The use of a quantitative assay to differentiate between Pneumocystis jirovecii colonisation and disease among hospitalised patients, South Africa 2015-2016
Date
2021
Authors
Yun, Jessica Ann Lai
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Abstract
Background: Pneumocystis jirovecii pneumonia (PJP) is responsible for infant and adult
mortality predominantly among those who are immunocompromised. Due to the use of
quantitative real-time polymerase chain reaction (qPCR) yielding a high sensitivity,
distinguishing between P. jirovecii colonisation and active clinical disease is challenging
because P. jirovecii DNA is detected regardless of its association with illness. By establishing
P. jirovecii qPCR DNA load cut-off value using beta-D-glucan (BDG) levels a biomarker of
fungal infection, P. jirovecii colonisation and disease could be differentiated. This could
prompt healthcare providers to initiate treatment in a timely manner as well as to assess the
“true” prevalence of PJP.
Aim: To establish P. jirovecii qPCR DNA load cut-off value for the differentiation between P.
jirovecii colonisation and disease among hospitalised patients admitted for severe respiratory
illness in South Africa, using BDG level as a biomarker of fungal infection.
Methodology: The study was a cross-sectional study design utilizing secondary data. Study
population included hospitalised patients enrolled in the pneumonia surveillance programme
from two of the five participating sentinel sites (KwaZulu-Natal and North West provinces)
from 2015-2016. A univariable and multivariable random effect (on surveillance hospital)
logistic regression was used to evaluate the demographic and clinical factors associated with
patients with and without available BDG level results. To determine the optimal P. jirovecii
DNA load cut-off value of the qPCR, using the BDG as the gold standard for fungal infection,
a receiver operating characteristic (ROC) analysis was conducted. Descriptive analysis to
describe the demographic, clinical and epidemiological characteristics of those who were P.
jirovecii diseased, colonised and negative was conducted.
Results: The total study population was 1542 patients enrolled in the pneumonia surveillance
programme who had sputum samples tested for P. jirovecii and had available qPCR results.
HIV was the most prevalent underlying condition with 70.43% (131/186) of patients living
with HIV had BDG results as compared to those without BDG results (51.48%, 680/1321). A
qPCR load cut-off value of 550 copies/5 μL was established which correctly classified 75% of
patients with PJP. This yielded a high specificity of 91.67% and a moderate sensitivity of
57.14%, respectively. The positive and negative predictive values were 100.0% and 81.6%,
respectively. Overall, the PJP prevalence was 5.12% and the prevalence of P. jirovecii
colonisation was 11.28% from 2015 to 2016. On multivariable analysis adjusting for
surveillance hospital clustering, patients who were infected with P. jirovecii compared to those
who were colonised with P. jirovecii were more likely to be female (Adjusted relative risk ratio
(aRRR) 1.91; 95% CI:1.04-3.49; p=0.04), to be living with HIV (aRRR 4.30; 95% CI:1.82-
10.16; p<0.01) and receive cotrimoxazole at time of hospital admission (aRRR 5.19; 95%
CI:2.24-12.06; p<0.001). Interestingly, patients who were negative for P. jirovecii compared
to those who were colonised with P. jirovecii were more likely to present with fever (aRRR,
1.44; 95% CI:1.01-2.05; p=0.04). Furthermore, patients infected with P. jirovecii compared to
those who were colonised with P. jirovecii were at an increased risk of in-hospital death (aRRR
2.63; 95% CI:0.98-7.05; p=0.05); therefore, indicating the severity of PJP.
Conclusions & recommendation: The cut-off value cannot be universally applied but rather
it may only be used in pilot studies with a similar population that consists of adults who are
living with HIV. Irrespective of not establishing a true gold standard for diagnostic tests to
diagnose PJP, the established cut-off value could be used in conjunction with clinical
presentation as well as to ruling in patients who have PJP (ruling diagnoses). Although the PCR
assays are considered to yield a higher sensitivity, more intensive specimen collection is needed
to estimate the true burden of PJP in South Africa
Description
A research report submitted in partial fulfilment of the requirements for the degree of Master of Science in Epidemiology to the Faculty of Health Sciences, School of Public Health, University of the Witwatersrand, Johannesburg, 2021