Comparison of phenotypic and genotypic diagnosis of acute human bocavirus 1 infection in children.

BACKGROUND: Diagnosis of human bocavirus 1 (HBoV1) has been based on qualitative PCRs detecting HBoV1 DNA or detection of HBoV1 mRNA.
OBJECTIVE: This study aims to assess whether a rapid and automated HBoV1 antigen test is suitable for diagnosis of acute HBoV1 infection. STUDY
DESIGN: HBoV1 antigen detection has been compared with quantitative HBoV1 DNA PCR and HBoV1 mRNA RT-PCR. RESULTS AND
CONCLUSION: We conclude that HBoV1 antigen detection has higher clinical specificity and positive predictive value than HBoV1 DNA qualitative PCRs, yet a lower sensitivity than HBoV1 mRNA detection. Additionally, HBoV1 antigen detection is beneficial in its rapidity and availability as a point-of-care test.

Background

Human bocavirus 1 (HBoV1) is a common cause of respiratory tract infections (RTI) in young children and causes serious lower RTI in children younger than six years of age [[1], [2], [3], [4], [5]]. The diagnosis of HBoV1 infections has traditionally been based on qualitative PCRs detecting viral DNA. Recent studies have however shown that the clinical value of this approach is low due to persistence of HBoV1 DNA in the airways, leading to frequent detection of low concentrations in healthy children [[6], [7], [8]]. Clinical specificity of HBoV1 mRNA detection has been shown to be superior to that of HBoV1 DNA detection [6,9,10]. Detection of HBoV1 antigen is another diagnostic approach. An automated HBoV1 antigen test based on immunometric detection has recently been developed and has shown promise in preliminary studies [11]. Antigen detection is well suited for automation and point-of-care diagnostics, and may be beneficial to making an HBoV1 diagnosis by providing rapid and specific results.

Objectives

The aim of this study was to compare HBoV1 antigen detection with HBoV1 mRNA RT-PCR and quantitative HBoV1 DNA detection (qPCR) in nasopharyngeal swabs (NPS) from children with RTI.

Study design

NPS (N = 632) were collected from children admitted for RTI to the Pediatric Emergency Department of Turku University Hospital from November 2015 to October 2016. All samples were stored in mariPOC RTI sample buffer at −20 °C throughout the study period.

The HBoV1 antigen test is part of a completely automated system for multianalyte identification of respiratory viruses and bacteria (influenzavirus A, influenzavirus B, respiratory syncytial virus, human coronavirus OC43, human metapneumovirus, human bocavirus 1, parainfluenzavirus type 1, parainfluenzavirus type 2, parainfluenzavirus type 3, adenovirus, Streptococcus pneumonia) and has IVD-CE approval for diagnostic testing (mariPOC® respi test, ArcDia International Ltd, Turku, Finland). It is an immunoassay based on separation-free two-photon excitation fluorometry and targets the HBoV1 capsid protein VP3 [12]. The mariPOC test reports strong positive samples in 20 min while low-positive and negative results are confirmed in 2 h. For mariPOC, NPSs were diluted in 1.3 mL of mariPOC RTI sample buffer for rapid prospective on-site testing on a random-access immunoassay analyzer. The test was in routine use at the Pediatric Emergency Department of Turku University Hospital for rapid point-of-care diagnostics of RTI throughout the study period. According to local guidelines, the mariPOC antigen test is used for all children hospitalized with acute RTI. This Pediatric Emergency Department serves a population of 70 000 children or adolescents (<16 years of age) and receives patients mostly by referral from primary care clinics. The majority of children hospitalized with RTIs are younger than 5 years of age and have lower RTIs.

The Turku University Hospital approved the study with a waiver of informed consent by the Ethics Committee, as the study involved reanalysis of de-identified samples collected as part of routine care.

All samples were analyzed in duplicate with HBoV1 DNA qPCR and HBoV1 mRNA reverse-transcription (RT)-PCR, as described previously [10,13]. The cutoff for a high viral load in DNA qPCR was >106 copies/mL.

Correlations between mRNA, high DNA load and antigen were measured by the phi coefficient of correlation. Sensitivity, specificity, and positive predictive value (PPV) for the antigen test were calculated by using mRNA RT-PCR and DNA qPCR as references.

Results

HBoV1 DNA, a high DNA load, mRNA, or antigen were detected in 96 (15.2%), 19 (3%), 17 (2.7%), and 13 (2.1%) of the 632 NPA samples, respectively (Table 1 ). Out of the 96 samples positive for DNA, only 19.8% had a high DNA load, 17.7% were positive for mRNA, and 13.5% for antigen. Of the samples with a high DNA load, 89.5% were positive for mRNA and 68.4% for antigen, whereas none of the low DNA-load samples harbored mRNA or antigen.

Table 1

Results of the HBoV1 diagnostic tests in relation to detection of HBoV1 mRNA and HBoV1 DNA quantity (N = 632).

		No.	HBoV1 mRNA+No. (%)	HBoV1 mRNA –No. (%)
HBoV1 antigen	Pos	13	13 (100)	0
	Neg	619	4 (0.6)	615 (99.4)
HBoV1 DNA quantity	Higha	19	17 (89.5)	2 (10.5)
	Lowa/neg	613	0	613 (100)
HBoV1 DNA presence	Pos	96	17 (17.7)	79 (82.3)
	Neg	536	0	536 (100)
			HBoV1 DNA high load	HBoV1 DNA low load or negative
		No.	No. (%)	No. (%)
HBoV1 antigen	Pos	13	13 (100)	0
	Neg	619	6 (1.0)	613 (99.0)
HBoV1 mRNA	Pos	17	17 (100)	0
	Neg	615	2 (0.3)	613 (99.7)

High HBoV1 DNA load: ≥106 copies/mL NPA; low load: <106 copies/mL NPA.

Positive HBoV1 antigen test results were available in 20 min for nine of the 13 samples, and the four remaining results were available in two hours.

Presence of HBoV1 antigen correlated well with both mRNA and high DNA load. The phi coefficient for the correlation between mRNA and antigen was 0.87, whereas it was 0.94 for mRNA and high DNA load.

Specificity and PPV of the HBoV1 antigen test were excellent, while the sensitivity was lower. With mRNA RT-PCR as reference, specificity, sensitivity, and PPV were 100%, 76.5%, and 100%, respectively. With high DNA load by qPCR as reference, the figures were 100%, 68.4%, and 100% (Table 2 ).

Table 2

Performances of the HBoV1 diagnostic tests, with mRNA RT-PCR or DNA qPCR as reference markers.

Diagnostic marker vs. reference	Specificity% (95 % CI)	Sensitivity% (95 % CI)	PPV% (95 % CI)
Antigen vs. mRNA	100 (99.4 – 100)	76.5 (52.7 – 90.4)	100 (77.2 – 100)
Antigen vs. DNA quantitya	100 (99.4 – 100)	68.4 (46.0 – 84.6)	100 (77.2 – 100)
DNA quantitya vs. mRNA	99.7 (98.8 – 99.9)	100 (81.6 – 100)	89.5 (68.6 – 97.1)
mRNA vs. DNA quantitya	100 (99.4 – 100)	89.5 (68.6 – 97.1)	100 (81.6 – 100)
DNA presence vs. mRNA	87.2 (84.3 – 89.6)	100 (81.6 – 100)	17.7 (11.4 – 26.5)

High HBoV1 DNA load: ≥106 copies/mL NPA. 95% CI, Confidence Interval (Wilson Score Method); PPV, positive predictive value.

Discussion

The widely used qualitative HBoV1 DNA PCRs yield results of low clinical specificity and very low PPV [1,2,6,7,14]. We confirmed this. With mRNA RT-PCR as a reference, the qualitative DNA PCR specificity and PPV were 87.2%, and 17.7%, respectively (Table 2).

Serological tests based on detection of HBoV1-specific IgM and increases of IgG in paired serum samples are well-established reference methods in clinical studies [1,2,6,14]. However, sera were not obtained from the children included in this study. Instead, we used mRNA detection and high DNA load as the basis for calculating specificities, sensitivities and PPVs. mRNA is a marker of actively transcribing virus during an ongoing infection, and a rapidly declining DNA load is the normal outcome after an acute infection. The HBoV1 mRNA test has performed well in studies, providing support for its use as reference [6,7,9,10]. Furthermore, a high DNA load (>104 or >106 copies/mL) in nasopharyngeal secretions has been shown to be a better diagnostic criterion than mere DNA detection by endpoint PCR [8,[14], [15], [16]]. The two tests together should therefore form a good basis for evaluating the antigen test.

Detection rates of 15% for HBoV1 DNA, 2.7% for mRNA and 3.0% for high DNA load (>106 copies/mL) in children with RTI are in agreement with previous reports in the literature [7,9,10,14,15]. Taken together, the results suggest that only about a fifth of the children with HBoV1 DNA in the nasopharynx has an acute HBoV1 infection.

In our study, the samples had been stored at −20 °C for up to two years, therefore some degradation of RNA cannot be excluded. The storage temperature recommended for RNA is usually −70 °C. The sensitivity of the HBoV1 antigen test could have been overestimated if RNA degradation had occurred. However, the mRNA positivity rate reflected that of the high DNA load, indicating that this effect was small. Furthermore, with an amplicon of 242 bp, the probability of degradation within the target sequence is low, explaining the limited effect of RNA degradation on the results.

Among adults with RTI, HBoV1 DNA is rarely detected. Sample sizes in studies on adults have been too small to conclude the clinical significance of HBoV1-DNA positivity [1,2].

We conclude that HBoV1 antigen detection is excellent for diagnosing acute HBoV1 infections in children. Compared with two independent reference assays, this antigen test shows substantially higher clinical specificity and PPV than the commonly used HBoV1 qualitative DNA PCRs. The sensitivity of the HBoV1 antigen test is lower than those of the PCRs, but it has the advantage of being a rapid test available for point-of-care use, with positive results usually available in 20 min, resulting in a faster diagnosis than traditional methods. This makes the test attractive for emergency room and outpatient use. A supplementary test, measuring HBoV1 DNA load, mRNA or specific antibodies, will be of value for higher sensitivity in an inpatient setting.

Ethical approval

Not required.

Funding

This work was supported by the Sigrid Jusélius Foundation, Finland [no grant number]; the Life and Health Medical Grant Association, Finland [no grant number]; the , China [201406170014]; the , Finland [no grant number]; the , Finland, under the project name ‘Get it done!’ [funding decision 534/14]; and BALTINFECT, within the European Union 7th framework programme [project no. 316275].

Competing interests: H.A, J.M.K., and J.O.K. are employed by ArcDia International Ltd. (Turku, Finland).

Competing interests: H.T. was previously employed by ArcDia International Ltd. (Turku, Finland).

CRediT authorship contribution statement

Nicola Isabelle Kols: Investigation, Writing - review & editing. Heli Aatola: Methodology, Validation. Ville Peltola: Investigation, Resources, Formal analysis. Man Xu: Investigation, Funding acquisition. Zaiga Nora-Krukle: Investigation, Resources, Formal analysis. Klaus Hedman: Investigation, Funding acquisition. Aurelija Zvirbliene: Investigation, Funding acquisition. Hanna Toivola: Methodology, Validation. Tytti Vuorinen: Investigation, Resources, Formal analysis. Juha M. Koskinen: Conceptualization, Methodology, Validation. Andrea H.L. Bruning: Methodology, Validation. Andreas Christensen: Investigation, Formal analysis, Writing - original draft, Writing - review & editing. Maria Söderlund-Venermo: Investigation, Formal analysis, Writing - review & editing, Funding acquisition. Janne O. Koskinen: Project administration, Conceptualization, Methodology, Validation, Writing - review & editing, Funding acquisition.

Declaration of Competing Interest

Competing interests: H.A, J.M.K., and J.O.K. are employed by ArcDia International Ltd. (Turku, Finland).

Competing interests: H.T. was previously employed by ArcDia International Ltd. (Turku, Finland).

N.I.K., V.P., M.X., Z.N.K., K.H., A.Z., T.V., A.H.L.B., A.C., and M.S.V.: no conflicts of interest.