StudyDesignSample sizeInterventionReason for exclusion
Abuekteish et al.19951Observational140BUD vs. BDPNo comparison group, cross- sectional analysis of 140 asthmatics with ICS treatment over 5 years.; no description of analysis; no adjustment for duration and dose of ICS;
Acun et al. 20052RCT100BUD vs. FPInsufficient reporting to allow for appraisal of methods and analysis; Results not reported.
Adachi et al., 20073RCT319CIC vs. BUDComparisons were between medium dose CIC, high dose CIC, and low dose BUD; no information on randomization scheme; no blinding (BUD group used a spacer whereas CIC groups did not); some baseline differences between groups; no information on attrition/dropouts for those who were randomized; no information on whether intention to treat or per protocol analysis used.
Agertoft et al. 19944Observational278BUD vs. controlAttrition NR, but high in other corresponding publication; high potential selection bias
Agertoft et al.20005Observational338BUD vs. controlHigh attrition and differential attrition; high potential for selection bias (mainly due to attrition); 97/270 in the BUD group had not yet attained adult height and were thus not analyzed.
Allen et al. 19946Meta-analysis810BUDLack of an appropriately described comprehensive, systematic literature search...
Anthracopoulos et al. 20077Observational641BUD vs. FPHigh potential for selection bias and confounding, very high attrition (low participation rate), unclear how patients were identified/selected/recruited, unclear if appropriate dosage comparison, open-label, unclear which confounders were adjusted for in the analyses (and no mention of parental height), analysis excluded children that required more than 36 months of ICS and those that entered puberty.
Aubier et al. 19998RCT503FP/SM vs. FP + SM vs. FPPoor reporting of methods and results of meaningful outcome
Bakhireva et al. 20079Observational96LTRAs vs. SABAs and controlSmall sample size (inadequate to detect differences in adverse events of interest).
Barnes et al. 200710RCT75MOM vs. BUDBaseline differences, lack of reporting of randomization, blinding, equal assessment of both groups,
Bleecker et al. 200611Pooled analysis183FP/SMPotential selection bias (from two different RCTs, just 183 (43%) of subjects had available genotype information; not clear how these were chosen; potential confounding, analyses don’t adjust for baseline SABA use or symptom scores which were slightly worse in the B16 Gly/Gly group; sample size-- studies not powered to detect differences among genotypes
Davis et al.12Meta-analysisNROmalizumabMethods not reported
Ferguson et al. 200713RCTBUD vs. FPAttrition high (> 40%), potential selection bias, less than 60% of subjects completed the 1 year study; did not account for greater # of steroid courses in BUD group (15 vs. 6); post- randomization exclusions
Kallen et al.14Observational2014BBUDPoor measurement and uncontrolled confounders
Karaman et al. 200715RCT67BUD vs. BUD+MOM vs. BUD+FMHigh attrition, masking not reported at any level, type or withdrawal/exclusion not reported and dropout rate significant, no ITT analysis, no explanation of why many randomized subjects not included in the analyses, no mention of statistical power
Lipworth et al. 199916Meta-analysisNRICSSearch terms not specified; meta-analysis methods not adequately reported; not independently reviewed; no report of publication bias, heterogeneity, or clear eligibility criteria; unclear how meta- analysis was carried out other than multiple regression.
Nong et al. 200117RCT77BDP vs. FPHigh potential for bias; Completer’s analysis; 22% post- randomization exclusions; incomplete inclusion/exclusion criteria; not sure it was actually randomized;
Ohaju-Obodo et al. 200518RCT109BUD vs. BDPHigh potential for selection and measurement bias; no blinding, analysis not described, unable to determine attrition, did not report randomization/allocation concealment methods
Palmer et al. 200619Observational546SMNo baseline data given for comparison of groups so unable to adequately assess potential for selection bias
Pauwels et al. 199820RCT340FP vs. BDPPoor reporting, confounding
Perng et al. 200421RCT49BUD vs. BUD+ ZAFHigh potential for selection bias and measurement bias
Riccioni et al. 200222RCT45BUD vs. MOMOpen-label, no ITT analysis, no reporting of majority of criteria for critical appraisal
Scott et al. 199923Pooled data670BUDPooled data analysis without a systematic literature search
Wardlaw et al. 200424RCT167MOM vs. FPNo blinding, randomisation method nr, no withdrawal information reported
Weiss et al. 200525RCT945BUD vs. TRAHigh potential for selection and measurement bias; all groups unblinded, not ITT analysis, ICS dosing was left to the discretion of the physician (starting dose and subsequent adjustments) making us unable to determine if the comparison is appropriate (nothing reported on actual dosing received.
Yurdakul et al. 200226RCT64BUD+FM vs. BUD+ZAFNot truly randomized---thus not really an RCT, allocation, blinding, etc. Nothing about withdrawals. Unable to determine if ITT analysis or what was done.

From: Appendix F, Studies of poor quality

Cover of Drug Class Review: Controller Medications for Asthma
Drug Class Review: Controller Medications for Asthma: Final Update 1 Report [Internet].
Jonas DE, Wines RCM, DelMonte M, et al.
Portland (OR): Oregon Health & Science University; 2011 Apr.
Copyright © 2011 by Oregon Health & Science University.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.