Mannocci 2002.

Methods	Randomised controlled trial, setting and start date unspecified, 3‐year duration. The setting was confirmed following telephone communication with the principal investigator as a single private practice and that the study participants were enrolled between 1997 and 1998.
Participants	Inclusion criteria: Single maxillary or mandibular premolar requiring endodontic treatment and crown build up Class II carious lesions without previous endodontic treatment and with preserved cusp structure In occlusal function after restoration Not used as abutment for fixed or removable partial dentures Loss of periodontal attachment (< 40%), assessed using the gingival index score (Loe 1963) Exclusion criteria: Spontaneous gingival bleeding (gingival index score = 3) Randomised: 117 (54 male, 63 female). Age range 35 to 55 years (mean 48 years) Teeth: maxillary first premolars (24), maxillary second premolars (57), first (3) and second (33) mandibular premolars Withdrawals/losses to follow‐up: Losses at specific recall time points: 1 year recall 5/60 (composite only) 2 year recall 12/60 (composite only) 3 year recall 10/60 (composite only) and 3/57 (composite and crown)
Interventions	Intervention: Group 1: orthograde endodontic treatment including a carbon fibre post, restored with adhesive techniques and composite (60 teeth) Comparison: Group 2: orthograde endodontic treatment including a carbon fibre post, restored with adhesive techniques and composite, and covered with full‐coverage metal‐ceramic crown (57 teeth) Composite restoration and core and crown build‐up identical for both groups (Light polymerising composite Z100, 3M). Crown preparation, impression, temporising and cementation according to standard clinical techniques All restorations carried out by a single operator Routine oral hygiene instruction from a dental hygienist
Outcomes	Clinical, radiographic and photographic evaluation by two calibrated examiners (not investigators); immediately before and after restoration and at 1, 2, 3 year recall. Outcomes: (as reported) Failure i.e. root fracture, post fracture, post decementation* Clinical and/or radiographic evidence of a marginal gap between tooth and restoration* Clinical evidence of secondary caries contiguous with the margins of the restoration* Clinical assessment: margins of the restoration with explorer and loops with fibreoptic illumination Photographic assessment: colour slides of the restorations with standard film Periapical radiographic assessment: standard paralleling technique Definition of failure: marginal gap between tooth and restoration determined by explorer radiographic evidence of a marginal gap between tooth and restoration secondary caries at the restoration margin, after the removal of the restoration root fracture noted after tooth extraction post fracture separation into two post parts post decementation separation of the post‐core (crown) restoration from tooth structure *Denotes outcomes prespecified in this review.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "selected subjects were randomly assigned to 1 of the following 2 experimental groups by tossing a coin" Page 298 Comment: probably done
Allocation concealment (selection bias)	Unclear risk	The method used to conceal the allocation sequence, that is to determine whether intervention allocations could have been foreseen in advance of, or during enrolment, was not reported Comment: insufficient information to permit a clear judgement. Telephone contact with principal investigator: no further information provided to enable any change to this assessment
Blinding of participants and personnel (performance bias) All outcomes	High risk	The nature of the interventions makes blinding of the participants or the trial investigators difficult, however no attempt was made to avoid performance bias.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Quote: "Evaluation of success or failure was performed by 2 examiners other than the operator" Page 300. Comment: the outcome assessors were not the care providers. They were calibrated and inter‐rater agreement for the specified outcomes was > 90%. Comment: As they were not blinded this presents a high risk of assessment bias.
Incomplete outcome data (attrition bias) All outcomes	High risk	Participants not available at the 2 and 3 year recall were reported but no reasons given Losses to follow‐up were not balanced across groups; and were large (> 20%) and not consistent at both recall time points in the composite‐only group Comment: although it was unclear from the report if these data were missing at random, this domain was judged as at a high risk of bias.
Selective reporting (reporting bias)	Low risk	Although data were sparse, all expected and prespecified outcomes appear to have been reported
Other bias	Unclear risk	Baseline measurements of caries, periodontal and endodontic outcomes were not reported. This presents an unclear risk of other bias (e.g. severity of illness bias)