PCR versus serology for diagnosing Mycoplasma pneumoniae infection: a systematic review & meta-analysis

Abstract

Background & objectives: Diagnosis for Mycoplasma pneumoniae usually relies on serological tests. PCR technology has some advantages but also limitations. The optimal selection for these tests still needs discussion. This paper reviews the overall diagnostic accuracy of PCR versus serological assays for diagnosis of M. pneumoniae infections and to identify factors associated with heterogeneity of results.

Methods: MEDLINE and Embase databases were searched. Articles meeting the selection criteria were retrieved for data collection and analysis. Studies were assessed for methodological quality using QUADAS. Hierarchial summary receiver operating characteristic (HSROC) model was used to estimate summary ROC curve.

Results: Initial meta-analysis showed a summary estimate of sensitivity (SEN) 0.62 (95% CI, 0.45-0.76), and specificity (SPE) 0.96 (95% CI, 0.93-0.98). Subgroup analyses were performed to identify factors associated with heterogeneity. For different gene targets, reference standards, subjects (children or adults) and different PCR types, these aspects can generate results of heterogeneity. The 16s rDNA target and adult subjects and real-time PCR may have better test results for PCR.

Interpretation & conclusions: Commercial PCR tests generated consistent results with high specificity but a lower and more variable sensitivity. The findings suggest commercial PCR tests having superiorities in diagnosing M. pneumoniae infections but still cannot replace serology. PCR plus serology could be good screening tests for reliable and accurate diagnosis of M. pneumoniae.

Fig. 1

Study selection scheme.

Fig. 2

Forest plot (shows initial meta-analysis results included all 14 studies).

Fig. 3

Bivariate boxplot with most studies clustering within the median distribution and 4 outliers (NO. 1–14 represents studies in Table 1, X–axis: Logit specificity, Y–axis: Logit sensitivity)

Fig. 4

HSROC curve of all 14 studies.

Fig. 5

Subgroup analysis 1 and forest plot for subgroup analysis 1.

Fig. 6.

Subgroup analysis 2 and forest plot for subgroup analysis 2.

Fig. 7

Subgroup analysis 3 and forest plot for subgroup analysis 3.

Fig. 8

Subgroup analysis 4 and forest plot for subgroup analysis 4

Fig. 9

Deeks’ funnel plot [The statistically non-significant P-value (0.81) for the slope coefficient suggests symmetry in the data and a relative low likelihood of publication bias].