The choice of study designs of diagnostic accuracy using Borrelia specific IgG and IgM antibodies for the diagnosis of Lyme borreliosis

BACKGROUND: Laboratory diagnosis of Lyme borreliosis (LB) is used in a variety of clinical settings where a range of other diagnoses may be considered. Therefore it is essential that diagnostic accuracy studies and litterature reviews consider information from different types of studies and choice of sample groups. The quality of patient selection is important to minimize risk of misclassification. This narrative review was inspired by systematic reviews where nearly all studies on the diagnostic accuracy for LB tests were determined as "biased" and having "low quality" based solely on study design considerations - not the clinical relevance.

OBJECTIVES: To propose flexible design and interpretation of studies used to assess diagnostic accuracy in clinical microbiology.

SOURCES: Criteria for rating the quality of studies were discussed among the ESCMID study group for LB (ESGBOR). The literature was searched for similar methodological discussions.

CONTENT: Knowledge of antibody reactivity in the background population across various clinical patient groups with and without infection should consider variations in clinical presentation and duration of disease. Case-control studies are the most frequently used design and were judged particularly instrumental in assessing serologic testing. However, clinical and epidemiological studies not specifically intended for diagnostic accuracy may also contribute estimates of sensitivity and specificity. Systematic reviews should focus on the application of the diagnostic assay for the individual patient in various clinical settings, rather than seeking an un-biased average. Different LB sample groups and controls for test panels are discussed.

IMPLICATIONS: Case-control (two-gate design) studies, case series, and seroprevalence studies representing the range of LB in different populations are necessary to assess the diagnostic accuracy of serological tests for LB. A broader range of studies should be considered for inclusion in systematic reviews of diagnostic accuracy.