Invasive Group B Streptococcal Disease in South Africa: Importance of Surveillance Methodology.

Data on neonatal group B streptococcal (GBS) invasive disease burden are needed to refine prevention policies. Differences in surveillance methods and investigating for cases can lead to varying disease burden estimates. We compared the findings of laboratory-based passive surveillance for GBS disease across South Africa, and for one of the provinces compared this to a real-time, systematic, clinical surveillance in a population-defined region in Johannesburg, Soweto. Passive surveillance identified a total of 799 early-onset disease (EOD, <7 days age) and 818 LOD (late onset disease, 7-89 days age) cases nationwide. The passive surveillance provincial incidence varied for EOD (range 0.00 to 1.23/1000 live births), and was 0.03 to 1.04/1000 live births for LOD. The passive surveillance rates for Soweto, were not significantly different compared to those from the systematic surveillance (EOD 1.23 [95%CI 1.06-1.43] vs. 1.50 [95%CI 1.30-1.71], respectively, rate ratio 0.82 [95%CI 0.67-1.01]; LOD 1.04 [95% CI 0.90-1.23] vs. 1.22 [95%CI 1.05-1.42], rate ratio 0.85 [95% CI 0.68-1.07]). A review of the few cases missed in the passive system in Soweto, suggested that missing key identifiers, such as date of birth, resulted in their omission during the electronic data extraction process. Our analysis suggests that passive surveillance provides a modestly lower estimate of invasive GBS rates compared to real time sentinel-site systematic surveillance, however, this is unlikely to be the reason for the provincial variability in incidence of invasive GBS disease in South Africa. This, possibly reflects that invasive GBS disease goes undiagnosed due to issues related to access to healthcare, poor laboratory capacity and varying diagnostic procedures or empiric antibiotic treatment of neonates with suspected sepsis in the absence of attempting to making a microbiological diagnosis. An efficacious GBS vaccine for pregnant women, when available, could be used as a probe to better quantify the burden of invasive GBS disease in low-middle resourced settings such as ours. From our study passive systems are important to monitor trends over time as long as they are interpreted with caution; active systems give better detailed information and will have greater representivity when expanded to other surveillance sites.