Cytomegalovirus (CMV) is a common cause of congenital infection worldwide1 and a leading nongenetic cause of sensorineural hearing loss.2 Most infants (85%–90%) with congenital CMV infection (cCMV) lack clinical abnormalities (“asymptomatic” cCMV) and are not identified in the newborn nursery.3 Approximately 10%–15% of these asymptomatic infants and about 50% of infants with clinical abnormalities at birth (“symptomatic” cCMV) will develop sequelae including sensorineural hearing loss beyond the neonatal period.4,5 In a large, multicenter study (CMV and Hearing Multicenter Screening or CHIMES study), we demonstrated that screening newborns for CMV infection by real-time PCR testing of saliva samples for CMV DNA was highly sensitive and specific.6 The objective of this study was to determine whether the real-time PCR assay identified more newborns with CMV than rapid culture of saliva specimens, the latter being the gold standard for identifying CMV-infected newborns.
MATERIALS AND METHODS
Infants born at 7 hospitals in the United States between June 2008 and March 2012 were enrolled prospectively in the National Institute on Deafness and Other Communication Disorders (NIDCD)-sponsored CHIMES study.6 Between June 2008 and December 2009, 35,334 newborns were screened for CMV by rapid culture and real-time PCR performed on saliva swabs placed in transport medium (liquid saliva specimens) as described previously.6 From January 2010 to March 2012, 36,905 newborns were screened for CMV by PCR of dried saliva specimens, followed by rapid culture of PCR-positive specimens.
The CMV testing protocols for PCR and rapid culture have been described.6 Positive screening results (PCR or rapid culture of newborn saliva) were confirmed by follow-up rapid culture and PCR testing of urine and saliva samples obtained within 3–6 weeks of birth. The concordance between PCR and rapid culture of newborn screening saliva samples was compared to determine whether use of saliva PCR for screening identifies more infants with cCMV.
Determination of Viral Load
Viral load values expressed as international units (IU/mL) based on calibration to the CMV WHO Standards,7 and the number of fluorescent cells on rapid culture were compared in discordant samples to determine whether the discordance between real-time PCR and rapid culture was secondary to low viral load.
As the real-time PCR and rapid culture are paired samples, the McNemar test was used to compare the results of the two assays. Viral load between concordant and discordant samples was analyzed using unpaired t test.
Results of Newborn CMV Screening
Of the 73,239 infants screened for CMV during the study period, 284 (0.4%) infants tested positive by PCR or rapid culture of saliva and were enrolled for confirmation of cCMV. The mean (± standard deviation) interval between sample collection and performance of PCR and rapid culture was 9.2 ± 5.3 days and 9.1 ± 5.5 days, respectively (P = 1). On confirmatory testing, 18 infants had negative PCR and rapid culture of saliva and urine samples and were considered to have false-positive screening results. Screening PCR and rapid culture assay results for the 266 infants confirmed to have cCMV were compared. Of these, newborn saliva samples from 252 (94.7%) were positive by both PCR and rapid culture. Discordant results between PCR and rapid culture were observed in 14 infants, and of these, 13 were PCR positive and rapid culture negative, whereas 1 was rapid culture positive and PCR negative (Fig. 1). The number of samples with discordant results between the two screening assays was significantly higher when tested by rapid culture than PCR (McNemar test; P = 0.003).
Duration Between Collection of Screening Samples and Testing
When compared to samples with concordant results, the mean duration (±standard deviation) from collection to testing was longer for samples with discordant results by both PCR (9.5 ± 5.3 vs. 12.4 ± 3.3 days; P = 0.05) and rapid culture (8.8 ± 5.4 vs. 14.6 ± 5 days; P = 0.02), respectively. Among samples with discordant results, there was no significant difference in duration between sample collection to performance of PCR or rapid culture (P = 1).
Viral Load in Discordant Specimens
The median viral load was not significantly different between discordant (1.86 × 105 IU/mL; range: 6.4 × 102 to 4.8 × 107 IU/mL) and concordant (2.5 × 106 IU/mL; range: 1 × 103 to 3 × 1010 IU/mL, P = 0.7) samples. The one sample that was rapid culture positive and PCR negative had only 4 positive fluorescent cells per well.
The results of this large newborn screening study showed that saliva real-time PCR identified more infants with cCMV compared to the traditional rapid culture technique. Of the 266 infants with confirmed cCMV, 13 infants were identified only by PCR. Based on the estimates that 20,000–30,000 infants with cCMV are born each year in the United States, between 900 and 1400 infected infants could be missed if rapid culture of saliva is used as the sole screening tool. Although rapid culture of saliva or urine has been the standard method for the identification of infants with cCMV,8,9 alternate methods are being evaluated for diagnostic testing and newborn screening because culture-based methods are labor-intensive and not amenable for high throughput capacity.10
Overall, 94.7% concordance was observed between real-time PCR and rapid culture of newborn saliva specimens. Of the 14 infants with discordance between PCR and rapid culture of saliva, 13 were missed by rapid culture, whereas only 1 sample was missed by PCR. Possible explanations for the discordance between PCR and rapid culture include a significant reduction in the amount of infectious virus during storage and transport of samples and the lag time between collection and testing leading to decreased sensitivity of rapid culture. It has been shown previously that virus titers in urine specimens decrease after 1 week even when kept at 4oC,11 which may also occur with saliva samples since the average time from sample collection to testing was >7 days. We found no significant difference in viral load between concordant and discordant samples. However, the discordance might be explained by the interval between sample collection and testing, which was found to be longer for discordant samples. While the longer interval does not appear to affect the results of PCR, it may have decreased the sensitivity of rapid culture.
This study has few limitations. The impact of storage on the time between collection and testing on PCR and rapid culture was not examined in this study. During the latter part of the study, rapid culture of newborn saliva samples was only performed on samples that tested positive by PCR. Therefore, it was not possible to determine the number of infected infants who would have identified only by rapid culture during this period. Also, saliva PCR yielded false-positive results in 18 infants in this study and emphasizes the importance of confirming cCMV infection in all infants who test positive on screening.
In conclusion, although both PCR and rapid culture identify most children with cCMV, PCR appears to identify more congenitally infected infants who would be missed by rapid culture. In addition, saliva samples are easy to obtain and the PCR technique would enable automation.
1. Kenneson A, Cannon MJ. Review and meta-analysis of the epidemiology of congenital cytomegalovirus (CMV) infection. Rev Med Virol. 2007;17:253–276
2. Morton CC, Nance WE. Newborn hearing screening–a silent revolution. N Engl J Med. 2006;354:2151–2164
3. Fowler KB, Dahle AJ, Boppana SB, et al. Newborn hearing screening: will children with hearing loss
caused by congenital cytomegalovirus infection be missed? J Pediatr. 1999;135:60–64
4. Williamson WD, Demmler GJ, Percy AK, et al. Progressive hearing loss
in infants with asymptomatic congenital cytomegalovirus infection. Pediatrics. 1992;90:862–866
5. Fowler KB, McCollister FP, Dahle AJ, et al. Progressive and fluctuating sensorineural hearing loss
in children with asymptomatic congenital cytomegalovirus infection. J Pediatr. 1997;130:624–630
6. Boppana SB, Ross SA, Shimamura M, et al.National Institute on Deafness and Other Communication Disorders CHIMES Study. Saliva
polymerase-chain-reaction assay for cytomegalovirus screening in newborns. N Engl J Med. 2011;364:2111–2118
7. Kraft CS, Armstrong WS, Caliendo AM. Interpreting quantitative cytomegalovirus DNA testing: understanding the laboratory perspective. Clin Infect Dis. 2012;54:1793–1797
8. Yamamoto AY, Mussi-Pinhata MM, Pinto PC, et al. Usefulness of blood and urine samples collected on filter paper in detecting cytomegalovirus by the polymerase chain reaction technique. J Virol Methods. 2001;97:159–164
9. Balcarek KB, Warren W, Smith RJ, et al. Neonatal screening for congenital cytomegalovirus infection by detection of virus in saliva
. J Infect Dis. 1993;167:1433–1436
10. Boppana SB, Ross SA, Novak Z, et al.National Institute on Deafness and Other Communication Disorders CMV and Hearing Multicenter Screening (CHIMES) Study. Dried blood spot real-time polymerase chain reaction assays to screen newborns for congenital cytomegalovirus infection. JAMA. 2010;303:1375–1382
11. Stagno S, Pass RF, Reynolds DW, et al. Comparative study of diagnostic procedures for congenital cytomegalovirus infection. Pediatrics. 1980;65:251–257