our review. Addi- 3 CA-LRTI secondary to M. pneumoniae. an area needing additional research. tional studies were identified through First-line treatment of children hospi- manual search of the bibliographies of METHODS talized with CAP currently includes a qualifying studies. Before manuscript b-lactam to treat common causative Search Design submission, a second search was per- bacterial agents such as Streptococ- This was a systematic review of all formed (August 2012–September 2013) cus pneumoniae and a macrolide to observational and randomized trials to ensure that the review was as up-to- provide additionally treatment of atypi- comparing antibiotics with spectrum of date as possible and any new studies 3 cal pathogens such as M. pneumoniae. activity for M. pneumoniae (eg, mac- were included in the analysis. – Studies of antibiotic use in the pre rolide, tetracycline, or quinolone class) IDSA guidelines era showed marked with placebo or antibiotics from any Study Selection variability in prescription practices,2 other class without spectrum of activ- Studies were considered eligible for probably because the efficacy ity against M. pneumoniae in children inclusion in the review if they met the of macrolides in the treatment of M. pneumoniae remains unclear and treatment recommendations, even in TABLE 1 Medical Subject Headings Used for Primary Literature Search 4,10 major textbooks, are variable. The Mycoplasma pneumoniae pediatrica antibiotica available evidence is also conflicting. pneumonia, mycoplasma infant anti-bacterial agents One small trial suggests that b-lactam respiratory tract infections child macrolidea pneumonia child, preschool roxithromycin use in children with CAP is more cost- recurrent respiratory tract infectiona adolescent erythromycin effective than macrolides,11 several respiratory infectiona telithromycin randomized controlled trials (RCTs) RRTI macrolides community-acquired infections azithromycin fi 12,13 suggest no difference in ef cacy, lower respiratory tract infectiona clarithromycin and 2 recent pediatric cohort studies lower respiratory infectiona suggest that b-lactam–macrolide com- a Represents open-ended term.

1082 BIONDI et al Downloaded from www.aappublications.org/news by guest on October 2, 2021 REVIEW ARTICLE following criteria: randomized or ob- findings to form a consensus. Dis- review authors that assumed all C. servational studies, included children crepancies were resolved through pneumoniae were in the “improved on #17 years of age, included $10 chil- discussion with the collaborating treatment” group.20 These were then dren who were diagnosed with CA-LRTI reviewers (B.A. and S.R.). When neces- subtracted from the total. This allowed secondary to laboratory-confirmed M. sary, the principal reviewers attempted an interpretation of the results that pneumoniae, data were provided for at to contact study authors to verify would ensure that, if a treatment effect least 1 outcome measuring improve- methods and extracted data, although was identified, it was not secondary to ment, and the study compared any no responses were received. C. pneumoniae. For comparison, the antibiotic with spectrum of activity analysis was also performed assuming against M. pneumoniae with either Validity and Quality Assessment that C. pneumoniae and M. pneumoniae placebo or another antibiotic without responded equally to treatment. Treat- Studies were assessed and graded for activity against M. pneumoniae. Stud- ment effects for dichotomous outcomes level of evidence based on criteria ies that examined children with were calculated as risk difference using published by the Oxford Center for chronic respiratory illnesses such as random effects modeling. Study het- Evidence-Based Medicine.18 An 18-item cystic fibrosis, bronchiectasis, bron- erogeneity was assumed for a P ,.10 study quality assessment tool (Sup- chopulmonary dysplasia, congenital and I2 $25%. If the primary outcome plemental Information) was developed heart disease, or immunodeficiency (clinical improvement) was assessed based on recommendations published were excluded, as were those that ex- in separate methods and separate time by the Agency for Healthcare Research amined children with nosocomial and periods, each set of outcomes was in- and Quality.19 Items on the quality as- congenital infections. cluded in the review and meta-analysis. sessment tool fell within 6 domains: Meta-analysis was performed by using Studies were considered eligible for subject selection, study performance Comprehensive Meta-Analysis Version meta-analysis if, in addition to meeting bias, detection bias, subject attrition, 2 (Biostat, Englewood, NJ). criteria for inclusion in the qualitative reporting bias, and financial conflict of review, information was either pro- interest. Items were marked as factor vided or could be extrapolated represent, not present, or unclear based RESULTS garding the total numberand outcomes on review of the manuscript. Each The initial database search identified of patients with M. pneumoniae within study was given a total quality score, 4667 articles. One other citation was each treatment group. A subgroup and for this purpose, labels of “not identified via bibliography review.4 analysis of the initial meta-analysis present” or “unclear” were grouped Most articles were excluded based on was performed that included only the together. Studies were considered to a preliminary screen of the abstracts, RCTs. be unblinded when it was not specifi- leaving 40 full-text articles to be cally stated that participants or inves- assessed for eligibility. Before sub- tigators were blinded. Three reviewers Identification of Trials and Data mission, a second search was per- (E.B., R.M., and B.A.) performed in- Extraction formed (August 2012–September 2013) dividual assessments on all included The principal investigators (E.B. and to ensure that no recent articles were studies independently and then dis- R.M.) independently screened each ci- missed in the review. This search cussed each until all 3 reviewers were tation identified through the initial revealed 275 new articles, 1 of which in agreement. search strategy as definitely, possibly, was assessed for full-text eligibility or clearly not meeting inclusion crite- but did not meet study criteria.2 Of the ria. Full-text articles of all studies def- Statistical Analyses 4943 total citations reviewed, 16 met initely or possibly meeting inclusion Because much of the information rel- criteria for inclusion in the qualitative criteria were obtained regardless of evant to this review was extrapolated synthesis.5,12,13,16,17,20–30 One RCT20 used primary language of the publication. from data presented in the included different outcome metrics at 2 differ- Reviewers(E.B.andR.M.)independently studies, the review authors attempted ent times and, for the purposes of this reviewed each full-text article and then to calculate statistical significance review, was considered 2 separate reached consensus before data ab- wherever possible using x2 or Fisher’s RCTs. Therefore, 17 studies were included. straction. For full-text articles meeting exact tests. In cases for which M. Four citations were non-English and were study criteria, reviewers (E.B. and R.M.) pneumoniae and Chlamydia pneumo- translated by the review authors (2 in extracted data for each study in- niae were grouped together, a “worst- Spanish,23,30 1 in French,21 and 1 in dependently and then discussed their case scenario” was performed by the German29). Eight studies provided

PEDIATRICS Volume 133, Number 6, June 2014 1083 Downloaded from www.aappublications.org/news by guest on October 2, 2021 data20,22,27,28 or enabled the review authors M. pneumoniae spectrum with non–M. tifiedbiascategory(12/17studies,71%) (E.B. and R.M.) to extract data13,16,23 pneumoniae spectrum antibiotics, and (Table 3). The most common cause of specifically comparing treatment with 2 case–control studies27,28 compared study performance bias was unblinded no treatment of M. pneumoniae in chil- macrolide treatment in patients with participants or observers. More than dren diagnosed with M. pneumoniae macrolide-sensitive and macrolide- half of the RCTs (6/11, 55%) reported and were included in the quantitative resistant M. pneumoniae. Esposito a potential conflict of interest, which in analysis; 5 were RCTs (Fig 1).13,16,20,23 et al20 included upper respiratory all cases was pharmaceutical spon- Therewere4294patientsenrolledinthe infections (URTIs) (199/352, 57%) in sorship with use of that company’s 17 included studies (Table 2). Of these, addition to CA-LRTI (153/352, 43%) but drug in the study. an aggregate 2648 patients could be did not provide a separate analysis, used to compare an agent treating M. and therefore all are included in this Spectrum-Specific Treatment in pneumoniae with an agent that did not review, as was done previously.10 One Pediatric CA-LRTI Secondary to treat M. pneumoniae in CA-LRTI. This study did not report the total number M. pneumoniae 17 number excludes comparator arms in of pediatric patients. Studies used Nine studies5,13,16,20,22,23,27,28 provided which antibiotics of the same class or various combinations of methods for enough detail to allow a comparison of spectrum were compared and patients M. pneumoniae detection; 5 used culture M. pneumoniae spectrum and non- 17,23,24,27,28 in whom therapy could not be de- of nasopharyngeal specimens, spectrum treatment in pediatric – – termined. One randomized trial pro- 15 used serology,5,12,13,16,17,20 26,28 30 patients with CA-LRTI secondary to M. vided data on a total of 155 patients but and 6 used polymerase chain reac- pneumoniae and included a total of 723 compared macrolide therapy with ei- tion assays of upper airway speci- such patients (Table 4). In terms of ther a b-lactam or another macrolide mens.12,20,23,25,26,28 No trials were overall clinical improvement, 4 of 5 and did not provide separate statis- placebo controlled, and a number of RCTs found no clinical benefit.13,16,20,23 tics.30 studies received pharmaceutical com- One retrospective cohort study identi- Ten RCTs12,13,17,20,23,24,26,29,30 and 3 co- pany funding.12,13,16,20,24,25 fied a significant improvement in du- hort studies5,22,25 compared a mac- When we evaluated for overall quality ration of fever in patients treated with rolide with a nonmacrolide in the among all studies included, detection macrolides, although clinically there treatment of CA-LRTI. One RCT16 com- bias, which included items such as was a large overlap in fever duration pared levofloxacin with a b-lactam, 1 validity and reliability of study out- range (3.0–6.8 vs 3.4–7.9 days; P = .04), retrospective cohort study21 compared comes, was the most frequently iden- and there was not a significant

FIGURE 1 Flowchart for included studies. a One article20 used 2 different outcome metrics at 2 different time periods and was therefore treated as 2 separate studies.

1084 BIONDI et al Downloaded from www.aappublications.org/news by guest on October 2, 2021 EITISVlm 3,Nme ,Jn 2014 June 6, Number 133, Volume PEDIATRICS TABLE 2 Study Characteristics and Quality, in Chronological Order Study Design/Length Patients Intervention/Comparator Relevant Outcomes Relevant Results Level of Evidence Sterner et al (1967) RCT/1.5 y N = 180 (unknown EES/cephaloridine Failure = daily fevers 8/87 (EES) vs 17/93 (cephaloridine) failures.a Mean 2b ages) with CAP for 5 d or fever $10 d fever = 2.7 vs 3.8 d. In those with MP, 0/13 vs 3/9 failures.a Does not separate adults and children. Ruhrmann et al (1982) RCT/1 y N = 120 (6 mo–14 yr) EES/amoxicillin Fever duration Mean fever = 2.6 6 4.1 d (EES) vs 2.4 6 1.9 d 2b with CAP (amoxicillin). Garo et al (1988) Retro, multisite N = 182 (mean age = MP spectrum/no spectrum Fever duration Mean fever = 3 d (spectrum) vs 7 d (no spectrum). 4 cohort/5 y 29 y) with MP 84% presented with atypical pneumonia. Does

Downloaded from not separate adults and children. Gomez Campdera RCT/2 y N = 155 (6 mo–16 y) AZM/co-amoxiclav Cured or improved At 3, 10, and 30 d, respectively: 78/82 (AZM) vs 66/73 2b et al (1996) with CAP (,5 y); EES (.5y) at 3, 10, or 30 d (comparator),a 80/82 vs 68/73,a and 80/82 vs 69/73.a No difference in duration of fever or cough. 15 total with MP. Gendrel et al (1997) Pro, cohort/2 y N = 104 (1.5–13 y) Macrolide/b-lactam Fever at 2–18 d In children with MP: 11/11 (macrolide) vs 2/32 4 a www.aappublications.org/news with CAP (b-lactam) became afebrile (P , .001). Harris et al (1998) Multisite, RCT/1.5 y N = 456 (6 mo–16 y) AZM/co-amoxiclav 1. Cured or improved In ,5 y group: clinical success in 114/125(AZM) 1b with CAP (,5 yr); EES ($5 yr) at 15–19 d vs 59/63 (co-amoxiclav) at 15–19 d,a and 97/114 2. Cured or improved at 4–6wk vs 41/48 at 4–6wk.1 35/310 vs 45/146 had adverse 3. Adverse events events (P , .001).a Sáez-Llorens Multisite, RCT/2 y N = 335 (6 mo–15 y) AZM/co-amoxiclav Cured or improved after 96/97 (AZM) vs 114/116 (co-amoxiclav) cured or 2b et al (1998) with CAP (,5 yr); EES ($5 yr) 3 d of treatment improved. In cases of MP, 9/9 and 5/5 cured or improved.a Wubbel et al (1999) RCT/2 y N = 174 (6 mo–16 y) AZM/co-amoxiclav 1. Cured 3 d after treatment 68/69 (AZM) vs 75/78 (comparator) cured. 10/69 2b with CAP (,5 yr); EES ($5 yr) 2. Adverse events (AZM) vs 33/49 (co-amoxiclav) had adverse a byguest on October2,2021 events (P , .001). Ferwerda et al (2001) Multisite, RCT/3 y N = 118 (3 mo–12 y) AZM/co-amoxiclav 1. Cured or improved at 10–13 d Success in 50/55 (AZM) vs 46/53 (co-amoxiclav) 1b with CAP 2. Cured or improved at 25–30 d at 10–13 d and 46/51 vs 43/50 at 25–30 d. 3. Adverse events 33/59 vs 41/58 adverse events.a Principi et al (2001) Pro, multisite N = 613 (2–14 y) Macrolide/no spectrum Cured or improved at 4–6 wk In subjects with MP or CP, success in 106/109 2b cohort/1 y with CAP (macrolide) vs 67/82 (comparator) (P , .001).a Kogan et al (2003) RCT/3 y N = 110 (30 d–14 y) Classic CAP: AZM/amoxicillin Clinical response: Mean fever = 1.7 d (AZM) vs 2.0 d (amoxicillin). 2b with CAP Fever duration By day 14 all 47 CXRs had improved by .75%. CXR .75% improved Esposito et al RCT/2 y N = 352 (1–14 y) with AZM and symptomatic/ Clinical success at 4–6 wk 76/76 (AZM) vs 88/114 (comparator) with MP 2b (2005), study A recurrent RTI symptomatic or CP had clinical success (P , .001). Esposito et al (2005), RCT/2 y N = 352 (1–14 y) with AZM and symptomatic/ #2 RTIs at 6 mo 53/71(AZM) vs 61/109 (comparator) with MP 2b study B recurrent RTI symptomatic or CP #2 RTIs (P = .01).a Bradley et al (2007) Multisite, RCT/2 y N = 738 (6 mo–16 y) Levoﬂoxacin/b-lactam Improved or cured at Success was 409/441(levoﬂoxacin) vs 138/147 2b with CAP (,5 yr); macrolide ($5 yr) 1–3 d and cured at 10–17 d (comparator) at 1–3 d and 439/503 vs 145/170 at 10–17 d.

Lu et al (2008) Retro, cohort/2 y N = 139 (8 mo–12 y) Macrolide/no macrolide Duration of fever Mean fever = 4.90 6 1.89 d (macrolide) 2b ARTICLE REVIEW with MP CA-RTI vs 5.63 6 2.22 d (no macrolide) (P = .04). Matsubara et al (2009) Retro, case– N =94(0–14 y) with MP Macrolide sensitive MP/ 1. Excellent or good 43/47 (sensitive) vs 5/22 (resistant) had positive 3b control/4 y CAP treated with resistant MP clinical response clinical response (P , .01). Mean fever = 1.5 macrolide 2. Mean fever duration vs 4.0 d (P , .01) and mean cough = 1.5 1085 3. Cough duration vs 4.0 d (P , .01). difference in other signs or symptoms, scenario,” thesamemeta-analysis,as-

3b white blood cell count, or C-reactive suming C. pneumoniae and M. pneu- Level of Evidence protein levels.5 One retrospective moniae responded equally to treatment, case–control study found an overall demonstrated a pooled risk difference clinical beneﬁt and 2 other studies of 0.12 (95% CI, 0.01 to 0.22). identiﬁed a decrease in fever duration, although these were rated as being DISCUSSION among the lowest quality of evidence in this review.22,27,28 Esposito et al2o (study The majority of studies included in our systematic review did not show a sig-

Mean DNA load A) was the only RCT to identify a clinical a niﬁcant clinical beneﬁtofM. pneumo-

cantly. fi pro, prospective; retro, retrospective; RTI, respiratory bene t, but the patient population in- fi Relevant Results .001). cluded children with URTI. Additionally, niae spectrum therapy in CA-LRTI. Of , fi

P the 9 studies that speci cally examined the data were categorized only as the issue of M. pneumoniae treatment “atypical” bacteria and did not sepa- in children with CA-LRTI secondary to afebrile ( decreased signi rate C. pneumoniae from M. pneumoniae.

8/8 (sensitive) vs 6/21 (resistant) became M. pneumoniae, almost all the pro- Twenty-seven patients had acute C. pneu- Mycoplasma pneumoniae; spective studies showed no clinical moniae infection only, and the “worst- beneﬁt. The remaining studies gener- case scenario” excludes all patients with ally suggest a statistical, but not nec- C. pneumoniae after assuming each im- essarily clinically relevant, decrease in proved on macrolide therapy. fever duration, and most of these are Two meta-analyses were attempted rated as low- or lowest-quality evi- using risk difference as the treatment dence. Our meta-analysis of RCTs sug-

starting therapy effect. The ﬁrst included all studies gests a small treatment beneﬁtin 1. Afebrile 2 d after 2. MP DNA load at 2 d listed in Table 4 that used dichotomous patients with CA-LRTI secondary to M. variables.13,16,20,22,23,27,28 However, there pneumoniae. However, the pooled ef- was a large degree of heterogeneity fect favoring treatment was driven

CXR, chest x-ray; EES, erythromycin ethylsuccinate; MP, (P , .001) in addition to publication bias primarily by the result of a single RCT and largely variable treatment effects. that included children with URTIs (57% A second analysis used only the of enrolled patients). Four of the 5 RCTs13,16,20,23 and demonstrated a pooled studies included in the meta-analysis MP/resistant MP Intervention/Comparator Relevant Outcomes risk difference of 0.12 (95% conﬁdence ﬁ

Macrolide sensitive did not show a bene tofM. pneumo-

Chlamydia pneumoniae; interval [CI], 20.04 to 0.20) (Fig 2). This niae spectrum therapy, and the sum- risk difference represents the absolute mary statistic is not statistically change in risk attributable to treat- signiﬁcant. The associated funnel plot ment with a macrolide. In our case, the demonstrates a paucity of small stud- risk in the treated group minus the risk ies showing a treatment effect, sug- Patients 15 y) with MP – in the control group was 12% (95% CI, gesting either publication bias against 24% to 20%). This ﬁnding suggests such studies, poor methodological de- CAP treated with macrolide =29(1

N that 12% of children treated with sign, or artifact heterogeneity among a macrolide will have more rapid clin- smaller studies. ical improvement, corresponding to The question of M. pneumoniae spec- – a number needed to treat of 8.33, but trum (specifically macrolide) use in the confidence interval overlapping 0% CAP, or other CA-LRTIs, is commonly control/5 y Design/Length negates statistical significance. There encountered in pediatric practice. Two remained significant heterogeneity large, retrospective cohort studies cant or not calculable unless otherwise stated. AZM, azithromycin; CP, fi between the studies (P =.02).The based on administrative data sets have funnel plot revealed potential for indirectly addressed the issue of M. Continued

Study publication bias against small studies pneumoniae treatment by comparing cance calculated by the review authors. ﬁ that show a treatment effect (Fig 3). To differences in length of hospital stay

Signi b Kawai et al (2012) Retro, case TABLE 2 Differences not signi a tract infection. provide comparison with the “worst-case among children receiving -lactam

1086 BIONDI et al Downloaded from www.aappublications.org/news by guest on October 2, 2021 REVIEW ARTICLE

TABLE 3 Bias Assessment for Individual Studies either database and therefore could Study Selection Performance Detection Attrition Reporting Conflict Overall not be assessed. Macrolides are among of Interest the most frequently prescribed classes Sterner et al (1967) x —————4/15 of antibiotics for respiratory tract —— —— — Ruhrmann x 5/15 infections in children, particularly in et al (1982) Garo et al (1988) x x x — x — 8/14 conditions for which antibiotics are not Gomez Campdera xxxx——7/17 clearly indicated.15 Additionally, na- et al (1996) tional guidelines intended to reduce Gendrel et al (1997) x x x — x — 8/14 Sáez-Llorens x — xx——6/15 practice variability in the management et al (1998) of CAP in children recommend consid- Harris et al (1998) ———x — x 4/15 eration of macrolide use but state that Wubbel et al (1999) x x x x x x 10/15 3 Ferwerda —————x 3/15 this is based on limited data. Our re- et al (2001) view found insufficient evidence to Principi et al (2001) —— x x x x 7/14 support or refute antibiotic use in CA- Kogan et al (2003) x x x x ——8/15 Esposito et al — xx—— x 6/15 LRTI secondary to M. pneumoniae. This (2005), study A suggests that macrolide use in CA-LRTI Esposito et al xxxx— x 8/15 could represent an area of opportunity (2005), study B Bradley et al (2007) — xx—— x 5/15 for reducing antibiotic use, which could Lu et al (2008) ——————2/14 have a significant impact on health care Matsubara ———x ——2/15 costs and the development of antimi- et al (2009) crobial resistance.32,33 Kawai et al (2012) — xx— x — 6/15 —, substantial bias not identified within the domain; x, substantial bias identified within the domain. Our review provides novel information in comparison with the previous review TABLE 4 Studies Comparing Spectrum With Nonspectrum Treatment of Children With Acute on the topic by including more enrolled Respiratory Infection With M. Pneumoniae Grouped by Outcome Terma patients (4294 vs 1912) and more Clinical Improvement at #5d studies (17 vs 7).10 Although the pre- fi Study Time Frame Outcome Treatment (n) Comparator (n) Pb vious review reports insuf cient data to draw conclusions, we suggest that Improved Total Improved Total the available literature does not cur- Sáez-Llorens et al (1998) 3 d Overall 9 9 5 5 ..99 rently support treatment of CA-LRTI Matsubara et al (2009) ,5 d Fever 43 47 5 22 ,.001 Kawai et al (2012) 2 d Fever 8 8 6 21 ,.001 secondary to M. pneumoniae. We examined observational studies, many of Study Time Frame Days of Feverc Days of Feverc which were assessed as having the Lu et al (2008) N/A Fever 4.9 6 1.89 5.63 6 2.22 .04 same level of evidence as the RCTs. In Clinical Improvement at .5d the absence of high-quality RCTs, Study Time Frame Outcome Treatment (n) Comparator (n) Pb experts suggest it is reasonable to include cohort and case–control studies Improved Total Improved Total in a systematic review, particularly Gendrel et al (1997) 2–18 d Fever 9 9 2 32 ,.001 when the quality of evidence is equiv- Harris et al (1998) 15–19 d Overall 21 21 9 9 ..99 34 fi Esposito et al (2005), 1 m Overall 49 49 88 114 ,.001 alent. Additionally, we identi ed 3 study Ad randomized trials that had not been Esposito et al (2005), 6 m Recurrent 26 45 61 109 .86 previously examined, 1 of which was study Bd Bradley et al (2007) 10–17 d Clinical 59 66 15 18 .44 assessed as having the highest level of 12,16,17 a Numbers often extrapolated by the review authors. evidence of any included study. b Includes P values calculated by the review authors. Although our review failed to identify c Results are mean 6 SD. d Worst-case scenario with all 27 C. pneumoniae patients placed in the treatment improved group and then excluded. clear therapeuticefficacyofmacrolides in pediatric CA-LRTI due to M. pneumoniae, monotherapy and children receiving length of stay among children receiving there are data in support of antibi- combination therapy with a macro- combination therapy. However, testing otic treatment. Military studies from lide.2,31 These studies found a shorter for M. pneumoniae was not recorded in the 1960s, performed on recruits in

PEDIATRICS Volume 133, Number 6, June 2014 1087 Downloaded from www.aappublications.org/news by guest on October 2, 2021 FIGURE 2 Meta-analysis of randomized trials comparing spectrum and nonspectrum treatment of M. pneumoniae in children. Studies were weighted by size and degree of heterogeneity. The overall effect (diamond) represents the pooled risk differences of the 5 studies. The confidence interval crosses 0.00, suggesting a lack of statistical significance. basic training, found that recruits necessarily represent the typical clini- often difficult to find a middle ground with atypical pneumonia more often cal situation encountered by pedia- between throwing out all summary had serologic evidence of M. pneumoniae tricians today, given their young adult, estimates and inappropriately com- infection (identified as Eaton’s agent homogenous population, nor can they bining studies, particularly in pediat- at the time) than asymptomatic con- account for mixed infections or mod- rics, where the paucity of high-quality trols.35,36 Additionally, in a random- ernized diagnostic testing. Nonetheless, RCTs is acute. We provided a summary ized, double-blind, placebo-controlled these studies describe principles of statistic for 1 broad study question, but study among 300 recruits with atypi- study design and conduct in this area we noted significant study heteroge- cal pneumonia due to M. pneumoniae, that can serve as a template for pedi- neity that we could not resolve with treatment with a tetracycline was found atric investigations. subgroup analysis. Therefore, our in- to decrease the duration of a number of One of the strengths of meta-analysis is terpretation of the entirety of the sta- clinical signs and symptoms faster than in the process of evaluating the avail- tistical testing associated with our placebo.37 These seminal studies do not able literature for heterogeneity. It is meta-analysis is that despite multiple

FIGURE 3 Funnel plot of standard errors by risk difference for all studies in the second meta-analysis. Publication bias is evident against small studies that show a treatment effect.

1088 BIONDI et al Downloaded from www.aappublications.org/news by guest on October 2, 2021 REVIEW ARTICLE studies, a clear treatment effect has not limitation should be balanced against CONCLUSIONS emerged. Furthermore, the study het- the fact that we abstracted quantitative Our systematic review provides in- data from more studies than previous erogeneity lends credence to the idea sufficient evidence to support con- meta-analyses on this subject. that a large RCT is necessary to provide clusions about the efficacy of macrolide guidance to pediatricians on this clin- Third, our assessment of treatment treatment of CA-LRTI due to M. pneu- ical question. effect was limited primarily to sub- moniae in children. Fever duration may There are several limitations to our jective outcome measures such as be decreased, but the clinical impact of review, most related to the quality of the symptomatic improvement or resolu- this effect is unclear, and we identified available data on the topic. First, most tion. This limitation is inevitable given few high-quality studies to support it. studies contained substantial bias or the primary endpoints used in the in- Our review represents the most com- financial conflict of interest. We at- cluded studies, but our bias assess- prehensive evaluation of treating M. fi tempted to account for this through in- ment speci cally addresses the pneumoniae in children with CA-LRTI to depth quality assessment and reporting subjective nature of the data. date. The findings from this review to allow a critical interpretation of Fourth, M. pneumoniae testing modal- highlight the need for high-quality, study results. In particular, detection ities have changed dramatically over prospective studies to assess the im- bias was commonly identified in the time, with variable testing character- pact of antibiotic therapy to treat CA- studies comparing spectrum with istics and lack of overall consistency. LRTI caused by M. pneumoniae in children. nonspectrum therapy in patients di- This limits the uniformity of the patient These studies should specifically ad- agnosed with M. pneumoniae. This bias population included in our review. Fur- dress the potential for confounding most often represented lack of blinding thermore, coinfection with a pathogen in mixed infections, timing of intervention of investigators or participants or lack addition to M. pneumoniae is common relative to symptom onset, and testing of a reliable outcome measure (ie, clini- but rarely identified via clinical testing, modalities that include a combination cal improvement). This quality issue making it difficult to draw specificcon- of serology and polymerase chain re- could potentially skew study results to- clusions about therapeutic effects.9,38 action assays. ward or away from the null hypothesis. Finally,nostudyincludedinouranalysis Second, we were unable to extract data took into account the duration of illness ACKNOWLEDGMENTS on outcomes for all studies, reducing before the start of antibiotic therapy. It The authors thank Dr Swati Murthy for the number of articles that could be is possible that the timing of in- her help with data collection and Dr Jill included in the meta-analysis and po- tervention relative to the start of Halterman for her review of the final tentially causing publication bias. This symptoms is related to the outcome. manuscript.

REFERENCES

1. Pati S, Lorch SA, Lee GE, Sheffler-Collins S, 4. Shah SS, Test M, Sheffler-Collins S, Weiss AK, among Japanese children. JInfect. 1988;16 Shah SS. Health insurance and length of Hall M. Macrolide therapy and outcomes (3):253–261 stay for children hospitalized with community- in a multicenter cohort of children hos- 8. Waites KB. New concepts of Mycoplasma acquired pneumonia. J Hosp Med. 2012;7 pitalized with Mycoplasma pneumoniae pneumoniae infections in children. Pediatr (4):304–310 pneumonia. JHospMed. 2012;7(4):311– Pulmonol. 2003;36(4):267–278 2. Ambroggio L, Taylor JA, Tabb LP, Newschaffer 317 9. Korppi M, Heiskanen-Kosma T, Kleemola M. CJ, Evans AA, Shah SS. Comparative ef- 5. Lu YJ, Chen TH, Lin LH, Shen CM, Huang CH. Incidence of community-acquired pneumonia fectiveness of empiric b-lactam mono- Macrolide use shortens fever duration in in children caused by Mycoplasma therapy and b-lactam–macrolide combination Mycoplasma pneumoniae infection in chil- pneumoniae: serological results of a pro- therapy in children hospitalized with dren: a 2-year experience. J Microbiol Immu- spective, population-based study in primary community-acquired pneumonia. J Pediatr. nol Infect. 2008;41(4):307–310 health care. Respirology.2004;9(1):109–114 2012;161(6):1097–1103 6. Defilippi A, Silvestri M, Tacchella A, et al. 10. Mulholland S, Gavranich JB, Gillies MB, 3. Bradley JS, Byington CL, Shah SS, et al. The Epidemiology and clinical features of My- Chang AB. Antibiotics for community- management of community-acquired pneu- coplasma pneumoniae infection in chil- acquired lower respiratory tract infec- monia in infants and children older than 3 dren. Respir Med. 2008;102(12):1762–1768 tions secondary to Mycoplasma pneumoniae months of age: clinical practice guidelines 7. Sakurai N, Nagayama Y, Honda A, Makuta in children. Cochrane Database Syst Rev. by the Pediatric Infectious Diseases Society M, Yamamoto K, Kojima S. Mycoplasma 2012;9:CD004875 and the Infectious Diseases Society of pneumoniae and other pathogens in the 11. Aurangzeb B, Hameed A. Comparative effi- America. Clin Infect Dis. 2011;53(7):e25–76 aetiology of lower respiratory tract infections cacy of amoxicillin, cefuroxime and

PEDIATRICS Volume 133, Number 6, June 2014 1089 Downloaded from www.aappublications.org/news by guest on October 2, 2021 clarithromycin in the treatment of tract infections. Pediatr Infect Dis J. 2005;24 29. Ruhrmann H, Blenk H. Erythromycin versus community-acquired pneumonia in chil- (5):438–444 amoxicillin for the treatment of pneumonia dren. J Coll Physicians Surg Pak. 2003;13 21. Garo B, Garré M, Quiot JJ, et al. Myco- in children (author’s transl) [in German]. (12):704–707 plasma pneumoniae infections. A multi- Infection. 1982;10(Suppl 2):S86–S91 12. Ferwerda A, Moll HA, Hop WC, et al. Efficacy, center retrospective study of 182 cases [in 30. Gomez Campdera JA, Navarro Gomez ML, safety and tolerability of 3 day azithromycin French]. Presse Med. 1988;17(29):1475– Hernandez Sampelayo T, Merello Godino C, versus 10 day co-amoxiclav in the treat- 1478 Sanchez Sanchez C. Azithromycin in the ment of children with acute lower re- 22. Gendrel D, Raymond J, Moulin F, et al Eti- treatment of ambulatory pneumonia in spiratory tract infections. J Antimicrob ology and response to antibiotic therapy of children. Acta Pediatr Esp. 1996;54(8):554– Chemother. 2001;47(4):441–446 community-acquired pneumonia in French 562 13. Harris JA, Kolokathis A, Campbell M, Cassell children. Eur J Clin Microbiol Infect Dis. 31. Leyenaar JK, Shieh MS, Lagu T, Pekow PS, GH, Hammerschlag MR. Safety and efficacy 1997;16(5):388–391 Lindenauer PK. Comparative effectiveness of azithromycin in the treatment of community- 23. Sáez-Llorens X, Castaño E, Wubbel L, et al. of ceftriaxone in combination with a mac- acquired pneumonia in children. Pediatr Importance of Mycoplasma pneumoniae rolide compared with ceftriaxone alone for Infect Dis J. 1998;17(10):865–871 and Chlamydia pneumoniae in children pediatric patients hospitalized with community acquired pneumonia [published 14. Toltzis P. Determining the best regimen for with community-acquired pneumonia [in – online ahead of print]. Pediatr Infect Dis J. children hospitalized with CAP. J Pediatr. Spanish]. Rev Med Panama. 1998;23(2):27 2013 2012;161(6):A1–A2 33 32. Carbon C, Isturiz R. Narrow versus broad 15. Hersh AL, Shapiro DJ, Pavia AT, Shah SS. 24. Wubbel L, Muniz L, Ahmed A, et al. Etiology spectrum antibacterials: factors in the se- Antibiotic prescribing in ambulatory pedi- and treatment of community-acquired lection of pneumococcal resistance to beta- atrics in the United States. Pediatrics. 2011; pneumonia in ambulatory children. Pediatr Infect Dis J. 1999;18(2):98–104 lactams. Drugs. 2002;62(9):1289–1294 128(6):1053–1061 25. Principi N, Esposito S, Blasi F, Allegra L; 33. Davey P, Brown E, Charani E, et al. Inter- 16. Bradley JS, Arguedas A, Blumer JL, Sáez-Llorens Mowgli Study Group. Role of Mycoplasma ventions to improve antibiotic prescribing X, Melkote R, Noel GJ. Comparative study of pneumoniae and Chlamydia pneumoniae in practices for hospital inpatients. Cochrane levofloxacin in the treatment of children with children with community-acquired lower Database Syst Rev. 2013;4:CD003543 community-acquired pneumonia. Pediatr In- respiratory tract infections. Clin Infect Dis. 34. Walker E, Hernandez AV, Kattan MW. Meta- fect Dis J. 2007;26(10):868–878 2001;32(9):1281–1289 analysis: its strengths and limitations. 17. Sterner G, Fransén H, Tunevall G, Svedmyr 26. Kogan R, Martínez MA, Rubilar L, et al. Cleve Clin J Med. 2008;75(6):431–439 A. A comparative trial of cephaloridine and Comparative randomized trial of azi- 35. Forsyth BR, Bloom HH, Johnson KM, Chanock erythromycin in hospitalized pneumonic thromycin versus erythromycin and amox- RM. Etiology of primary atypical pneumonia patients. Postgrad Med J. 1967;43(Suppl icillin for treatment of community-acquired in a military population. JAMA. 1965;191: – – 43):43 ,53 54 pneumonia in children. Pediatr Pulmonol. 364–368 18. Oxford Centre for Evidence-Based Medicine. 2003;35(2):91–98 36. Chanock RM, Mufson MA, Bloom HH, James The Oxford levels of evidence 2. 2013. 27. Matsubara K, Morozumi M, Okada T, et al A WD, Fox HH, Kingston JR. Eaton agent Available at: www.cebm.net/index.aspx? comparative clinical study of macrolide- pneumonia. JAMA. 1961;175:213–220 o=5653. Accessed August 29, 2013 sensitive and macrolide-resistant Myco- 37. Kingston JR, Chanock RM, Mufson MA, et al. 19. Viswanathan M, Ansari MT, Berkman ND, plasma pneumoniae infections in pediatric Eaton agent pneumonia. JAMA. 1961;176:118– et al. Assessing the Risk of Bias of In- patients. J Infect Chemother. 2009;15(6): 123 dividual Studies in Systematic Reviews of 380–383. 38. Michelow IC, Olsen K, Lozano J, Duffy LB, Health Care Interventions. Rockville, MD: 28. Kawai Y, Miyashita N, Yamaguchi T, et al. McCracken GH, Hardy RD. Diagnostic utility Methods Guide for Effectiveness and Com- Clinical efficacy of macrolide antibiotics and clinical significance of naso- and oro- parative Effectiveness Reviews; 2008 against genetically determined macrolide- pharyngeal samples used in a PCR assay to 20. Esposito S, Bosis S, Faelli N, et al. Role of resistant Mycoplasma pneumoniae pneu- diagnose Mycoplasma pneumoniae infection in atypical bacteria and azithromycin therapy monia in paediatric patients. Respirology. children with community-acquired pneumonia. for children with recurrent respiratory 2012;17(2):354–362 J Clin Microbiol. 2004;42(7):3339–3341

1090 BIONDI et al Downloaded from www.aappublications.org/news by guest on October 2, 2021 Treatment of Mycoplasma Pneumonia: A Systematic Review Eric Biondi, Russell McCulloh, Brian Alverson, Andrew Klein, Angela Dixon and Shawn Ralston Pediatrics 2014;133;1081 DOI: 10.1542/peds.2013-3729 originally published online May 26, 2014;

Updated Information & including high resolution figures, can be found at: Services http://pediatrics.aappublications.org/content/133/6/1081 References This article cites 36 articles, 3 of which you can access for free at: http://pediatrics.aappublications.org/content/133/6/1081#BIBL Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.aappublications.org/site/misc/Permissions.xhtml Reprints Information about ordering reprints can be found online: http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on October 2, 2021 Treatment of Mycoplasma Pneumonia: A Systematic Review Eric Biondi, Russell McCulloh, Brian Alverson, Andrew Klein, Angela Dixon and Shawn Ralston Pediatrics 2014;133;1081 DOI: 10.1542/peds.2013-3729 originally published online May 26, 2014;

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://pediatrics.aappublications.org/content/133/6/1081

Data Supplement at: http://pediatrics.aappublications.org/content/suppl/2014/05/20/peds.2013-3729.DCSupplemental

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2014 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on October 2, 2021