Evaluation of Cytomegalovirus Infections Transmitted via Breast Milk in Preterm Infants With a Real-Time Polymerase Chain Reaction Assay
Objective. Preterm infants are at greater risk of symptomatic cytomegalovirus (CMV) infection than term infants. Breast milk is the main source of perinatal CMV infections. This study evaluated the kinetics of CMV load in breast milk and the rate of postnatal CMV transmission via breast milk from mothers to their preterm infants.
Methods. This was a prospective study of 30 mothers and their 43 preterm infants. The infants either had a gestational age of <34 weeks or weighed <2000 g at birth. Breast milk, serum, and urine samples were collected every 2 weeks until discharge, and screened for CMV infection using a real-time PCR assay. Most of the breast milk had been preserved at −20°C before feeding to the preterm infants.
Results. Twenty-four mothers (24 of 30, 80%), who had 34 preterm infants, were CMV immunoglobulin G positive. Twenty-one (87.5%) of the 24 seropositive mothers, who had 30 preterm infants, had detectable CMV deoxyribonucleic acid (DNA) in breast milk during the study period. Most breast milk became positive for CMV DNA 2 weeks after delivery. Viral DNA copy numbers increased until they peaked at 4 to 6 weeks. Afterward, the CMV DNA copy numbers decreased. Of the 30 infants who were fed CMV DNA-positive breast milk, CMV infection was confirmed in 3 infants. However, they had no clinical symptoms of CMV infection.
Conclusions. Despite the high rate of CMV DNA in breast milk, symptomatic infections in the preterm infants did not occur. These results might be associated with the method of breast milk preservation and the population we studied. CMV infections transmitted via breast milk feeding did not have much impact on preterm infants in our institutes.
Cytomegalovirus (CMV) is the most common cause of intrauterine and perinatal infections worldwide.1 Perinatal infections occur during exposure to genital secretions at birth or via breast milk. CMV shed into breast milk is the main source of CMV infection during the first year of life.2 In contrast to congenital infections, CMV infections transmitted via breast milk are usually asymptomatic, especially in term infants.2,3 Recently, Hamprecht et al4 reported that the transmission rate of CMV from seropositive mothers to preterm infants via breast milk was 37% in Germany. In their study, half of the infected infants had neutropenia, thrombocytopenia, or hepatopathy, and 4 infants had sepsis-like symptoms.5 If their observations were correct, avoiding breastfeeding or inactivating the virus in breast milk would be necessary for preterm infants. However, postnatal symptomatic CMV infections in preterm infants are seldom observed in Japan, except for transfusion-acquired infections, despite the fact that a relatively high proportion of Japanese women are seropositive for CMV.6,7
To assess the differences in the rate of CMV transmission via breast milk in Germany and Japan, we performed a prospective study using a real-time polymerase chain reaction (PCR) assay. The real-time PCR assay is a useful technique that can be used to process large numbers of samples rapidly, and it allows direct measurement of the viral load without handling the tube in the postamplification steps, where contamination of deoxyribonucleic acid (DNA) carryover can occur.8 The CMV load in infants’ urine, serum, and breast milk can be quantified using the real-time PCR assay, which is a rapid, sensitive system with a wide dynamic range.9
This study evaluated the kinetics of CMV load in breast milk and the rate of CMV transmission via breast milk from mothers to their preterm infants.
The study population consisted of preterm infants and their mothers who were admitted to Nagoya University Hospital, the Japanese Red Cross Nagoya First Hospital, and Ogaki Municipal Hospital, between June 1, 2001, and March 31, 2002. The study included preterm infants with a gestational age of <34 complete weeks or with a weight of <2000 g at birth. Infants with congenital CMV infection, that is, infants in whom CMV DNA was detected in a urine or serum sample during the first week postpartum, and those who had other congenital viral infections, severe congenital anomaly syndrome, congenital heart disease, or hydrops fetalis, as well as infants who were fed on donor breast milk, were excluded from this study. Eighty-two infants were admitted in this age category during the study period. Twenty-six infants were excluded from the study for the following reasons: 17 were fed donor breast milk, 5 refused enrollment, congenital viral infection in 1, congenital heart disease in 1, hydrops fetalis in 1, and fed only formula milk in 1. The remaining 56 infants and their 42 mothers were enrolled, but 13 infants and their 12 mothers withdrew from the study. Finally, 43 preterm infants (7 twins, 3 triplets) and their 30 mothers were analyzed.
Enteral feeding of the preterm infants was started when their general condition stabilized. Raw breast milk was sometimes used when their mothers were hospitalized (for 6–11 days). About half of the infants received raw breast milk at least once. After the infants’ mothers were discharged, all their breast milk was preserved at −20°C. The infants were given supplemental formula milk when there was insufficient breast milk.
Peripheral blood and urine were collected from the infants at birth and every 2 weeks until discharge. In addition, colostrum was collected from their mothers. Breast milk was collected every 2 weeks until the infants were discharged. For CMV antibody tests, peripheral blood was taken from the mothers during the first week postpartum. Peripheral blood was centrifuged and serum was prepared. If the infants received transfusions during the study period, some of the blood products were also collected for the assay. All samples were frozen at −20°C.
During the study period, the infants’ clinical symptoms (abdominal distension, apnea-bradycardia, hepatomegaly, and petechiae) were carefully evaluated every day by their respective attending doctors. In addition, blood samples were taken regularly every week, and the following laboratory parameters were assessed: white blood cell count, absolute number of neutrophils, thrombocyte count, and liver enzyme activity. These evaluations and laboratory tests were scheduled and performed without informing the attending doctors of the results of the breast milk PCR tests.
CMV immunoglobulin (Ig)G and IgM antibodies were determined using enzyme immunoassay kits (Denka Seiken Inc, Tokyo, Japan).
Preparation of DNA
DNA was extracted from 200 μL of the infant’s serum, blood product, or unfractionated breast milk using a QIAamp Blood kit (Qiagen, Hilden, Germany), and eluted in 100 μL of distilled water. DNA was extracted from 140 μL of the infant’s urine using a QIAamp Viral RNA kit (Qiagen) according to the manufacturer’s instructions.
Real-Time Quantitative PCR Assay
The PCR primers and probe used for this assay were from the immediate early (IE) gene. Real-time PCR was conducted using a TaqMan PCR kit (PE Applied Biosystems, Foster City, CA), as described previously.9,10 Briefly, 10 μL of the solution containing DNA extracted from the samples were added to the PCR mixture. After activation of AmpliTaq Gold for 10 minutes at 95°C, 50 cycles of 15 seconds at 95°C, and 1 minute at 62°C were conducted in a Model 7700 sequence detector (PE Applied Biosystems). Real-time fluorescence measurements were taken and the threshold cycle (CT) value for each sample was calculated by determining the point at which the fluorescence exceeded a threshold limit (10 × the standard deviation of the base line). pGEM-IE, which contains the IE gene, was used as a positive control.9 A standard graph was constructed; the CT values from the clinical samples were plotted on the standard curve, and the copy number was calculated automatically using Sequence Detector version 1.6 (PE Applied Biosystems), a software package for data analysis. Samples were defined as negative when the CT values exceeded 50 cycles. The DNA copy numbers of the serum, urine, and breast milk were expressed per milliliter. The detection limit of this assay was ∼100 copies per milliliter.
The ethics committee of each hospital approved this study, and informed consent was obtained from every mother.
Thirty mothers and their 43 preterm infants (7 twins, 3 triplets) met the entry criteria. The median gestational age at birth was 31.1 weeks (range 27.2–36.7 weeks) and the median birth weight was 1360 g (range 714-1827 g). Ten infants were delivered vaginally and 33 were delivered by cesarean section. The median length of hospital stay was 64 days (range 23–105 days). The median number of infant serum/urine samples, and breast milk samples was 4 (range 2–7) and 3 (range 2–7), respectively.
Twenty-four mothers (24 of 30, 80%) were positive for CMV IgG antibodies. In 21 (21 of 24, 87.5%) of these seropositive mothers, CMV DNA was detected in breast milk during the study period. In all 6 seronegative mothers, the breast milk was negative for CMV DNA, and CMV DNA was not detected in serum or urine samples from their 9 infants during the observation period.
The CMV DNA copy numbers in breast milk measured using the real-time PCR assay are shown in Fig 1. Most of the colostrum samples, which were collected within 1 week of delivery, were negative for CMV DNA (number positive/number tested = 2 of 10, 20%). Most breast milk samples became positive for CMV DNA 2 weeks after delivery (number positive/number tested = 17 of 18, 94%) and the viral DNA copy numbers increased until 4 to 6 weeks after delivery, when the quantity of CMV DNA peaked at 104∼106 copies/mL. Subsequently, the viral DNA copy numbers decreased.
Of the 30 infants who were fed CMV DNA-positive breast milk, 3 (10%) had detectable CMV DNA. The clinical and virologic characteristics of these infants are shown in Table 1. None of the 3 infants had clinical symptoms associated with CMV infection. The maximum CMV DNA copy numbers in their breast milk samples were 9.3 × 104, 6.8 × 104, and 8.3 × 103 copies/mL, respectively. There were 12 noninfected infants whose maximum CMV DNA copy number in breast milk exceeded these 3 values. Five infants, including 1 infected infant, received blood transfusions during the study period. The blood products were examined for CMV DNA using the real-time PCR assay, but none was detected.
In this study, we detected CMV DNA in breast milk from 21 (87.5%) of 24 seropositive mothers. In previous studies, the rate of virolactia in unfractionated milk ranged from 9% to 32%.2,3 Several studies have suggested that cell-free milk whey might play a more important role in CMV infection via breast milk,1,4,11,12 and detected CMV DNA most reliably from milk whey, and not from unfractionated milk or milk cells. It is thought that the lipid component of unfractionated milk inhibits PCR.11 Although we used unfractionated milk, the rate of CMV DNA detection was high. Our preliminary data showed that using unfractionated milk did not inhibit PCR (data not shown). Therefore, we did not feel that it was necessary to fractionate breast milk into milk whey and cells in our system.
In our study, CMV DNA was first detected in breast milk 2 weeks after delivery in most cases. The highest levels of secretory IgA are present during the first few days of lactation and IgA may protect against CMV infection by coating and blocking viral attachment sites and by neutralizing viral particles.1 Some cytokines that interfere with viruses, including carrier proteins such as lactoferrin, and some non-Ig antiviral factors are secreted in breast milk. The presence of these substances may be partially responsible for the small amount of CMV shedding into the colostrum and early breast milk.12
We used a real-time PCR assay to quantify CMV DNA. The traditional, qualitative PCR assay is too sensitive to monitor or diagnose CMV-related disease.9 Therefore, it is difficult to decide when the infection occurred. We were able to evaluate the kinetics of CMV DNA in breast milk using the real-time PCR assay. Most of the breast milk samples collected during the first week after delivery were negative for CMV DNA by the real-time PCR assay. Two weeks later, CMV DNA was detectable; the copy number increased until 4 or 6 weeks after delivery and then decreased. The incubation time of CMV infection is between 30 and 120 days.2,4 Therefore, mother-to-infant infections transmitted via breast milk should not occur until at least 6 weeks after delivery. In our 3 infectious cases, 2 infants had detectable CMV DNA in urine 8 and 12 weeks after delivery, but 1 infant had detectable CMV DNA 4 weeks after delivery. Of the 3, only this case was delivered vaginally, and it may have acquired the CMV infection from contact with secretions during the birth process. Furthermore, 1 infected infant who received blood transfusions may have acquired the CMV infection from the blood products. We did not perform a molecular characterization of the PCR products of the 3 infected maternal-infants pairs. If we had, we could have confirmed the CMV transmission via breast milk.
Hamprecht et al4 reported that the transmission of CMV from breastfeeding mothers to their preterm infants could result in symptomatic CMV infections, such as sepsis-like symptoms,5 and that the early onset of symptomatic infections occurred only in extremely immature preterm infants.13 In our study, we observed CMV transmission in only 3 of 30 (10%) preterm infants, although the detection rate and quantity of CMV DNA in breast milk was high (87.5%). None of these cases had clinical symptoms related to CMV infection. The discrepancy between our study and their reports may be associated with the study population. The gestational age and birth weight in our study were somewhat greater than in their study (gestational age: median 31 vs 29 weeks; birth weight: median 1360 vs 1100 g). Further study of extremely premature infants will be necessary to clarify the relationship between breastfeeding and symptomatic CMV infections.
Another possible cause of the discrepancy is the difference in breast milk storage. In our institutes, the breast milk was preserved at −20°C before feeding. Hamprecht et al4 previously reported that 17 of 67 (25%) preterm infants acquired CMV infections. In their unit, the raw breast milk was refrigerated at 4 to 10°C for a maximum of 12 hours before feeding.13 CMV infections have not been observed in CMV seronegative preterm infants fed banked human milk that was either pasteurized or frozen.14,15 Another study failed to detect CMV after heating (72°C for 10 seconds), whereas freezing only partially reduced CMV infectivity, especially when the virus load was high.16 Therefore, the transmission rate of CMV at our institutes might have been affected by the use of frozen milk.
We evaluated the kinetics of CMV DNA in breast milk using a real-time PCR assay. The CMV DNA copy number peaked 4 to 6 weeks after delivery. Although the positive rate of CMV DNA in breast milk was high, no symptomatic infections occurred in the preterm infants that we studied. These results might be caused by the method of breast milk preservation and the population that we studied. CMV infection via breast milk feeding did not have much impact on preterm infants in our institutes.
- Received August 14, 2002.
- Accepted December 12, 2002.
- Address correspondence to Ayako Yasuda, MD, Department of Pediatrics, Nagoya University Graduate School of Medicine, 65, Tsurumai-cho, Shouwa-ku, Nagoya 466-8550, Japan. E-mail:
- ↵Dworsky M, Yow M, Stagno S, Pass RF, Alford CA. Cytomegalovirus infection of breast milk and transsmission in infancy. Pediatrics.1983;72 :295– 299
- ↵Maschen J, Speer CP, Jahn G, et al. Inactivation of cytomegalovirus (CMV) in breast milk. European Society for Pediatric Research. Pediatr Res.1999;45 :924
- Copyright © 2003 by the American Academy of Pediatrics