Expert evaluation of different infection types in dairy cow quarters naturally infected with Staphylococcus aureus or...

The purpose of this study was to improve the diagnostic recommendations for Staphylococcus aureus and Streptococcus agalactiae control using bacterial culture (BC), polymerase chain reaction (PCR) and somatic cell count (SCC) as diagnostic methods. The study was carried out in three steps: firstly, diagnostic test patterns for naturally infected quarters with Staph. aureus (24 quarters) and Strep. agalactiae (16 quarters) were created by sampling the quarters each day for 21 days and analysing the daily quarter milk samples using BC, PCR and SCC. Secondly, 30 mastitis experts were asked to group and describe the diagnostic test patterns and to establish a diagnosis for each group. The experts' statements regarding the groups they established were subsequently examined using qualitative content analysis to assign "infection types" to the statements. Lastly, the test performance was estimated for BC, PCR and SCC using generalised logistic regression models with the interpreted statements as a reference for infection. The experts mainly identified the Staph. aureus quarter-patterns as persistent infections, while some had more dynamic patterns. Strep. agalactiae quarter-patterns mainly involved persistent infection, yet some appeared hard to diagnose and were assigned to almost all different infection types, while experts did not agree on the interpretation. Estimates of Se for detection of Staph. aureus infection were 95.9% [93.7; 97.3] for BC, 99.5% [98.3; 99.8] for PCR, and 96.1% [94.0; 97.5] for SCC. The corresponding Sp estimates were 74.5% [65.7; 81.7], 66% [57.2; 73.8] and 43.7% [36.2; 51.5] for BC, PCR and SCC, respectively. The Se estimates of BC and PCR for Strep. agalactiae infection were 100% [83.5; 100] and 99.9% [99.6; 100], respectively, whereas the Se of SCC detecting Strep. agalactiae infection was only 34.3% [26.4; 43.3]. This indicated that Strep. agalactiae-positive BC and PCR test results were more important than SCC results to the experts when diagnosing a quarter as infected. The Sp estimates of BC, PCR and SCC for Strep. agalactiae infection were 99% [72.8; 100], 97.7% [62.1; 99.9], and 65.7% [56.7; 73.7], respectively. We conclude that PCR and BC are highly sensitive in the detection of persistent and new infections as defined by the experts, although the Se was not always 100%. An accepted lower Sp suggests that experts place less emphasis on false-positive results. We recommend that efforts are made to develop consistent terminology to characterise intramammary infections over time so that the course of infection can be taken into account at diagnosis.