Pediatric and Developmental Pathology 17, 176–180, 2014 DOI: 10.2350/14-01-1432-OA.1 ª 2014 Society for Pediatric Pathology
Evaluation of Serial Urine Viral Cultures for the Diagnosis of Cytomegalovirus Infection in Neonates and Infants KAREN M. CHISHOLM,1 NATALI AZIZ,2 MICHAL MCDOWELL,2 FRANCES P. GUO,1 NIVEDITA SRINIVAS,3 WILLIAM E. BENITZ,3 MARY E. NORTON,2 KATHLEEN GUTIERREZ,3 ANN K. FOLKINS,1 AND BENJAMIN A. PINSKY1,4* 1
Department Department 3 Department 4 Department CA, USA 2
of Pathology, Stanford University School of Medicine, Stanford, CA, USA of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford, CA, USA of Pediatrics, Stanford University School of Medicine, Stanford, CA, USA of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, Stanford,
Received January 19, 2014; accepted March 5, 2014; published online March 11, 2014.
ABSTRACT Cytomegalovirus (CMV) is the most common cause of congenital infection worldwide. Urine viral culture is the standard for CMV diagnosis in neonates and infants. The objectives of this study were to compare the performance of serial paired rapid shell vial cultures (SVC) and routine viral cultures (RVC), and to determine the optimal number of cultures needed to detect positive cases. From 2001 to 2011, all paired CMV SVC and RVC performed on neonates and infants less than 100 days of age were recorded. Testing episodes were defined as sets of cultures performed within 7 days of one another. A total of 1264 neonates and infants underwent 1478 testing episodes; 68 (5.4%) had at least one episode with a positive CMV culture. In episodes where CMV was detected before day 21 of life, the first specimen was positive in 100% (16/16) of cases. When testing occurred after 21 days of life, the first specimen was positive in 82.7% (43/52) of cases, requiring three cultures to reach 100% detection. The SVC was more prone to assay failure than RVC. Overall, when RVC was compared to SVC, there was 86.0% positive agreement and 99.9% negative agreement. In conclusion, three serial urine samples are necessary for detection of CMV in specimens collected between day of life 22 and 99, while one sample may be sufficient on or before day of life 21. Though SVC was more sensitive than RVC, the risk of SVC failure supports the use of multimodality testing to optimize detection.
This work was presented at the Society for Pediatric Pathology Fall Meeting in Salt Lake City, UT, USA, September 27, 2013. *Corresponding author, e-mail: [email protected]
Key words: cytomegalovirus, hearing loss, sensorineural, virus cultivation
INTRODUCTION Cytomegalovirus (CMV) is the most common congenitally acquired viral infection worldwide, with a birth prevalence of 0.6 to 0.7% [1,2]. Approximately 10% to 15% of neonates are symptomatic at birth, with signs and symptoms including jaundice, hepatomegaly, splenomegaly, thrombocytopenia, microcephaly, and low birth weight [2–5]. Of these symptomatic infants, 40% to 90% will later develop permanent disabilities, including hearing loss, vision loss, and cognitive impairments [2– 6]. Of the 85% to 90% of infants who are asymptomatic at birth, 10% to 15% will develop permanent disabilities, predominantly hearing loss, but also cognitive or neurologic impairments [2,4,6]. Overall, congenital CMV infection is the most common nonhereditary cause of sensorineural hearing loss, affecting 10% to 15% of congenitally infected CMV individuals, and accounting for 15% to 20% of cases of congenital hearing loss [7]. When congenital infection is suspected, viral culture is performed ideally within the first two to three weeks of life [5,6,8,9]. Otherwise, differentiating congenital from early postnatal infection is difficult. Postnatal transmission can occur due to exposure to maternal genital tract secretions during delivery, maternal breast milk ingestion, and transfusions with CMV seropositive blood [8]. While postnatal infection may be asymptomatic, it also can cause a septic syndrome that may include hepatomegaly, splenomegaly, thrombocytopenia, and pneumonia [5]. However, compared to those with congenital CMV, infants with postnatal CMV exposure are not at risk for
sensorineural hearing loss or other neurologic or cognitive deficiencies [10]. The standard methods for the diagnosis of CMV infection in neonates and infants are urine or saliva viral culture [8,9]. Detection of CMV can be accomplished through routine viral culture (RVC) methods or centrifugation-enhanced rapid shell vial cultures (SVC). The RVC has lengthy turn-around times, taking usually 1 to 2 weeks, but sometimes as long as 4 to 6 weeks, to develop cytopathic effect (CPE). The SVC methods have been found to be as specific as, and even more sensitive than, conventional viral cultures and require only 12 to 72 hours [11–13]. However, neonatologists and pediatricians often order multiple cultures to ensure CMV detection in infected infants. Real-time polymerase chain reaction (PCR) also may be used for CMV detection in neonatal urine, though neither sensitivity nor specificity show a statistically significant difference compared to SVC [14]. In patients with congenital CMV infection, treatment with ganciclovir can prevent hearing deterioration and may improve neurodevelopmental outcomes [15,16]. As such, prompt and accurate diagnosis of CMV infection is critical for appropriate evaluation and timely management.
METHODS Patient selection With Stanford University Medical Center Institutional Review Board approval, data on all urine viral cultures performed on infants under 100 days of age, from January 2001 through December 2011, were collected. Inclusion required that specimens were submitted for paired CMV SVC and RVC. Patient name, date of birth, medical record number, and date of cultures were recorded to tabulate and group cultures by patient. Testing episodes were defined as sets of paired cultures that were performed within 7 days of one another. Viral cultures Urine specimens were collected in sterile containers and stored at 2u to 8uC before testing. For RVC, urine specimens were mixed 1:2 with antimicrobial mix containing gentamicin (150 mg/mL), vancomycin (1000 mg/mL), and amphotericin B (35 mg/mL). Of the processed urine, 0.25 mL was inoculated onto each traditional tube culture monolayer including human foreskin (HF) fibroblasts (Diagnostic Hybrids, Athens, OH, USA) and human embryonic lung MRC-5 fibroblasts (Viromed, Minnetonka, MN, USA) containing 1.0 to 1.5 mL of the manufacturer’s culture medium. The RVCs were incubated for 28 days at 35uC. Cell monolayers with demonstrable CPE were scraped, washed with phosphate buffered saline (PBS; Sigma-Aldrich, St. Louis, MO, USA), spotted to Cel-Line Supercured HTC slides (Thermo Fisher Scientific, Waltham, MA, USA), and fixed for 10 minutes in acetone. The cells then were stained for the presence of CMV by indirect immunofluorescence using anti-CMV antibodies targeting the immediate early antigen (Millipore, Billerica, MA,
Table 1.
Number of specimens submitted per episode
Specimens per episode
Episodes
%
1 2 3 4 5 6 Total
649 235 512 72 6 4 1478
43.9 15.9 34.6 4.9 0.4 0.3 100%
USA) according to the manufacturer’s specifications. The number of days to identifiable CPE was noted for all positive cultures. For SVC, two human fibroblast vials (one HF and one MRC-5 from the vendors described above) were each inoculated with 0.25 mL processed urine after aspiration of the culture medium. The vials were centrifuged at 3000 rpm for 30 minutes, the monolayers refed, and the vials incubated at 35uC. At 24 hours, the cell monolayer from one vial (HF) was washed with PBS, fixed with acetone, and stained as above with the same antibody reagent used for RVC. The other vial was processed (MRC-5) at 48 hours. The SVC results were reported only if the monolayers were .60% confluent, the negative control showed no specific fluorescence, and the positive control, cultured CMV AD169 reference strain (American Type Culture Collection, Manassas, VA, USA), exhibited the expected apple green nuclear fluorescence. A positive SVC result on a patient sample was reported if at least one cell demonstrated this characteristic staining. The SVCs that demonstrated intense nonspecific fluorescence or excessive destruction of the monolayer by specimen toxicity or contamination were reported as unsatisfactory. When either paired SVC or RVC was positive, the specimen was considered CMV culture positive. Statistics Contingency tables were evaluated with Fisher’s exact test, 2-tailed, and the Mann-Whitney test was used to analyze days to CPE in RVC (GraphPad, La Jolla, CA, USA). In addition, Student’s t-test was used for comparing positive and negative episodes (Excel; Microsoft, Redmond, WA, USA). Positive percent agreement and negative percent agreement were calculated for SVC and RVC concordance, and evaluated by McNemar’s test.
RESULTS Culture results Over the course of the study, 2997 urine specimens from 1264 neonates and infants less than 100 days of age were submitted for CMV detection by paired RVC and SVC. Testing episodes, defined as sets of paired cultures that were performed within 7 days of one another, totaled 1478, with a range from 1 to 6 specimens per episode, and a mean of 2 specimens per episode. Table 1 details the
CMV SERIAL URINE VIRAL CULTURE
177
Table 2. Shell vial culture and routine viral culture concordance Shell vial culture
Routine vial culture Positive Negative
Positive
Negative
98 16
2 2873
A total of 7 paired shell vial cultures had equipment failure, toxic interference, quality assurance failure, or laboratory error for 24- and 48hour shell vials, and 1 routine viral culture had laboratory error. The total number of paired cultures with available results was 2989.
Figure 1. Comparison of days to cytopathic effect (CPE) for specimens collected on or before 21 days of life and specimens collected after 21 days of life (Mann-Whitney test, P , 0.0001). The solid horizontal lines indicate the median number of days to CPE.
number of specimens per episode submitted for paired RVC and SVC. Of all neonates and infants, 68 (5.4%) were positive for CMV. Compared to positive infants (range of tests per episode, 1–4; mean, 1.92), the number of tests per episode in negative infants (range of tests per episode, 1–6; mean, 2.03) was not statistically significant (Student t-test, P 5 0.347). Of the 68 CMV-positive infants, 16 (23.5%) had positive cultures on or before 21 days of life. These 16 neonates had specimens obtained between days 1 and 20 of life. Of these neonates 14 were cultured at less than 7 days old, while the other two neonates were 17 and 20 days of age. For all 16 of these neonates, the first urine culture submitted of the episode always was positive, 100% (16/16). In addition, in all of these cases, the 24hour SVC and the RVC from the first specimen always were positive. Due to multiple specimens submitted, a total of 28 cultures was performed in these infants. Only two specimens were SVC and RVC negative, both the second specimens submitted for two neonates. For the remaining 26 cultures, SVC and RVC were positive in all pairs. As shown in Figure 1, the number of days to identifiable CPE for RVC ranged from 1 to 18, with a mean of 4.3 days and median of 2.5. A total of 52 infants of 68 (76.5%) had positive cultures after 21 days of life, ranging from day 25 to day 99 of life. In these infants, the first specimen of each episode was positive in 82.7% (43/52) of cases. In 11.5% (6/52) of cases, CMV was not detected until the second specimen, and in 5.8% (3/52) of cases CMV was not detected until the third specimen. Of the infants’ first positive specimens, 76.9% (40/52) of culture pairs demonstrated SVC and RVC positivity, while only SVC
178
CHISHOLM ET AL.
was positive in 21.1% (11/52) and only RVC was positive in 1.9% (1/52). In these 52 infants, 54 total episodes were positive, since two infants had two episodes each that were positive. For these 54 episodes, a total of 107 urine specimens was submitted, with negative culture pairs accounting for 15.9% (17/107). For all positive RVCs, days to identifiable CPE ranged from 1 to 28, with a mean of 9.6 days and median of 9 days. As depicted in Figure 1, positive RVCs in infants who were CMV-positive at less than 21 days of life took significantly less time to develop CPE than RVCs performed in infants more than 21 days of life (Mann-Whitney test, P , 0.0001). Culture failure rates All 2997 urine specimens included in this study were tested by RVC and SVC. One RVC was not reported due to laboratory error (1/2997, 0.03%). Of the 2997 study specimens, 2972 had SVCs read at both 24 and 48 hours, while 19 had SVCs read only at 48 hours, and 6 had SVCs read only at 24 hours. Of the total 5969 SVCs performed, results were not reported in 109 due to toxic interference (30 in the 24-hour and 79 in the 48-hour SVC, 1.8%), 40 due to quality assurance failure (21 in the 24-hour and 19 in the 48-hour SVC, 0.67%), 28 due to laboratory error (10 in the 24-hour and 18 in the 48-hour SVC, 0.47%), and two due to laboratory equipment failure (1 in the 24hour and 1 in the 48-hour SVC, 0.03%). As such, SVC was more prone to failure (179/5969, 3.0%) than RVC (1/2997, 0.03%, Fisher’s exact test, P , 0.0001). SVC and RVC concordance Of the 2997 specimens, CMV was detected in 116 specimens from SVC and/or RVC. Eight specimens had unavailable results (due to toxic interference, quality assurance failure, laboratory error, or equipment failure) for 24- and 48-hour SVC or RVC. Using SVC as reference, the positive percent agreement between SVC and RVC was 86.0% (98/114), and the negative percent agreement was 99.9% (2873/2875, Table 2). The results of the SVC and RVC tests were significantly different (McNemar’s test, P 5 0.0013). All discordant specimens were from infants more than 21 days of age (21 versus .21 days of age, Fisher’s exact test, P 5 0.0118). The
two cultures that were only positive by RVC occurred in the same patient; this patient had two other specimens submitted in the same episode that were positive by RVC and SVC, suggesting that these were not false-positive RVCs, but false-negative SVCs. Of the 16 specimens in which only SVCs were positive, 9 specimens were taken in an episode in which other specimens also were culturepositive, 2 specimens were from infants with subsequent CMV-positive urine specimens by qualitative PCR, and 1 specimen was from an infant with a CMV-positive nasopharyngeal swab by RVC; these findings suggested that these positive SVCs were not false-positives, but that the RVCs were false-negatives. The remaining 4 positive SVC specimens were either the only specimens collected in an episode [2], or the patient had a previous specimen that was negative by both SVC and RVC [2], and, hence, the accuracy of the result could not be confirmed or refuted.
DISCUSSION Congenital CMV is an important diagnosis due to the potential attenuation of long-term sequelae with timely ganciclovir therapy, and supportive management with more frequent audiology and neurodevelopmental evaluations [15,17]. The current study showed that a single urine sample may sufficiently diagnose cases of congenital CMV infection. For all neonates with urine samples submitted on or before 21 days of life, the first SVC and the first RVC were positive. While RVC in some cases required more than 10 days to show CPE (range, 1– 18 days), the 24-hour SVC was positive in 100% of first urine specimens. Conversely, neonates and infants with urine submitted after 21 days of life needed up to three urine samples to reach 100% detection. Only 82.7% of cases had first specimen positivity, 23.1% of which had discordance between SVC and RVC results. In addition, the 24-hour SVC was negative in some of these cases, though the 48hour SVC was positive. Also, the RVC took significantly longer to develop CPE compared to specimens collected from neonates less than 21 days of life. The finding of more consistent CMV detection between SVC and RVC in those patients diagnosed on or before 21 days of life compared to those diagnosed after 21 days of life, as well as shorter time intervals to CPE in RVC, is consistent with previous studies. Congenitally infected neonates, whether they are symptomatic or not, excrete virus at birth [18]. In those with postnatal CMV infection, viral excretion begins three to 12 weeks after birth [18,19]. In addition, the CMV urine viral load is known to be higher in congenital infection compared to postnatal infection [18,20]. Thus, earlier time to positivity is expected in congenital infection, as is longer time to positivity and decreased concordance between cultures in postnatal infection. Of note, while the 21-day threshold increases the likelihood that a CMV infection was acquired congeni-
tally, it is not absolute. In this study, one neonate was identified whose first CMV testing episode included two negative cultures, one on day 1 and another on day 4 of life, before a positive second testing episode occurring on day 17 of life, consistent with postnatal infection. Furthermore, the positive RVC from the second testing episode did not demonstrate CPE until day 18. This case illustrates the importance of early CMV testing to help rule out congenital CMV infection. As of yet, widespread newborn screening for congenital CMV has not been implemented. In the event that timely testing is not performed, some groups have proposed to look back into the neonatal or fetal time period to test for CMV. For example, dried blood spots collected at birth or formalinfixed paraffin-embedded placental tissues have been used to make the diagnosis of congenital CMV [21,22]. Of the positive specimens, SVC identified 16 cases that were negative by RVC, while RVC identified 2 cases that were negative by SVC. Previous studies have shown that SVC more often is positive than the paired RVC, though the current study is the first to evaluate these methods exclusively in urine from neonates and infants. From urine collected from a variety of patient groups, Gleaves and colleagues [11] identified 32/109 specimens in which SVC was positive, but RVC was negative. Similarly, Myers and Amsterdam [23] identified 55 specimens that were SVC-positive, but RVC-negative in patients with a history of renal transplant or immunodeficiency; urine specimens accounted for 28 of these 55 specimens. While SVC was more often CMV-positive in this study than RVC, this method also was more prone to toxic interference, quality assurance failure, or laboratory error. However, the 3% SVC failure rate in this study is not large, especially in comparison with historical rates of 4% to 10.7% [24–26]. The rate of failure and the sensitivity of the assay led Paya and colleagues [25] to recommend at least two vials be inoculated per urine specimen, as was performed in this study, for maximum detection of CMV. In this study, we demonstrated that three serial urine samples are necessary for full detection of CMV in specimens collected between days of life 22 and 99, while one sample is sufficient on or before day of life 21. While SVC was more sensitive than RVC, the risk of SVC failure supports the use of multimodality testing to optimize CMV detection in neonates and infants. Though many laboratories perform paired SVC and RVC for the diagnosis of congenital CMV, the use of PCR has emerged, including the testing of urine and saliva [14,27]. de Vries and colleagues [14] compared CMV PCR with SVC on neonatal urine and demonstrated a trend towards increased PCR sensitivity, though the difference did not reach statistical significance. Due to the length of time RVC takes to develop CPE, PCR may be a useful and timely replacement of this conventional method. Furthermore, a multimodal testing approach should be considered given that prompt and accurate diagnosis of congenital CMV infection will allow for appropriate therapeutic management and follow-up.
CMV SERIAL URINE VIRAL CULTURE
179
Future studies evaluating culture and nucleic acid amplification methods for identification of CMV infection in neonates and infants may prove useful for the development of optimal diagnostic strategies. ACKNOWLEDGMENTS We thank the staff of the Stanford Clinical Virology Laboratory for their continued exceptional work and dedication. We also thank Sharon Chen for her critical reading of the manuscript. REFERENCES 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. 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. 3. Boppana SB, Pass RF, Britt WJ, Stagno S, Alford CA. Symptomatic congenital cytomegalovirus infection: neonatal morbidity and mortality. Pediatr Infect Dis J 1992;11:93–99. 4. Lombardi G, Garofoli F, Stronati M. Congenital cytomegalovirus infection: treatment, sequelae and follow-up. J Matern Fetal Neonatal Med 2010;23(suppl 3):45–48. 5. Coll O, Benoist G, Ville Y, et al. Guidelines on CMV congenital infection. J Perinat Med 2009;37:433–445. 6. Stagno S, Britt W. Cytomegalovirus infections. In: Remington JS, Klein JO, Wilson CB, Baker CJ, eds. Infectious Diseases of the Fetus and Newborn Infant, 6th ed. Philadelphia, PA: Saunders, 2006;739–781. 7. Grosse SD, Ross DS, Dollard SC. Congenital cytomegalovirus (CMV) infection as a cause of permanent bilateral hearing loss: a quantitative assessment. J Clin Virol 2008;41:57–62. 8. Ross SA, Novak Z, Pati S, Boppana SB. Overview of the diagnosis of cytomegalovirus infection. Infect Disord Drug Targets 2011;11: 466–474. 9. Revello MG, Gerna G. Diagnosis and management of human cytomegalovirus infection in the mother, fetus, and newborn infant. Clin Microbiol Rev 2002;15:680–715. 10. Kurath S, Halwachs-Baumann G, Muller W, Resch B. Transmission of cytomegalovirus via breast milk to the prematurely born infant: a systematic review. Clin Microbiol Infect 2010;16:1172–1178. 11. Gleaves CA, Smith TF, Shuster EA, Pearson GR. Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J Clin Microbiol 1985;21: 217–221. 12. Alpert G, Mazeron MC, Colimon R, Plotkin S. Rapid detection of human cytomegalovirus in the urine of humans. J Infect Dis 1985; 152:631–633.
180
CHISHOLM ET AL.
13. Thiele GM, Bicak MS, Young A, Kinsey J, White RJ, Purtilo DT. Rapid detection of cytomegalovirus by tissue culture, centrifugation, and immunofluorescence with a monoclonal antibody to an early nuclear antigen. J Virol Methods 1987;16:327–338. 14. de Vries JJ, van der Eijk AA, Wolthers KC, et al. Real-time PCR versus viral culture on urine as a gold standard in the diagnosis of congenital cytomegalovirus infection. J Clin Virol 2012;53:167– 170. 15. Kimberlin DW, Lin CY, Sanchez PJ, et al. 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. 16. Shin JJ, Keamy DG Jr, Steinberg EA. Medical and surgical interventions for hearing loss associated with congenital cytomegalovirus: a systematic review. Otolaryngol Head Neck Surg 2011; 144:662–675. 17. American Academy of Pediatrics JCoIH. Year 2007 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics 2007;120:898–921. 18. Stagno S, Reynolds DW, Tsiantos A, Fuccillo DA, Long W, Alford CA. Comparative serial virologic and serologic studies of symptomatic and subclinical congenitally and natally acquired cytomegalovirus infections. J Infect Dis 1975;132:568–577. 19. Reynolds DW, Stagno S, Hosty TS, Tiller M, Alford CA Jr. Maternal cytomegalovirus excretion and perinatal infection. N Engl J Med 1973;289:1–5. 20. Nijman J, van Loon AM, de Vries LS, et al. Urine viral load and correlation with disease severity in infants with congenital or postnatal cytomegalovirus infection. J Clin Virol 2012;54:121–124. 21. de Vries JJ, Barbi M, Binda S, Claas EC. Extraction of DNA from dried blood in the diagnosis of congenital CMV infection. Methods Mol Biol 2012;903:169–175. 22. Folkins AK, Chisholm KM, Guo FP, McDowell M, Aziz N, Pinsky BA. Diagnosis of congenital CMV using PCR performed on formalin-fixed, paraffin-embedded placental tissue. Am J Surg Pathol 2013;37:1413–1420. 23. Myers JB, Amsterdam D. The laboratory diagnosis of cytomegalovirus infections. Immunol Invest 1997;26:383–394. 24. Stirk PR, Griffiths PD. Use of monoclonal antibodies for the diagnosis of cytomegalovirus infection by the detection of early antigen fluorescent foci (DEAFF) in cell culture. J Med Virol 1987; 21:329–337. 25. Paya CV, Wold AD, Ilstrup DM, Smith TF. Evaluation of number of shell vial cell cultures per clinical specimen for rapid diagnosis of cytomegalovirus infection. J Clin Microbiol 1988;26:198–200. 26. Leland DS, Hansing RL, French ML. Clinical experience with cytomegalovirus isolation using conventional cell cultures and early antigen detection in centrifugation-enhanced shell vial cultures. J Clin Microbiol 1989;27:1159–1162. 27. Boppana SB, Ross SA, Shimamura M, et al. Saliva polymerasechain-reaction assay for cytomegalovirus screening in newborns. N Engl J Med. 2011;364:2111–2118.
Evaluation of Serial Urine Viral Cultures for the Diagnosis of Cytomegalovirus Infection in Neonates and Infants KAREN M. CHISHOLM,1 NATALI AZIZ,2 MICHAL MCDOWELL,2 FRANCES P. GUO,1 NIVEDITA SRINIVAS,3 WILLIAM E. BENITZ,3 MARY E. NORTON,2 KATHLEEN GUTIERREZ,3 ANN K. FOLKINS,1 AND BENJAMIN A. PINSKY1,4* 1
Department Department 3 Department 4 Department CA, USA 2
of Pathology, Stanford University School of Medicine, Stanford, CA, USA of Obstetrics and Gynecology, Stanford University School of Medicine, Stanford, CA, USA of Pediatrics, Stanford University School of Medicine, Stanford, CA, USA of Medicine, Division of Infectious Diseases and Geographic Medicine, Stanford University School of Medicine, Stanford,
Received January 19, 2014; accepted March 5, 2014; published online March 11, 2014.
ABSTRACT Cytomegalovirus (CMV) is the most common cause of congenital infection worldwide. Urine viral culture is the standard for CMV diagnosis in neonates and infants. The objectives of this study were to compare the performance of serial paired rapid shell vial cultures (SVC) and routine viral cultures (RVC), and to determine the optimal number of cultures needed to detect positive cases. From 2001 to 2011, all paired CMV SVC and RVC performed on neonates and infants less than 100 days of age were recorded. Testing episodes were defined as sets of cultures performed within 7 days of one another. A total of 1264 neonates and infants underwent 1478 testing episodes; 68 (5.4%) had at least one episode with a positive CMV culture. In episodes where CMV was detected before day 21 of life, the first specimen was positive in 100% (16/16) of cases. When testing occurred after 21 days of life, the first specimen was positive in 82.7% (43/52) of cases, requiring three cultures to reach 100% detection. The SVC was more prone to assay failure than RVC. Overall, when RVC was compared to SVC, there was 86.0% positive agreement and 99.9% negative agreement. In conclusion, three serial urine samples are necessary for detection of CMV in specimens collected between day of life 22 and 99, while one sample may be sufficient on or before day of life 21. Though SVC was more sensitive than RVC, the risk of SVC failure supports the use of multimodality testing to optimize detection.
This work was presented at the Society for Pediatric Pathology Fall Meeting in Salt Lake City, UT, USA, September 27, 2013. *Corresponding author, e-mail: [email protected]
Key words: cytomegalovirus, hearing loss, sensorineural, virus cultivation
INTRODUCTION Cytomegalovirus (CMV) is the most common congenitally acquired viral infection worldwide, with a birth prevalence of 0.6 to 0.7% [1,2]. Approximately 10% to 15% of neonates are symptomatic at birth, with signs and symptoms including jaundice, hepatomegaly, splenomegaly, thrombocytopenia, microcephaly, and low birth weight [2–5]. Of these symptomatic infants, 40% to 90% will later develop permanent disabilities, including hearing loss, vision loss, and cognitive impairments [2– 6]. Of the 85% to 90% of infants who are asymptomatic at birth, 10% to 15% will develop permanent disabilities, predominantly hearing loss, but also cognitive or neurologic impairments [2,4,6]. Overall, congenital CMV infection is the most common nonhereditary cause of sensorineural hearing loss, affecting 10% to 15% of congenitally infected CMV individuals, and accounting for 15% to 20% of cases of congenital hearing loss [7]. When congenital infection is suspected, viral culture is performed ideally within the first two to three weeks of life [5,6,8,9]. Otherwise, differentiating congenital from early postnatal infection is difficult. Postnatal transmission can occur due to exposure to maternal genital tract secretions during delivery, maternal breast milk ingestion, and transfusions with CMV seropositive blood [8]. While postnatal infection may be asymptomatic, it also can cause a septic syndrome that may include hepatomegaly, splenomegaly, thrombocytopenia, and pneumonia [5]. However, compared to those with congenital CMV, infants with postnatal CMV exposure are not at risk for
sensorineural hearing loss or other neurologic or cognitive deficiencies [10]. The standard methods for the diagnosis of CMV infection in neonates and infants are urine or saliva viral culture [8,9]. Detection of CMV can be accomplished through routine viral culture (RVC) methods or centrifugation-enhanced rapid shell vial cultures (SVC). The RVC has lengthy turn-around times, taking usually 1 to 2 weeks, but sometimes as long as 4 to 6 weeks, to develop cytopathic effect (CPE). The SVC methods have been found to be as specific as, and even more sensitive than, conventional viral cultures and require only 12 to 72 hours [11–13]. However, neonatologists and pediatricians often order multiple cultures to ensure CMV detection in infected infants. Real-time polymerase chain reaction (PCR) also may be used for CMV detection in neonatal urine, though neither sensitivity nor specificity show a statistically significant difference compared to SVC [14]. In patients with congenital CMV infection, treatment with ganciclovir can prevent hearing deterioration and may improve neurodevelopmental outcomes [15,16]. As such, prompt and accurate diagnosis of CMV infection is critical for appropriate evaluation and timely management.
METHODS Patient selection With Stanford University Medical Center Institutional Review Board approval, data on all urine viral cultures performed on infants under 100 days of age, from January 2001 through December 2011, were collected. Inclusion required that specimens were submitted for paired CMV SVC and RVC. Patient name, date of birth, medical record number, and date of cultures were recorded to tabulate and group cultures by patient. Testing episodes were defined as sets of paired cultures that were performed within 7 days of one another. Viral cultures Urine specimens were collected in sterile containers and stored at 2u to 8uC before testing. For RVC, urine specimens were mixed 1:2 with antimicrobial mix containing gentamicin (150 mg/mL), vancomycin (1000 mg/mL), and amphotericin B (35 mg/mL). Of the processed urine, 0.25 mL was inoculated onto each traditional tube culture monolayer including human foreskin (HF) fibroblasts (Diagnostic Hybrids, Athens, OH, USA) and human embryonic lung MRC-5 fibroblasts (Viromed, Minnetonka, MN, USA) containing 1.0 to 1.5 mL of the manufacturer’s culture medium. The RVCs were incubated for 28 days at 35uC. Cell monolayers with demonstrable CPE were scraped, washed with phosphate buffered saline (PBS; Sigma-Aldrich, St. Louis, MO, USA), spotted to Cel-Line Supercured HTC slides (Thermo Fisher Scientific, Waltham, MA, USA), and fixed for 10 minutes in acetone. The cells then were stained for the presence of CMV by indirect immunofluorescence using anti-CMV antibodies targeting the immediate early antigen (Millipore, Billerica, MA,
Table 1.
Number of specimens submitted per episode
Specimens per episode
Episodes
%
1 2 3 4 5 6 Total
649 235 512 72 6 4 1478
43.9 15.9 34.6 4.9 0.4 0.3 100%
USA) according to the manufacturer’s specifications. The number of days to identifiable CPE was noted for all positive cultures. For SVC, two human fibroblast vials (one HF and one MRC-5 from the vendors described above) were each inoculated with 0.25 mL processed urine after aspiration of the culture medium. The vials were centrifuged at 3000 rpm for 30 minutes, the monolayers refed, and the vials incubated at 35uC. At 24 hours, the cell monolayer from one vial (HF) was washed with PBS, fixed with acetone, and stained as above with the same antibody reagent used for RVC. The other vial was processed (MRC-5) at 48 hours. The SVC results were reported only if the monolayers were .60% confluent, the negative control showed no specific fluorescence, and the positive control, cultured CMV AD169 reference strain (American Type Culture Collection, Manassas, VA, USA), exhibited the expected apple green nuclear fluorescence. A positive SVC result on a patient sample was reported if at least one cell demonstrated this characteristic staining. The SVCs that demonstrated intense nonspecific fluorescence or excessive destruction of the monolayer by specimen toxicity or contamination were reported as unsatisfactory. When either paired SVC or RVC was positive, the specimen was considered CMV culture positive. Statistics Contingency tables were evaluated with Fisher’s exact test, 2-tailed, and the Mann-Whitney test was used to analyze days to CPE in RVC (GraphPad, La Jolla, CA, USA). In addition, Student’s t-test was used for comparing positive and negative episodes (Excel; Microsoft, Redmond, WA, USA). Positive percent agreement and negative percent agreement were calculated for SVC and RVC concordance, and evaluated by McNemar’s test.
RESULTS Culture results Over the course of the study, 2997 urine specimens from 1264 neonates and infants less than 100 days of age were submitted for CMV detection by paired RVC and SVC. Testing episodes, defined as sets of paired cultures that were performed within 7 days of one another, totaled 1478, with a range from 1 to 6 specimens per episode, and a mean of 2 specimens per episode. Table 1 details the
CMV SERIAL URINE VIRAL CULTURE
177
Table 2. Shell vial culture and routine viral culture concordance Shell vial culture
Routine vial culture Positive Negative
Positive
Negative
98 16
2 2873
A total of 7 paired shell vial cultures had equipment failure, toxic interference, quality assurance failure, or laboratory error for 24- and 48hour shell vials, and 1 routine viral culture had laboratory error. The total number of paired cultures with available results was 2989.
Figure 1. Comparison of days to cytopathic effect (CPE) for specimens collected on or before 21 days of life and specimens collected after 21 days of life (Mann-Whitney test, P , 0.0001). The solid horizontal lines indicate the median number of days to CPE.
number of specimens per episode submitted for paired RVC and SVC. Of all neonates and infants, 68 (5.4%) were positive for CMV. Compared to positive infants (range of tests per episode, 1–4; mean, 1.92), the number of tests per episode in negative infants (range of tests per episode, 1–6; mean, 2.03) was not statistically significant (Student t-test, P 5 0.347). Of the 68 CMV-positive infants, 16 (23.5%) had positive cultures on or before 21 days of life. These 16 neonates had specimens obtained between days 1 and 20 of life. Of these neonates 14 were cultured at less than 7 days old, while the other two neonates were 17 and 20 days of age. For all 16 of these neonates, the first urine culture submitted of the episode always was positive, 100% (16/16). In addition, in all of these cases, the 24hour SVC and the RVC from the first specimen always were positive. Due to multiple specimens submitted, a total of 28 cultures was performed in these infants. Only two specimens were SVC and RVC negative, both the second specimens submitted for two neonates. For the remaining 26 cultures, SVC and RVC were positive in all pairs. As shown in Figure 1, the number of days to identifiable CPE for RVC ranged from 1 to 18, with a mean of 4.3 days and median of 2.5. A total of 52 infants of 68 (76.5%) had positive cultures after 21 days of life, ranging from day 25 to day 99 of life. In these infants, the first specimen of each episode was positive in 82.7% (43/52) of cases. In 11.5% (6/52) of cases, CMV was not detected until the second specimen, and in 5.8% (3/52) of cases CMV was not detected until the third specimen. Of the infants’ first positive specimens, 76.9% (40/52) of culture pairs demonstrated SVC and RVC positivity, while only SVC
178
CHISHOLM ET AL.
was positive in 21.1% (11/52) and only RVC was positive in 1.9% (1/52). In these 52 infants, 54 total episodes were positive, since two infants had two episodes each that were positive. For these 54 episodes, a total of 107 urine specimens was submitted, with negative culture pairs accounting for 15.9% (17/107). For all positive RVCs, days to identifiable CPE ranged from 1 to 28, with a mean of 9.6 days and median of 9 days. As depicted in Figure 1, positive RVCs in infants who were CMV-positive at less than 21 days of life took significantly less time to develop CPE than RVCs performed in infants more than 21 days of life (Mann-Whitney test, P , 0.0001). Culture failure rates All 2997 urine specimens included in this study were tested by RVC and SVC. One RVC was not reported due to laboratory error (1/2997, 0.03%). Of the 2997 study specimens, 2972 had SVCs read at both 24 and 48 hours, while 19 had SVCs read only at 48 hours, and 6 had SVCs read only at 24 hours. Of the total 5969 SVCs performed, results were not reported in 109 due to toxic interference (30 in the 24-hour and 79 in the 48-hour SVC, 1.8%), 40 due to quality assurance failure (21 in the 24-hour and 19 in the 48-hour SVC, 0.67%), 28 due to laboratory error (10 in the 24-hour and 18 in the 48-hour SVC, 0.47%), and two due to laboratory equipment failure (1 in the 24hour and 1 in the 48-hour SVC, 0.03%). As such, SVC was more prone to failure (179/5969, 3.0%) than RVC (1/2997, 0.03%, Fisher’s exact test, P , 0.0001). SVC and RVC concordance Of the 2997 specimens, CMV was detected in 116 specimens from SVC and/or RVC. Eight specimens had unavailable results (due to toxic interference, quality assurance failure, laboratory error, or equipment failure) for 24- and 48-hour SVC or RVC. Using SVC as reference, the positive percent agreement between SVC and RVC was 86.0% (98/114), and the negative percent agreement was 99.9% (2873/2875, Table 2). The results of the SVC and RVC tests were significantly different (McNemar’s test, P 5 0.0013). All discordant specimens were from infants more than 21 days of age (21 versus .21 days of age, Fisher’s exact test, P 5 0.0118). The
two cultures that were only positive by RVC occurred in the same patient; this patient had two other specimens submitted in the same episode that were positive by RVC and SVC, suggesting that these were not false-positive RVCs, but false-negative SVCs. Of the 16 specimens in which only SVCs were positive, 9 specimens were taken in an episode in which other specimens also were culturepositive, 2 specimens were from infants with subsequent CMV-positive urine specimens by qualitative PCR, and 1 specimen was from an infant with a CMV-positive nasopharyngeal swab by RVC; these findings suggested that these positive SVCs were not false-positives, but that the RVCs were false-negatives. The remaining 4 positive SVC specimens were either the only specimens collected in an episode [2], or the patient had a previous specimen that was negative by both SVC and RVC [2], and, hence, the accuracy of the result could not be confirmed or refuted.
DISCUSSION Congenital CMV is an important diagnosis due to the potential attenuation of long-term sequelae with timely ganciclovir therapy, and supportive management with more frequent audiology and neurodevelopmental evaluations [15,17]. The current study showed that a single urine sample may sufficiently diagnose cases of congenital CMV infection. For all neonates with urine samples submitted on or before 21 days of life, the first SVC and the first RVC were positive. While RVC in some cases required more than 10 days to show CPE (range, 1– 18 days), the 24-hour SVC was positive in 100% of first urine specimens. Conversely, neonates and infants with urine submitted after 21 days of life needed up to three urine samples to reach 100% detection. Only 82.7% of cases had first specimen positivity, 23.1% of which had discordance between SVC and RVC results. In addition, the 24-hour SVC was negative in some of these cases, though the 48hour SVC was positive. Also, the RVC took significantly longer to develop CPE compared to specimens collected from neonates less than 21 days of life. The finding of more consistent CMV detection between SVC and RVC in those patients diagnosed on or before 21 days of life compared to those diagnosed after 21 days of life, as well as shorter time intervals to CPE in RVC, is consistent with previous studies. Congenitally infected neonates, whether they are symptomatic or not, excrete virus at birth [18]. In those with postnatal CMV infection, viral excretion begins three to 12 weeks after birth [18,19]. In addition, the CMV urine viral load is known to be higher in congenital infection compared to postnatal infection [18,20]. Thus, earlier time to positivity is expected in congenital infection, as is longer time to positivity and decreased concordance between cultures in postnatal infection. Of note, while the 21-day threshold increases the likelihood that a CMV infection was acquired congeni-
tally, it is not absolute. In this study, one neonate was identified whose first CMV testing episode included two negative cultures, one on day 1 and another on day 4 of life, before a positive second testing episode occurring on day 17 of life, consistent with postnatal infection. Furthermore, the positive RVC from the second testing episode did not demonstrate CPE until day 18. This case illustrates the importance of early CMV testing to help rule out congenital CMV infection. As of yet, widespread newborn screening for congenital CMV has not been implemented. In the event that timely testing is not performed, some groups have proposed to look back into the neonatal or fetal time period to test for CMV. For example, dried blood spots collected at birth or formalinfixed paraffin-embedded placental tissues have been used to make the diagnosis of congenital CMV [21,22]. Of the positive specimens, SVC identified 16 cases that were negative by RVC, while RVC identified 2 cases that were negative by SVC. Previous studies have shown that SVC more often is positive than the paired RVC, though the current study is the first to evaluate these methods exclusively in urine from neonates and infants. From urine collected from a variety of patient groups, Gleaves and colleagues [11] identified 32/109 specimens in which SVC was positive, but RVC was negative. Similarly, Myers and Amsterdam [23] identified 55 specimens that were SVC-positive, but RVC-negative in patients with a history of renal transplant or immunodeficiency; urine specimens accounted for 28 of these 55 specimens. While SVC was more often CMV-positive in this study than RVC, this method also was more prone to toxic interference, quality assurance failure, or laboratory error. However, the 3% SVC failure rate in this study is not large, especially in comparison with historical rates of 4% to 10.7% [24–26]. The rate of failure and the sensitivity of the assay led Paya and colleagues [25] to recommend at least two vials be inoculated per urine specimen, as was performed in this study, for maximum detection of CMV. In this study, we demonstrated that three serial urine samples are necessary for full detection of CMV in specimens collected between days of life 22 and 99, while one sample is sufficient on or before day of life 21. While SVC was more sensitive than RVC, the risk of SVC failure supports the use of multimodality testing to optimize CMV detection in neonates and infants. Though many laboratories perform paired SVC and RVC for the diagnosis of congenital CMV, the use of PCR has emerged, including the testing of urine and saliva [14,27]. de Vries and colleagues [14] compared CMV PCR with SVC on neonatal urine and demonstrated a trend towards increased PCR sensitivity, though the difference did not reach statistical significance. Due to the length of time RVC takes to develop CPE, PCR may be a useful and timely replacement of this conventional method. Furthermore, a multimodal testing approach should be considered given that prompt and accurate diagnosis of congenital CMV infection will allow for appropriate therapeutic management and follow-up.
CMV SERIAL URINE VIRAL CULTURE
179
Future studies evaluating culture and nucleic acid amplification methods for identification of CMV infection in neonates and infants may prove useful for the development of optimal diagnostic strategies. ACKNOWLEDGMENTS We thank the staff of the Stanford Clinical Virology Laboratory for their continued exceptional work and dedication. We also thank Sharon Chen for her critical reading of the manuscript. REFERENCES 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. 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. 3. Boppana SB, Pass RF, Britt WJ, Stagno S, Alford CA. Symptomatic congenital cytomegalovirus infection: neonatal morbidity and mortality. Pediatr Infect Dis J 1992;11:93–99. 4. Lombardi G, Garofoli F, Stronati M. Congenital cytomegalovirus infection: treatment, sequelae and follow-up. J Matern Fetal Neonatal Med 2010;23(suppl 3):45–48. 5. Coll O, Benoist G, Ville Y, et al. Guidelines on CMV congenital infection. J Perinat Med 2009;37:433–445. 6. Stagno S, Britt W. Cytomegalovirus infections. In: Remington JS, Klein JO, Wilson CB, Baker CJ, eds. Infectious Diseases of the Fetus and Newborn Infant, 6th ed. Philadelphia, PA: Saunders, 2006;739–781. 7. Grosse SD, Ross DS, Dollard SC. Congenital cytomegalovirus (CMV) infection as a cause of permanent bilateral hearing loss: a quantitative assessment. J Clin Virol 2008;41:57–62. 8. Ross SA, Novak Z, Pati S, Boppana SB. Overview of the diagnosis of cytomegalovirus infection. Infect Disord Drug Targets 2011;11: 466–474. 9. Revello MG, Gerna G. Diagnosis and management of human cytomegalovirus infection in the mother, fetus, and newborn infant. Clin Microbiol Rev 2002;15:680–715. 10. Kurath S, Halwachs-Baumann G, Muller W, Resch B. Transmission of cytomegalovirus via breast milk to the prematurely born infant: a systematic review. Clin Microbiol Infect 2010;16:1172–1178. 11. Gleaves CA, Smith TF, Shuster EA, Pearson GR. Comparison of standard tube and shell vial cell culture techniques for the detection of cytomegalovirus in clinical specimens. J Clin Microbiol 1985;21: 217–221. 12. Alpert G, Mazeron MC, Colimon R, Plotkin S. Rapid detection of human cytomegalovirus in the urine of humans. J Infect Dis 1985; 152:631–633.
180
CHISHOLM ET AL.
13. Thiele GM, Bicak MS, Young A, Kinsey J, White RJ, Purtilo DT. Rapid detection of cytomegalovirus by tissue culture, centrifugation, and immunofluorescence with a monoclonal antibody to an early nuclear antigen. J Virol Methods 1987;16:327–338. 14. de Vries JJ, van der Eijk AA, Wolthers KC, et al. Real-time PCR versus viral culture on urine as a gold standard in the diagnosis of congenital cytomegalovirus infection. J Clin Virol 2012;53:167– 170. 15. Kimberlin DW, Lin CY, Sanchez PJ, et al. 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. 16. Shin JJ, Keamy DG Jr, Steinberg EA. Medical and surgical interventions for hearing loss associated with congenital cytomegalovirus: a systematic review. Otolaryngol Head Neck Surg 2011; 144:662–675. 17. American Academy of Pediatrics JCoIH. Year 2007 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics 2007;120:898–921. 18. Stagno S, Reynolds DW, Tsiantos A, Fuccillo DA, Long W, Alford CA. Comparative serial virologic and serologic studies of symptomatic and subclinical congenitally and natally acquired cytomegalovirus infections. J Infect Dis 1975;132:568–577. 19. Reynolds DW, Stagno S, Hosty TS, Tiller M, Alford CA Jr. Maternal cytomegalovirus excretion and perinatal infection. N Engl J Med 1973;289:1–5. 20. Nijman J, van Loon AM, de Vries LS, et al. Urine viral load and correlation with disease severity in infants with congenital or postnatal cytomegalovirus infection. J Clin Virol 2012;54:121–124. 21. de Vries JJ, Barbi M, Binda S, Claas EC. Extraction of DNA from dried blood in the diagnosis of congenital CMV infection. Methods Mol Biol 2012;903:169–175. 22. Folkins AK, Chisholm KM, Guo FP, McDowell M, Aziz N, Pinsky BA. Diagnosis of congenital CMV using PCR performed on formalin-fixed, paraffin-embedded placental tissue. Am J Surg Pathol 2013;37:1413–1420. 23. Myers JB, Amsterdam D. The laboratory diagnosis of cytomegalovirus infections. Immunol Invest 1997;26:383–394. 24. Stirk PR, Griffiths PD. Use of monoclonal antibodies for the diagnosis of cytomegalovirus infection by the detection of early antigen fluorescent foci (DEAFF) in cell culture. J Med Virol 1987; 21:329–337. 25. Paya CV, Wold AD, Ilstrup DM, Smith TF. Evaluation of number of shell vial cell cultures per clinical specimen for rapid diagnosis of cytomegalovirus infection. J Clin Microbiol 1988;26:198–200. 26. Leland DS, Hansing RL, French ML. Clinical experience with cytomegalovirus isolation using conventional cell cultures and early antigen detection in centrifugation-enhanced shell vial cultures. J Clin Microbiol 1989;27:1159–1162. 27. Boppana SB, Ross SA, Shimamura M, et al. Saliva polymerasechain-reaction assay for cytomegalovirus screening in newborns. N Engl J Med. 2011;364:2111–2118.
