Classification of pregnancy related non-specific low back pain and pelvic girdle pain: a systematic review

Abstract

Background: Pregnancy related non-specific low back pain (PNSLBP) and pelvic girdle pain (PGP) affect almost half of pregnant women, causing substantial disability for some. These diagnoses, which may represent several possible individual or co-occurring conditions, lack the specificity required to guide personalized treatment decisions. A reliable and valid classification is needed to direct effectiveness research and inform treatment strategies.Objectives: To describe the psychometric properties of pathoanatomic classifications of PNSLBP and PGP as well as to assess the risk of bias of supporting evidence.Methods: Two independent reviewers systematically searched PubMed, Embase, Cochrane and PEDro databases and conducted a manual search of reference list of eligible articles reporting office-based examination for classifying PNSLBP/PGP. QUADAS-2 tool was used to assess the risk of bias and applicability of reviewed studies.Results: Ten studies were eligible. Inter-rater reliability was 87% for PNSLBP and 84.6% for PGP. Major sources of potential bias included the lack of reference standard diagnostic tests, insufficient blinding, non-systematic order of index test application, and unclear participant selection procedures. Patient flow throughout studies and the interval between testing (flow and timing) showed the lowest risk of bias. The applicability of the reference standard had the highest concern, while participant selection and index test had the lowest.Conclusion: Although acceptable inter-rater reliability has been demonstrated, there is no current evidence to support the validity of PNSLBP and PGP classification systems. Future studies should consider including reference standard diagnostic tests, random order sequencing of index tests, rigorous selection criteria, and blinded assessors.