A multicenter comparative study of two classification systems for radial polydactyly

Abstract

Background: The aim of this study was to compare type occurrence and reliability of the Wassel and Rotterdam classifications for radial polydactyly.

Methods: The authors classified a large population of radial polydactyly patients from two European clinics using both classification systems, and compared the incidences of the different types to a population derived from a systematic literature review. The authors further assessed intraobserver and interobserver reliability of both classification systems in a test-retest design with seven observers, using kappa statistics.

Results: Forty percent of the 520 cases with available radiographs could not be classified using the Wassel classification, whereas all cases could be classified using the Rotterdam classification. All unclassifiable cases had aberrant components; the majority were of the triphalangeal (63 percent), deviating (43 percent), or hypoplastic (39 percent) kind. Types III, IV, and VI occurred more often when using the Rotterdam classification. Intraobserver and interobserver reliability was comparable for both classification systems (κ=0.87 versus κ=0.83, and κ=0.65 versus κ=0.70). Types II and IV had the lowest reliability in both the Wassel and Rotterdam classifications (κ=0.30 to 0.59). Aberrant components indicating deviation and hypoplasia had the lowest reliability in the Rotterdam classification (κ=0.19 to 0.45).

Conclusions: The Rotterdam classification has broader classification possibilities and similar intraobserver and interobserver reliability compared with the Wassel classification. Although it is more complex and the aberrant components should be more strictly defined to increase its clinical relevance, we recommend using the Rotterdam classification.

Clinical question/level of evidence: Diagnostic, I.