Cytomegalovirus (CMV) is the most common cause of congenital infection in developed countries (0.45%–0.7% of live births).1,2 Nonetheless, universal newborn screening has not been implemented for congenital CMV (cCMV) infection. The importance of an early accurate diagnosis of this condition resides in the possibility to improve hearing and neurodevelopmental outcomes by prompt antiviral therapy, in infants who are symptomatic at birth even without central nervous system involvement.3–5
The diagnosis of cCMV in the newborn is confirmed by identification of the virus through positive culture findings or by molecular techniques in various clinical samples (blood, urine, saliva, cerebrospinal fluid) during the first 2 weeks of life,6 with urine remaining the gold standard sample. Beyond that period, isolation of viral DNA in dried blood spots (DBSs) collected at birth for newborn screening by real-time polymerase chain reaction (rt-PCR) is a useful retrospective diagnostic tool in some cases.7,8 Several European studies have reported sensitivities of 99%–100% for the diagnosis of cCMV in DBS samples.7,9–11 In contrast, Boppana et al2 reported sensitivities of 28% and 34%, depending on the polymerase chain reaction (PCR) technique used, and in a recent meta-analysis of 15 studies, a wide range of sensitivities was reported.7 The sensitivities were higher in retrospective studies suggesting higher sensitivities in patients with long-term sequelaes. De Vries et al8 compared 8 CMV DNA extraction methods from DBS (using a single amplification method) and reported sensitivities of 32%–73% with a single testing method and higher values with triplicate testing. A direct relationship was found between the sensitivity and the viral load (VL) logarithmic category. In a study performed by Soetens et al,12 the best extraction method together with the best amplification method (rt-PCR) raised the sensitivity to 82%. A preliminary study performed by our group in Spain found a sensitivity of 50% for DBS PCR in a small series of 14 patients,13 and this led us to perform a nationwide study in a larger sample.
The aim of the present study was to evaluate the reliability of our rt-PCR technique to retrospectively detect CMV DNA in DBS of neonates with confirmed cCMV included in the Spanish Registry of cCMV (REDICCMV) Infection.
MATERIALS AND METHODS
Patients born between January 2007 and January 2016 with confirmed cCMV diagnosed by positive rt-PCR or viral culture in samples of any body fluid (urine, blood, cerebrospinal fluid) within the first 2 weeks of life, and included in REDICCMV, were invited to participate. The follow-up period was 6 months or longer in all patients.
REDICCMV is a multicenter, ambispective, observational national registry of infants with cCMV in Spain. At the time of the study (November 2016), 378 patients from 40 hospitals had been included. The recorded data cover maternal and prenatal information and the infant’s clinical, laboratory and imaging data, antivirals received and clinical, neurodevelopmental and hearing outcomes.
The registry data are collected and managed using the Research Electronic Data Capture application (Vanderbilt University Medical Center, Nashville, Tennessee) tools, a secure web-based application designed to support data capture for research studies,14 hosted at Hospital Universitario 12 de Octubre (Madrid, Spain). The registry was approved by the Institutional Review Board of Hospital Universitario 12 de Octubre.
Infants in whom cCMV had definitively been ruled out by CMV PCR urine testing at birth were used as negative controls. CMV in urine had been assessed because of clinical criteria at birth (eg, prematurity, low birth weight, microcephaly, petechial exanthema) and tested negative in all cases. All controls were recruited in a single center between 2014 and 2016.
The data collected for this study included sex, age in days at the cCMV infection diagnosis, symptoms at birth, sensorineural hearing loss (SNHL) and neurologic impairment at birth and during follow-up (6 months minimum), DBS size (in mm), time elapsed between DBS collection and sample processing, plasma VL at birth when available and use of antiviral therapy. Infants who showed intrauterine growth retardation, pathologic neuroimaging, hearing loss, thrombocytopenia, petechiae, microcephaly, liver and/or spleen enlargement, hypotonia, sepsis, jaundice, elevated liver enzymes or chorioretinitis were considered as symptomatic.
DBSs were collected by staff from the local newborn metabolic screening laboratories and sent to the coordinating center, where they were analyzed in the Microbiology Department’s molecular diagnosis laboratory.
The type of the filter paper used to collect blood and the diameter of the DBS differed between the participating regions in Spain (Table, Supplemental Digital Content 1, http://links.lww.com/INF/D218).
Two full blood spots, obtained by cutting the paper with a sterile scalpel, were placed together in a sample tube, and 1 mL of phosphate saline buffer was added. The tubes were placed in a vortex mixer to facilitate transfer of the dried blood to the solution. A 400-μL amount of this sample was used for DNA extraction in an automated system (EasyMag; bioMérieux, Marcy, l’Étoile, France), with a minor modification of the standard protocol (200 μL of silica was added). According to the manufacturer, the lysis of the sample was done in a 2 mL tube of Nuclisens EasyMag lysis buffer and were left for 20 minutes at room temperature. Ten μL eluate derived from 100 μL DNA eluate was used as the input volume into the PCR reaction. Amplification was carried out by an rt-PCR technique (RealStar CMV, Altona, Germany) in the Smartcycler thermal cycler (Cepheid, Sunnyvale, CA). The rt-PCR assay includes a heterologous amplification system (internal control) to identify possible rt-PCR inhibition and to confirm the integrity of the reagents. According to the manufacturer, the limit of detection of the technique is 0.42 IU/μL and the limit of quantification is 1 IU/μL. VL is expressed in IU/mL and in base-10 logarithmic scale. As there was a predilution step, the blood had a dilution factor of 6.67 (1000/150), so the limit of detection of the technique in this sample was 0.7 IU/μL.
Categorical variables were compared with the χ2 test and Fisher exact test and expressed as the number and percentage [with 95% confidence intervals]. The Mann-Whitney U test was used to compare continuous variables, and the Spearman rank correlation coefficient to evaluate relationships between them. Continuous variables were expressed as the median and interquartile range (IQR). Statistical analyses were performed using SPSS version 20 (IBM, Armonk, New York). A p value of <0.05 was considered statistically significant.
Parental consent was obtained in all cases and controls, and the local ethics committees from all participating centers gave their approval for the study.
Ten hospitals contributing to the REDICCMV registry from 5 autonomous communities in Spain agreed to participate in the study, which included 103 cCMV patients and 81 controls.
In the patient group, 52% were female, 33.3% were premature, and 26.2% were small for gestational age. Median gestational age was 38 weeks (IQR: 36–39), and median weight was 2647 grams (IQR: 2232–3045). Sixty-five (63.1%) infants showed birth signs or symptoms consistent with cCMV (Table 1), and 64 (62.7%) received antiviral treatment (ganciclovir alone, ganciclovir followed by valganciclovir or valganciclovir alone). The median follow-up time was 35.4 months (IQR: 15.7–50.8). At last follow-up, 23 (24%) patients had neurologic impairment and 27 (27.8%) had SNHL. The median time elapsed between DBS collection at birth and sample processing was 23.5 (IQR: 12.8–42) months. Forty-four 8-mm diameter samples were tested, as well as one 9 mm, thirty 12 mm, twenty-five 13 mm and three 15 mm. VL at birth (bVL) data were available in 95 patients and detectable in 82 (86.3%): median bVL was 2364 IU/mL (3.4 log10 copies/mL, IQR: 2.5–4) in symptomatic and 1548 IU/mL (3.2 log10 copies/mL, IQR: 2.2–3.5) in asymptomatic neonates.
Fifty-eight (56.3%) DBS samples from the patient group tested positive by CMV DNA rt-PCR. Among the 81 control samples, 2 tested positive and 79 negative. Therefore, the performance of the CMV DNA rt-PCR assay in neonatal DBS for the diagnosis of cCMV was as follows: sensitivity 0.56 (95% confidence interval: 0.47–0.65), specificity 0.98 (0.91–0.99), positive likelihood ratio 22.81 (5.74–90.58) and negative likelihood ratio 0.45 (0.36–0.60).
The only variable significantly associated with negative CMV DNA results on DBS testing was plasma bVL (P = 0.017) (Table, Supplemental Digital Content 2, http://links.lww.com/INF/D219) (Fig. 1). No associations were found for any of the clinical features studied, the time elapsed between collection and processing of the samples or the size or the geographical origin of the DBS. The sensitivity of the assay did not differ between analyses of symptomatic and asymptomatic neonates: 0.54 (0.42–0.65) and 0.6 (0.45–0.74), respectively (P = 0.51).
Sensitivity was further calculated according to bVL category, with the following findings: 0.4 [0.22–0.61] for 2–3 log10 bVL (20 patients), 0.68 [0.48–0.83] for 3–4 log10 bVL (25 patients) and 0.78 [0.55–0.91] for 4–5 log10 bVL (18 patients) (Fig. 2).
This study, investigating the usefulness of rt-PCR for detecting CMV DNA in DBS, includes the largest series of neonates with confirmed cCMV reported to date and provides an indication of daily clinical practice in our country. The epidemiologic data of the cCMV patients recruited were similar to those reported in previous studies,6,15 but 63% of patients were symptomatic at birth. This value is higher than would be expected in the natural history of cCMV infection (10%–15%)1, 6, 15, 16 and was the reason why many patients received antiviral therapy (62.1%). At follow-up, 24% had neurologic impairment and 27.8% had SNHL. These major differences compared with findings in previous studies are likely because of the fact that our cohort was mainly diagnosed based on clinical suspicion of infection and not during population-based screening.
The overall sensitivity of our assay (56%) was far lower than has been reported in studies by several European groups,7,12,17–19 but the specificity remained very high. These results are worthy of note because if the decision to treat a symptomatic baby beyond the first 2 weeks of life is based on DBS rt-PCR results, then 44% of infants with cCMV in our cohort would have been missed. Furthermore, not all cCMV patients have a detectable bVL; hence, even if a 100% sensitive technique was available, some patients would still be missed. These findings suggest that retrospective identification of cCMV patients by this technique may be suboptimal and prompt us to consider whether universal cCMV screening in urine or saliva may be a more reliable option.20,21 Recent studies have described PCR assays for use with saliva samples,22,23 which would open a new possibility for hypothetical universal screening.
On analysis of the patients’ specific clinical features, none of the factors studied showed a correlation with the DBS rt-PCR results. The only patient variable that was significantly associated with the DBS findings was the plasma VL at birth: patients with a low bVL tested negative on DBS rt-PCR more often than those with higher levels. Some studies have reported10,18,24 that patients with birth symptoms have a higher DBS VL, although Soetens et al12 found no differences in DBS VL results between infants with and without sequelae. In accordance with this last study, we found no correlations between the DBS results and the presence of birth symptoms, SNHL at follow up or neurologic impairment. The studies published about the relationship between bVL and outcome are contradictory. While Lanari et al25 found differences in the outcomes related with bVL, Ross et al26 could not demonstrate a relationship between bVL and SNHL, except for asymptomatic children with less than 3500 ge/mL who presented a more favorable outcome, in a larger study.
A standard DBS sample contains about 50 µL of whole blood in a 12-mm2 diameter spot of filter paper.27 However, as was mentioned, the type of paper differed between the different regions of Spain. Although this may have resulted in different amounts of blood in the spots, which could potentially have affected the performance of the assay, we found no differences related to the sample origin or DBS size. In contrast, in a meta-analysis, Wang et al7 reported differences between studies using large vs small DBS areas, but they defined a large area as >25 mm2, equivalent to a 5.6-mm diameter DBS, which is much smaller than the spots used in our study. In keeping with the results of Soetens et al,12 but contrary to Christoni et al,28 we found that the time elapsed between the birth date and DBS processing did not affect the rt-PCR results because of theoretical changes in DNA integrity.
Regarding the extraction method, which could be thought to be suboptimal because there is a predilution step that can decrease the limit of detection, when comparing our results by VL logarithmic categories with those reported by de Vries et al,8 our results showed higher sensitivity for the low and medium VL categories and slightly lower sensitivity for the high category. The predilution step increases the limit of detection from 0.42 IU/μL to 0.7 IU/μL, but we used 400 μL of sample for the extraction, to increase the sensitivity. So, the effect of the predilution step sensitivity is minimized. Thus, the lower sensitivity we observed could be explained, at least partially, by a lower plasma bVL in our series (median bVL was 3.3 log10) and not only by the extraction method. Of note, even with the best extraction technique in the study by de Vries et al8 (the unmodified Barbi protocol), sensitivity in the 2–3 log10 VL category was 0.2, lower than ours (0.4). Manual extraction methods are reported to be the most sensitive12 but are more difficult to use in clinical laboratories and involve a risk of external sample contamination.
It has been reported that rt-PCR may have higher sensitivity than standard PCR techniques.12 Atkinson et al27 reported a new amplification technique (single tube nested PCR) that increased overall sensitivity from 69% to 81%. Whether this new technique would have increased the sensitivity in our study, in which median bVL was low, warrants further investigation. Moreover, as some percentage of cCMV patients have an undetectable bVL (between 5.8%29 and 26%30), 100% sensitivity does not seem to be a realistic goal.
The fact that our study was performed following daily clinical practice is both a limitation and strength of the study. The DBS collection technique may have differed in the various participating autonomous regions of the country, and plasma bVL was quantified in different laboratories. Hence, our sensitivity data should be viewed in relation to routine clinical practice in a cohort with relatively low plasma bVL. Another limitation is that the study population comes from a cohort of well-defined patients rather than universal screening; therefore, some paucisymptomatic and asymptomatic patients may have been missed. However, the absence of a correlation between birth symptoms and the test results seems to minimize the impact of this limitation. In contrast with Wang et al,7 who reported higher sensitivities in studies involving symptomatic patients compared with those with universal screening design, we found lower sensitivity in a nonscreening-based study. The lower sensitivity we observed could be explained, at least partially, by a lower bVL in our series (median bVL was 3.3 log10), as previously mentioned.
The authors thank all the patients and their families for their contribution, Celine Cavallo for English language support and Santiago Pérez-Hoyos for statistical support. The authors are also grateful to Dr. Daniel Blázquez-Gamero for his contribution in Spanish Registry of Congenital Cytomegalovirus coordination.
1. Dollard SC, Grosse SD, Ross DS. New estimates of the prevalence of neurological and sensory sequelae and mortality associated with congenital cytomegalovirus infection
. Rev Med Virol. 2007;17:355–363.
2. 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.
3. Kimberlin DW, Lin CY, Sánchez PJ, et al; National Institute of Allergy and Infectious Diseases Collaborative Antiviral Study Group. Effect of ganciclovir therapy on hearing in symptomatic congenital
cytomegalovirus disease involving the central nervous system: a randomized, controlled trial. J Pediatr. 2003;143:16–25.
4. Kimberlin DW, Jester PM, Sánchez PJ, et al. Valganciclovir for symptomatic congenital
cytomegalovirus disease. N Engl J Med. 2015;372:933–943.
5. Oliver SE, Cloud GA, Sánchez PJ, et al; National Institute of Allergy, Infectious Diseases Collaborative Antiviral Study Group. Neurodevelopmental outcomes following ganciclovir therapy in symptomatic congenital
cytomegalovirus infections involving the central nervous system. J Clin Virol. 2009;46(suppl 4):S22–S26.
6. Britt W. Wilson CB, Nizet V, Remington JS, Klein JO, Maldonado Y. Cytomegalovirus. In: Infectious Diseases of the Fetus and Newborn. 2011:7th ed. Philadelphia; Saunders, 706–755.
7. Wang L, Xu X, Zhang H, et al. Dried blood spots PCR assays to screen congenital cytomegalovirus infection
: a meta-analysis. Virol J. 2015;12:60.
8. de Vries JJC, Claas ECJ, Kroes ACM, et al. Evaluation of DNA extraction methods for dried blood spots in the diagnosis of congenital cytomegalovirus infection
. J Clin Virol. 2009;46:S37–S42.
9. Snijdewind IJ, van Kampen JJ, Fraaij PL, et al. Current and future applications of dried blood spots in viral disease management. Antiviral Res. 2012;93:309–321
10. Leruez-Ville M, Ngin S, Guilleminot T, et al. Detection of cytomegalovirus DNA on dried blood spots collected from infants infected with HIV: an in-house method adaptable in resource-limited settings. J Virol Methods. 2013;193:503–507.
11. Barbi M, MacKay WG, Binda S, et al. External quality assessment of cytomegalovirus DNA detection on dried blood spots. BMC Microbiol. 2008;8:2.
12. Soetens O, Vauloup-Fellous C, Foulon I, et al. Evaluation of different cytomegalovirus (CMV) DNA PCR protocols for analysis of dried blood spots from consecutive cases of neonates with congenital
CMV infections. J Clin Microbiol. 2008;46:943–946.
13. Can we rule out a congenital cytomegalovirus infection
when the result of polymerase chain reaction
in dried blood spots is negative. Enferm Infecc Microbiol Clin. 2014;32:570–573.
14. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381.
15. Boppana SB, Ross SA, Fowler KB. Congenital cytomegalovirus infection
: clinical outcome. Clin Infect Dis. 2013;57(suppl 4):S178–S181.
16. Goderis J, De Leenheer E, Smets K, et al. Hearing loss and congenital
CMV infection: a systematic review. Pediatrics. 2014;134:972–982.
17. Leruez-Ville M, Vauloup-Fellous C, Couderc S, et al. Prospective identification of congenital cytomegalovirus infection
in newborns using real-time polymerase chain reaction
assays in dried blood spots. Clin Infect Dis. 2011;52:575–581.
18. Vauloup-Fellous C, Ducroux A, Couloigner V, et al. Evaluation of cytomegalovirus (CMV) DNA quantification in dried blood spots: retrospective study of CMV congenital
infection. J Clin Microbiol. 2007;45:3804–3806.
19. Barbi M, Binda S, Primache V, et al. Congenital cytomegalovirus infection
in a northern Italian region. NEOCMV Group. Eur J Epidemiol. 1998;14:791–796.
20. Kadambari S, Luck S, Davis A, et al. Clinically targeted screening for congenital
CMV - potential for integration into the National Hearing Screening Programme. Acta Paediatr. 2013;102:928–933.
21. Gantt S, Dionne F, Kozak FK, et al. Cost-effectiveness of universal and targeted newborn screening for congenital cytomegalovirus infection
. JAMA Pediatr. 2016;170:1173–1180.
22. Yamamoto AY, Mussi-Pinhata MM, Marin LJ, et al. Is saliva as reliable as urine for detection of cytomegalovirus DNA for neonatal screening of congenital
CMV infection? J Clin Virol. 2006;36:228–230.
23. 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.
24. Walter S, Atkinson C, Sharland M, et al. Congenital
cytomegalovirus: Association between dried blood spot viral load and hearing loss. Arch Dis Child Fetal Neonatal Ed. 2008.
25. Lanari M, Lazzarotto T, Venturi V, et al. Neonatal cytomegalovirus blood load and risk of sequelae in symptomatic and asymptomatic congenitally infected newborns. Pediatrics. 2006;117:e76–e83.
26. Ross SA, Novak Z, Fowler KB, et al. Cytomegalovirus blood viral load and hearing loss in young children with congenital
infection. Pediatr Infect Dis J. 2009;28:588–592.
27. Atkinson C, Emery VC, Griffiths PD. Development of a novel single tube nested PCR for enhanced detection of cytomegalovirus DNA from dried blood spots. J Virol Methods. 2014;196:40–44.
28. Christoni Z, Syggelou A, Soldatou A, et al. Evaluation of a modified extraction protocol increasing sensitivity in quantification of CMV viremia in Guthrie cards. J Clin Virol. 2012;55:360–362.
29. Boppana SB, Fowler KB, Pass RF, et al. Congenital cytomegalovirus infection
: association between virus burden in infancy and hearing loss. J Pediatr. 2005;146:817–823.
30. Bradford RD, Cloud G, Lakeman AD, et al; National Institute of Allergy and Infectious Diseases Collaborative Antiviral Study Group. Detection of cytomegalovirus (CMV) DNA by polymerase chain reaction
is associated with hearing loss in newborns with symptomatic congenital
CMV infection involving the central nervous system. J Infect Dis. 2005;191:227–233.