AIMS (I): To investigate the effects of study characteristics on the reported success rates of secondary root canal treatment (2 degrees RCT or root canal retreatment); and (ii) to investigate the effects of clinical factors on the success of 2 degrees RCT. Longitudinal human clinical studies investigating outcome of 2 degrees RCT which were published upto the end of 2006 were identified electronically (MEDLINE and Cochrane database 1966-2006 Dec, week 4). Four journals (Dental Traumatology, International Endodontic Journal, Journal of Endodontics, Oral Surgery Oral Medicine Oral Pathology Endodontics Radiology), bibliographies of all relevant papers and review articles were hand-searched. Two reviewers (Y-LN, KG) independently assessed and selected the studies based on specified inclusion criteria and extracted the data onto a pre-designed proforma, independently. The criteria were: (i) Clinical studies on 2 degrees RCT; (ii) Stratified analyses available for 2 degrees RCT where 1 degrees RCT data included; (iii) Sample size given and larger than 10; (iv) At least 6-month post-operative review; (v) Success based on clinical and/or radiographic criteria (strict = absence of apical radiolucency; loose = reduction in size of radiolucency); and (vi) Overall success rate given or could be calculated from the raw data. Three strands of evidence or analyses were used to triangulate a consensus view. The reported findings from individual studies, including those excluded for quantitative analysis, were utilized for the intuitive synthesis which constituted the first strand of evidence. Secondly, the pooled weighted success rates by each study characteristic and potential prognostic factor were estimated using the random effect model. Thirdly, the effects of study characteristics and prognostic factors (expressed as odds ratios) on success rates were estimated using fixed and random effects meta-analysis with DerSimonean and Laird's methods. Meta-regression models were used to explore potential sources of statistical heterogeneity. Study characteristics considered in the meta-regression analyses were: decade of publication, study-specific criteria for success (radiographic, combined radiographic & clinical), unit of outcome measure (tooth, root), duration after treatment when assessing success ('at least 4 years' or '<4 years'), geographic location of the study (North American, Scandinavian, other countries), and qualification of the operator (undergraduate students, postgraduate students, general dental practitioners, specialist or mixed group). Of the 40 papers identified, 17 studies published between 1961 and 2005 were included; none were published in 2006. The majority of studies were retrospective (n = 12) and only five prospective. The pooled weighted success rate of 2 degrees RCT judged by complete healing was 76.7% (95% CI 73.6%, 89.6%) and by incomplete healing, 77.2% (95% CI 61.1%, 88.1%). The success rates by 'decade of publication' and 'geographic location of study' were not significantly different at the 5% level. Eighteen clinical factors had been investigated in various combinations in previous studies. The most frequently and thoroughly investigated were 'periapical status' (n = 13), 'size of lesion' (n = 7), and 'apical extent of RF' (n = 5) which were found to be significant prognostic factors. The effect of different aspects of primary treatment history and re-treatment procedures has been poorly tested. The pooled estimated success rate of secondary root canal treatment was 77%. The presence of pre-operative periapical lesion, apical extent of root filling and quality of coronal restoration proved significant prognostic factors with concurrence between all three strands of evidence whilst the effects of 1 degrees RCT history and 2 degrees RCT protocol have been poorly investigated.