Page 74 - ADULT-ONSET ASTHMA PREDICTORS OF CLINICAL COURSE AND SEVERITY

ADULT-ONSET ASTHMA – PREDICTORS OF CLINICAL COURSE AND SEVERITY
RESULTS
SEARCH AND SELECTION
The searches retrieved 2,919 unique records, all of them providing tles and/or abstracts in English language. Among these, we found 21 eligible diagnos c accuracy studies (Figure 1). Another 18 studies ful lled the eligibility criteria for the enrichment sample. Contac ng their authors led to eight addi onal inclusions. We could also include data from two studies from our own department, and iden ed one more through a conference poster. No addi onal studies were iden ed by scanning reference lists and searching trial registries. Overall, we included 24 studies performed in adults, and eight in children (Appendix 2).
STUDY CHARACTERISTICS
Detailed characteris cs of included studies are provided in Appendix 3. In summary, all studies used a single set of inclusion criteria (cohort studies) and the number of pa ents included in the analysis of diagnos c accuracy varied from 24 to 566 in adults, and from 27 to 150 in children. The mean/median age ranged from 27.0 to 59.8 in adults, and from 6.8 to 13.0 in children.
In all cases, study par cipants had been recruited in secondary or ter ary care facili es and both males and females had been included. Studies in adults included asthma pa ents with varying severity: mild-moderate (n=4; 17%), mild-severe (n=4; 17%), moderate-severe (n=4; 17%), severe (n=5; 21%), or not reported (n=7; 29%). In children, asthma severity was mild (n=1; 13%), mild-severe (n=1; 13%), moderate-severe (n=1; 13%), severe (n=2; 25%), or not reported (n=3; 38%). In adults, 12 studies (50%) included current non-smokers only, one study (4%) current smokers only, and 11 studies (46%) included both.
Two studies in adults (8%) evaluated cor costeroid (inhaled and/or oral) untreated pa ents only, 11 studies (46%) evaluated cor costeroid-treated pa ents only, and 11 studies (46%) included both treated and untreated pa ents. In children, these numbers were one (13%), three (38%), and four (50%), respec vely. There were large between-study di erences in atopy and asthma severity status.
In adults, 21 studies (88%) used only sputum as the reference standard, whereas two studies (8%) used sputum and EBB, and one study (4%) used BAL and EBB. In children, sputum was the reference standard in four studies (50%), BAL in two studies (25%), BAL and EBB in one study (13%), and sputum, BAL and EBB in one study (13%).
RISK OF BIAS
Detailed results of the QUADAS-2 assessment are provided in Appendix 4. All but ve studies (84%) showed risk of bias in at least one domain, o en because thresholds for index test posi vity had not been prede ned (n=21; 66%), or because more than 10% of the pa ents had been excluded because of missing reference standard results (n=14; 44%). In addi on,
72