OBJECTIVETo evaluate the accuracy of partial- mouth periodontal examination (PMPE) protocols frequently used in epidemiological periodontal surveys.

METHODSArticles in English or Chinese published before Jan 31st 2014 were searched, which compared the results of PMPE protocols with those of gold-standard protocol, i.e.full-mouthmesialbuccal-midbuccal-distobuccal-mesiolingual-midlingual-distolingual (MB-B-DB-ML-L-DL) protocol.

RESULTSTwelve articles were included and nine that frequently used PMPE protocols were evaluated. All these protocols underestimated the prevalence scores. For prevalence of probing depth (PD) ≥ 4 mm, 6 mm and attachment loss (AL) ≥ 4 mm, 6 mm, smaller amount of underestimation was observed in community periodontal index of treatment needs (CPITN) teeth (-12.6%--3.5%), full-mouth MB-B-DB (-16.1%--3.5%), full-mouth MB-B-DL (-10.8%--6.1%) and half-mouth MB-B-DB-ML-L-DL (-23.6%--2.0%) protocols. For severity and extent estimates, half-mouth MB-B-DB-ML- L-DL provided lowest biased results (relative bias: -1.0%- 1.1% for severity and -6.7%-0.1% for extent). Full-and half-mouth MB-B-DB also performed well, with relative bias within ± 5.0% in most cases. CPITN overestimated the severity and extent of periodontal disease, the relative bias of which amounted to 42.3% and 38.1%, respectively.

CONCLUSIONSHalf-mouth MB-B-DB-ML-L-DL and full-mouth MB-B-DB protocols caused lower biased results in prevalence, severity and extent estimates of PD and AL.