Beruflich Dokumente
Kultur Dokumente
DOI: 10.1097/INF.0000000000001630
Tomoyuki Fujii, MD, PhD1, Akira Oka, MD, PhD2, Ichiro Morioka, MD, PhD3, Hiroyuki
Moriuchi, MD, PhD4, Shin Koyano, MD, PhD5, Hideto Yamada, MD, PhD6, Shigeru Saito,
MD, PhD7, Hiroshi Sameshima, MD, PhD8, Takeshi Nagamatsu, MD, PhD1, Shinya Tsuchida,
D
MD, PhD2, and Naoki Inoue, PhD9*, for the Japanese Congenital Cytomegalovirus Study
Group #
TE
Departments of 1 Obstetrics and Gynecology and 2 Paediatrics, The University of Tokyo,
#
Members of the Japanese Congenital Cytomegalovirus Study Group who contributed to this
Abbreviated title: Risk-Based Newborn Screening and In Vitro Diagnostics for Congenital
CMV
Running head: Screening Approach & In Vitro Diagnostics for Congenital CMV
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
COI and Funding support: None of the authors has any conflict of interest to declare. This
work was supported by a Grant for the Research on Child Development and Diseases
(15gk0110003h0103) from the Agency for Medical Research and Development (AMED), the
Japanese Government.
ACKNOWLEDGMENTS
We would like to thank our medical and nursing colleagues as well as the newborns and their
D
parents who agreed to take part in this study. We would like to also thank the following
members of the Japanese Congenital Cytomegalovirus Study Group for their contributions: H.
TE
Samejima (Samejima Bonding Maternity Clinic, Saitama), S. Yamaguchi (Yamaguchi
Univ.), M. Itoh (Univ. Toyama), Y. Kawagoe (Univ. Miyazaki). All individuals listed are part
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
ABSTRACT
authorities, we evaluated the clinical risks and performance of diagnostic assays developed by
D
cCMV infection (n=575) were screened for the presence of CMV DNA in urine specimens
collected onto filter paper placed in their diapers using the PCR-based assay reported
TE
previously. Liquid urine specimens were obtained from all of 20 CMV-positive newborns
and 107 of the CMV-negative newborns identified in the screening. We used these 127
specimens, as well as 12 from cCMV cases identified in a previous study and 41 from healthy
newborns, to compare the performance of two commercial assays and one in-house assay.
EP
Results: The risk-based screening allowed the identification of cCMV cases at least 10-fold
more efficiently than our previous universal screening, although there appears to be a limit to
during pregnancy was found frequently in mothers of cCMV newborns, CMV-IgM alone is
C
not an effective diagnostic marker. The urine-filter based assay and the three diagnostic
Conclusions: Although risk-based and universal newborn screening strategies for cCMV
infection each have their respective advantages and disadvantages, urine-filter based assay
A
newborn screening
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
INTRODUCTION
countries and is associated with significant clinical consequences, not only at birth but also
developmental delays 1. Our retrospective studies demonstrated that fetal cCMV infection
was associated with in 12-15% of cases of severe SNHL and 25% of cases of developmental
delay of unknown cause, with one half of the sequelae being late-onset 2,3. Early
D
identification of cCMV infection may lead to new treatment options with antiviral agents 4
and early intervention in infants with SNHL to aid language development 5. Therefore, it is
TE
important to establish newborn cCMV screening programs.
Traditionally, the diagnosis of cCMV infection has been performed by CMV isolation
from urine specimens collected within 3 weeks of birth. As the virus isolation is laborious,
EP
PCR detection of CMV DNA in urine specimens is used as the gold standard assay. However,
(qPCR) assay using urine specimens collected on filter discs (urine-filter) that can be used
directly as a PCR template without additional purification or elution steps 6,7, and
C
demonstrated the feasibility of our assay for relatively large-scale screening programs 810. To
risk factors for cCMV infection, and evaluated the significance of each risk factor in the
reagents developed by two commercial companies using urine specimens obtained from the
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
MATERIALS and METHODS
Study design
As shown in Fig. 1, this study consisted of two parts. Part A sought to establish the
significance and limitations of newborn CMV screening based on clinical manifestations and
features indicative of cCMV at five study sites (Tokyo, Kobe, Nagasaki, Toyama, and
Miyazaki) from Oct. 1, 2014 to Nov. 30, 2015. Five hundred and seventy-five newborns who
D
exhibited at least one of the clinical manifestations and serological indications listed in Table
2, including fetal and neonatal symptoms as well as maternal features of possible CMV
TE
infection during pregnancy, were enrolled for CMV screening using the urine-filter based
qPCR assay.
companies (Shino-Test Corporation and QIAGEN Japan), and the results of the study were
EP
intended to be used in the application of these two commercial assays as In-Vitro Diagnostics
negative specimens for enrollment of the Part B study were determined based on suggestions
by the regulatory authorities ahead of the study. Urine specimens in a liquid form were
C
collected both from all CMV-positive newborns (n=20) and a portion of the CMV-negative
newborns identified during screening at two of the study sites (n=107) as well as from a
C
number of healthy newborns (n=41). The selection of the CMV-negative newborns was done
randomly from the two study sites so as to make the numbers almost equal between the sites
A
(n=53 in the Tokyo area, and n=54 in the Kobe area). For the selection, the Excel RAND
function was used. Collection and use of the human subject materials was approved by the
Ethical Committee on Human Subjects of each participating institute and of the two
companies. Informed consent from the parents of the respective newborns was obtained at
enrollment for collection of liquid urine specimens for evaluation of the assays developed by
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
the two companies. All specimens collected for screening and evaluation of the assays were
anonymized and only the attending physicians can link the assay results to their patients. In
addition, the original code numbers were randomized and recoded to ensure that the assay
Clinical evaluation
Clinical information for the newborns, including birth weight, gestation age, clinical
D
manifestations, and abnormal laboratory findings, were extracted from their medical records.
Audiological testing was performed using auditory brainstem responses and/or auditory
TE
steady-state responses, and brain imaging was performed by echography, computed
tomography and/or magnetic resonance imaging. The maternal medical records were
examined for abnormalities during pregnancy. Serologic tests for CMV-specific IgG and IgM
for the pregnant mothers were performed at a commercial laboratory (SRL Inc., Japan) using
EP
EIA kits produced by DENKA SEIKEN Co. Ltd (Niigata, Japan). The CMV-IgM kit is based
Urine samples were collected from newborns onto filter paper within 1 week after birth, and
C
the presence of CMV-DNA was assessed as described previously 7 with minor modifications,
including the use of No.4A filter paper (ADVANTEC, Tokyo, Japan) instead of FTA-Elute
C
Card and removal of the procedure for washing the filter discs with water prior to qPCR
reaction. These modifications allowed for reduction in cost, potential for minor skin
A
irritations (change in filter paper) and labor (removal of the washing procedure). qPCR
assays with filter paper spiked with CMV-positive or -negative urine specimens validated that
DNA-based assays
The in-house real-time qPCR assay using liquid urine samples was performed as described
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
previously 2. This was one of the assays used for the collaborative study to evaluate the
proposed 1st WHO International Standard for CMV. The assay produced by QIAGEN was
also based on a qPCR assay, and used reagents included in the artus CMV RGQ MDx kit,
which is approved for monitoring CMV viral loads in transplant patients in the US. Two
hundred copies per reaction was used as cut-off for the positive detection of CMV DNA in
this study. The assay developed by Shino-Test Corp. was based on the asymmetric isothermal
D
amplification method, known as the Smart Amplification Process (SMAP), originally
developed by Riken 11, and its conditions were similar to those described previously for CMV
detection 12 apart from the use of a different set of primers (IVD application submission no.
DNA.
Statistical analyses TE
5122858000105). Incubation for 60 min was used as a cut-off point for the detection of CMV
EP
Statistical significance in the prevalence of cCMV between study areas was evaluated using
the Chi-square test. Correlations between viral copy numbers obtained by two different
RESULTS
C
Screening of 575 newborns for cCMV infection based on the clinical manifestations listed in
C
Table 2 as risk factors identified 20 newborns with cCMV infection across the 5 study sites.
As 3 of the 5 study sites participated in the previous study based on the universal screening of
A
newborns for cCMV infection 8, we compared the detection frequencies of cCMV infection
between the risk-based and universal screening approaches. As shown in Table 1, the risk-
based screening provided an at least 10-fold increase in the detection of newborns with
cCMV. Although there was ~2-fold difference in the prevalence of cCMV infection between
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Association between clinical manifestations and cCMV infection
We obtained clinical information for the 20 CMV-positive newborns as well as for 107 CMV-
negative infants, which corresponded to about one-fourth of the total population of CMV-
negative newborns identified during the urine-filter-based screening at the Tokyo and Kobe
study sites (Table 2, manifestations of each patient are shown in on-line only Table,
D
symptomatic cCMV, defined as those with typical clinical manifestations (hearing
TE
with the abnormalities in brain images, was 60%, with 14 of the 20 CMV-positive newborns
exhibiting at least one of the clinical manifestations (average 4.1 manifestations; excluding
maternal indications). When compared with the clinical manifestations of the 107 CMV-
more than one-third of the cCMV cases exhibited intrauterine growth restriction, low birth
weight, and jaundice that required treatment, these manifestations were also identified in the
C
mothers across the 5 study sites delivered cCMV newborns (see the equation in the Footnote 1
A
at the end of text). Among the 50 CMV-negative and 10 CMV-positive newborns from
CMV-IgM positive mothers, 9 and 6 of them, respectively, had additional indications for
cCMV infection (Table 3). Due to the limited number of specimens, it is hard to predict any
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Comparison of the DNA-based assays for cCMV infection
Urine specimens in liquid form were collected from 180 newborns (Fig. 1) and used for the
specimens from 41 healthy newborns and from 107 newborns with clinical manifestations
who were found to be CMV-negative in the screening assay were all CMV-negative, and
those from the 20 newborns with clinical manifestations who were found to be CMV-positive
D
in the screening assay and from 12 CMV-positive newborns identified in the previous study
were all CMV-positive in all of the three assays; i.e., the qPCR assay developed by QIAGEN,
TE
the SMAP-based assay by Shino-Test, and our in-house qPCR assay.
This study did not focus on the determination of the quantitative correlation among the
assays as the DNA-based assays developed by the two companies are intended to clinically
diagnose the presence or absence of congenital CMV infection. However, to identify any
EP
potential limitations in each assay, quantitative measurements by the three assays were
compared. As shown in Fig. 2A, there was a good correlation for the in-house and QIAGEN
qPCR assays (r=0.93). In contrast, the in-house qPCR and the Shino-Test SMAP assay
showed a weaker correlation (r=0.66) (Fig. 2B). There was one obvious outlier (case 71),
C
marked with an arrow, and exclusion of this outlier resulted in a better correlation (r=0.86).
As the target sequence region for the SMAP assay of the outlier contained one base alteration
C
in the detection primer sequence, it is plausible that the alteration decreased the detection
efficiency of the SMAP reaction. Further, as the sequences for the primers in the target region
A
of the SMAP assay were well conserved among 19 Japanese clinical isolates and 4 laboratory
strains 13, the base alteration identified in this study appears to be very rare.
DISCUSSION
There were three major outcomes of our study. First, CMV screening based on clinical
manifestations was able to improve the efficiency of CMV screening by at least 10-fold,
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
although one-third of cCMV cases may be overlooked as discussed below. Second, as the
DNA-based CMV assays developed by two companies and our in-house qPCR assays
provided consistent results, both the commercial assays are considered to be reliable for use as
an In Vitro diagnostics for the determination of cCMV infection in newborns. Finally, the
urine-filter-based assay is a reliable tool for newborn CMV screening, with the 107 CMV-
negative and 20 CMV-positive results obtained in the screening tests of newborns with
D
clinical manifestations being completely consistent with those obtained by the 3 other
confirmatory assays.
TE
We previously reported that i) typical clinical manifestations, defined as any of
jaundice, were observed in 22.7% (15/66) of the infected newborns, while ii) abnormalities in
brain images, including intracranial calcifications, ventricular dilation, and abnormal lesions,
EP
were identified in 10 out of 58 cases, so that iii) the proportion of symptomatic cases with
typical clinical manifestations and/or abnormalities in brain images was 30.3 % (20/66) 8.
The proportion of symptomatic cCMV cases in this study was two-fold (60%) that in the
previous study, while the risk-based selection allowed the at least 10-fold more efficient
C
identification of cCMV cases. These findings suggest that a risk-based screening approach is
feasible in situation in which there are insufficient resources for the introduction of universal
C
screening. The ~2-fold difference in the prevalence of cCMV infections between the study
sites, although it was not statistically significant, could be due to the difference in the ratio of
A
type-1 study sites (primary obstetric clinics and municipal hospitals) against type-2 sites
(university-associated and government hospitals) between the study areas as discussed in our
previous publication 8.
However, there is an obvious drawback in the risk-based approach; that is, the potential
to overlook asymptomatic cases. Among the 66 cCMV cases in the previous study, 10 cases
10
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
presented with one or two non-specific manifestations, including low birth weight (n=8),
jaundice (n=2), and mild liver function abnormality (n=2), in addition to the 20 cases with
symptomatic cCMV. Thus, 45.5% (30/66) of cCMV cases could be identified by clinical
manifestations observable in the fetuses and newborns. This study found that the detection of
CMV IgM or seroconversion in the mothers, in the absence of any other clinical
manifestations, could identify 30% (6/20) of cCMV cases. Taken together, a simple
D
calculation based on this information shows that the screening approach based on a
combination of clinical manifestations and serology may be able to identify about two-thirds
TE
(see the equation in the Footnote 2 at the end of text) of cCMV cases; in other words, such an
approach may overlook one-third of cCMV cases, particularly asymptomatic cases or those
The clinical manifestations used for the screening were selected based on previously
EP
reported observational studies 1419. Among them, thrombocytopenia, petechiae and
hydrocephalus were the most useful predictors for cCMV infection. SNHL is the most
common feature of cCMV infection, and its early detection and subsequent therapeutic
intervention using antiviral agents remain important clinical issues. Although universal
C
thirds of all newborns in Japan. In addition, late-onset SNHL is well known to be associated
C
with the clinical course of cCMV, and the DNA-based diagnosis of cCMV infection during
neonatal period is crucial. Examination of brain images during the early phase of infancy are
A
not routinely indicated in the absence of other clinical manifestations. Actually, all of the 7
newborns with SNHL and 6 with brain image abnormalities had at least 2 additional
manifestations. Therefore, most of newborns with SNHL or brain image abnormalities could
be selected and subjected to the screening test based on apparent clinical signs, leading the
correct diagnosis. The finding of a small positive predictive value for CMV-IgM positivity in
11
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
pregnant mothers for cCMV infection is consistent with the results of our previous study 20.
provide a useful way to increase the detection of cCMV cases in the population. The
introduction and standardization of serologic assays for CMV-IgG avidity is likely to increase
the efficacy of CMV serology. Nevertheless, it is well documented that cCMV infection can
be caused not only primary infection but also reinfection and reactivation 2124.
D
As CMV-specific antigenemia assay and IgM serology using umbilical or neonatal blood
specimens can detect only a portion of newborns with cCMV infection 8, and since CMV
TE
DNA copy numbers in urine specimens are significantly higher than those in blood specimens
from newborns with cCMV infection 6, DNA-based assays using urine specimens are
expected to perform better than any other assays. Although saliva specimens contain the
amount of CMV DNA similar to urine specimens, there is a potential for contamination with
EP
CMV DNA from breast milk 25. The performance of all three assays in the diagnosis of
cCMV was adequate as expected. However, the quantitative analysis of the performance of
the assays revealed that some urine specimens had lower copy numbers of CMV DNA. As
those specimens were mainly from the previous study and had shown high copy numbers
C
immediately after collection, it is considered likely that the freeze-thawing of the urine
materials a couple of times for use in other studies degraded the DNA.
C
based assay followed by either of the two commercial assays as an In Vitro diagnostics
A
affords appropriate identification of cCMV infection in newborns, although the risk-based and
12
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
FOOTNOTES
2) 20 cCMV newborns were identified from 575 newborns (20 CMV+, 555 CMV -)
Thus, (20x10/13)/(20x10/13+555x50/96)=0.051
2. Among the 66 cCMV cases identified by the universal screening program in the previous
D
study,
TE
2) Cases w/o clinical manifestations but whose cCMV could be identified only by
serology=(66-30)x30%=10.6
Thus, (33+10.6)/66 = 62% of cCMV cases in the previous study could be presumably
identified.
EP
C
C
A
13
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
REFERENCES
1. Pass RF. Cytomegalovirus. In: Knipe DM, Howley PM, eds. Fields VIrology. 4th.
D
3. Koyano S, Inoue N, Nagamori T, et al. Dried umbilical cords in the retrospective
TE
Clin Infect Dis. 2009;48:e93-e95.
2009;63:862-867.
EP
5. Yoshinaga-Itano C. Early intervention after universal neonatal hearing screening:
using newborn urine samples collected on filter paper: feasibility and outcomes from a
14
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
10. Matsuo K, Morioka I, Oda M, et al. Quantitative evaluation of ventricular dilatation
11. Mitani Y, Lezhava A, Kawai Y, et al. Rapid SNP diagnostics using asymmetric
2007;4:257-262.
D
12. Kohda C, Chiba N, Shimokoba K, et al. A simple smart amplification assay for the
TE
2014;208:160-165.
1980;66:758-762.
15. Kenneson A, Cannon MJ. Review and meta-analysis of the epidemiology of congenital
C
16. Boppana SB, Ross SA, Fowler KB. Congenital cytomegalovirus infection: clinical
C
17. Leung AKC, Sauve RS, Davies HD. Congenital cytomegalovirus infection. J Natl Med
A
Assoc. 2003;95:213-218.
18. Naing ZW, Scott GM, Shand A, et al. Congenital cytomegalovirus infection in
19. Griffiths PD, Walter S. Cytomegalovirus. Curr Opin Infect Dis. 2005;18:241-245.
15
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
20. Sonoyama A, Ebina Y, Morioka I, et al. Low IgG avidity and ultrasound fetal
1928-1933.
21. Nagamori T, Koyano S, Inoue N, et al. Single cytomegalovirus strain associated with
fetal loss and then congenital infection of a subsequent child born to the same mother. J
D
22. Ikuta K, Minematsu T, Inoue N, et al. Cytomegalovirus (CMV) glycoprotein H-based
TE
congenital CMV infection and their mothers. J Clin Virol. 2013;58:474-478.
23. Boppana SB, Rivera LB, Fowler KB, Mach M, Britt WJ. Intrauterine transmission of
2001;344:1366-1371.
EP
24. Mussi-Pinhata MM, Yamamoto AY, Moura Brito RM, et al. Birth prevalence and
cytomegalovirus infection using saliva can be influenced by breast feeding. Arch Dis
16
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
FIGURE LEGENDS
Figure 1. Study design and numbers of newborns enrolled and diagnosed. This study
consisted of two parts. Part A was for screening based on clinical manifestations, and Part B
an in-house PCR assay. Each dot represents one urine specimen. A) In-house qPCR vs.
D
QIAGEN qPCR. B) In-house qPCR vs. Shino-Test SMAP-based assay. The arrow indicates
TE
EP
C
C
A
17
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Table 1. Comparison between the universal and risk-based screening approaches
this study previous study fold of enrichment
screening strategy risk-based universal
D
TE
EP
C
C
A
18
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Table 2. Clinical manifestations and abnormalities among the newborns
clinical manifestations considered as number of newborns analyzed
consequece of cCMV infection CMV- (n=107) CMV+ (n=20)
fetuses
intrauterine growth restriction 20 8
ultrasound abnormality 5 6
newborns
low birth weight 38 9
premature birth 23 4
liver function abnormality 19 4
splenohepatomegaly 1 2
anemia 5 3
D
jaundice that may require treatments 30 7
pneumonia/respiratory disorder 22 3
thrombocytopenia 3 6
petechiae/ecchymosis 0 5
TE
hearing impairment 4 6
chorioretinitis 0 1
microcephaly 7 5
hydrocephalus 2 6
meningoencephalitis 0 0
intracerebral calcification 0 6
EP
mother during pregnancy
fever of unknown origin 14 1
detection of CMV-specific IgM 50 (n=96) 10 (n=13)
seroconversion of CMV-specific IgG 1 (n=95) 4 (n=15)
19
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Table 3 Relationship of CMV-IgM-positivity with numbers of additional manifestations
numbers of cases
CMV-
cCMV numbers of manifestations other than CMV-IgM
specific IgM total
0 1 2 3
+ 10 4 1 1 4
+ - or 3 0 1 2
ND 7 2 0 5
+ 50 41 5 3 1
- - or 46 12 11 23
ND 11 1 1 9
D
+:present, -:absent, ND:not determined
TE
EP
C
C
A
20
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Figure 1
D
TE
EP
C
C
A
21
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
Figure 2
D
TE
EP
C
C
A
22
Copyright 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.