Yunhao Liu1,2*, Jennifer B. Griffin3, Atis Muehlenbachs4, Stephen J. Rogerson5, Anya J. Bailis6, Rajni Sharma7, David J. Sullivan8, Antoinette K. Tshefu9, Sarah H. Landis10, Jean‑Marie M. Kabongo11, Steve M. Taylor12 and Steven R. Meshnick1
Abstract
Background: Placental histopathology has been considered the gold standard for diagnosis of malaria during pregnancy. However, in under-resourced areas placental tissue is often improperly fixed and processed; the resulting formalin pigment is difficult to distinguish from malaria pigment. This study examines two alternative diagnostic methods: polymerase chain reaction (PCR) and a novel immunohistochemistry (IHC)-based method using an antibody against histidine-rich protein 2 (HRP2).
Methods: Placental histopathology from 151 pregnant women in Kinshasa was assessed by two blinded microscopists and compared with peripheral blood PCR and IHC for HRP2. The Cohen’s kappa coefficients were calculated to assess the test agreement. The sensitivity and specificity of individual tests were calculated using PCR or IHC as the reference standard as well as latent class analysis (LCA).
Results: PCR and IHC correlated fairly well. The correlation between the two blinded microscopists was poor, as there was widespread formalin pigment. Using LCA, all of the tests had high specificities. The most sensitive test was IHC (67.7 %), with PCR as second-best (56.1 %).
Conclusions: PCR and/or IHC are suitable diagnostics when the presence of formalin pigment substantially compromises placental histopathology.
Keywords: Malaria, Histopathology, PCR, IHC, Latent class analysis, Molecular epidemiology
Malaria Journal
Liu et al. Malar J (2016) 15:272
