A collaborative study of cytomegalovirus antibodies in mothers and young children in 19 countries*

Bulletin of the World Health Organization, 59 (4): 605-610 (1981) A collaborative study of cytomegalovirus antibodies in mothers and young children in 19 countries* U. KRECH,1 & J. TOBIN 2 Regional variations in the incidence of cytomegalovirus (CMV) infections in mothers andyoung children were investigated by testing cord blood andserum samplesfrom infants and children up tofouryears oldfor thepresence ofCMVantibodies in 19 different regions of the world. The samples were tested by both the local virus laboratory and the reference laboratory, using the same batch ofcomplement-fixing CMVantigen and techniques which were validated by comparison of the titres recordedfor samples ofcoded sera sent to each laboratory. The incidence ofCMV antibodies variedfrom 44-100% in mothers andfrom 3-95% in young children. The number ofchildren with CMVantibodies increased with age in five areas; suggesting that there was some child to child transmission ofCMV infection between children in these regions. In the other regions, the absence ofany significant age- related increase indicated that the mainpathway ofCMVinfection in early life was by trans- missionfrom mothers to their infants. The significance of thesefindings is discussed. The high prevalence of cytomegalovirus (CMV) complement-fixing antibodies throughout the world indicates that CMV infections are very common, although the patterns of infection may vary in differ- ent regions. It has been difficult to compare the results of previous studies because of variations in the reagents and techniques employed by the different investigators. This study was designed to produce comparable data from different regions of the world on the incidence of CMV complement-fixing anti- bodies in mothers and their children in the first four years of life. The results were then analysed to compare the pattern of CMV infections in different regions and to identify the mode of transmission in infants and young children. MATERIALS AND METHODS Sera Blood samples were taken from umbilical cords (maternal) and newborn babies in the local maternity hospital, and from infants and young children admit- ted to the local hospital because of illnesses other than * This paper reports the results of a collaborative study organized by, and supported by a grant from, the World Health Organization. The work was coordinated by Professor Dr U. Krech. The labora- tories participating in the collaborative study (and the responsible officers) are listed in an annex on page 610. I Institute for Medical Microbiology, Frohbergstrasse 3, CH-9000 St Gallen, Switzerland. 2 Wilhelm Dunn School of Pathology, Oxford, England. those due to congenital abnormalities. The serum samples were normally divided into aliquots, one being evaluated by the local laboratory and the other by the reference laboratory at St Gallen, Switzerland. However, specimens from 5 participating centres were tested only by the reference laboratory, and the specimens from 2 other areas were tested only in the local laboratory, whose results had agreed well with those from St Gallen. Approximately 100 cord bloods and between 110 and 300 serum samples from children were tested in each location. Reagents Lyophilized complement-fixing CMV antigen was prepared by the reference laboratory and supplied to all collaborating investigators. In addition, each laboratory was provided with samples of three different, coded, lyophilized human sera and was asked to test them for CMV complement-fixing antibodies. Most laboratories used a locally adapted technique based on a method published by the American Public Health Association (1). RESULTS Comparative antibody titrations with coded sera The coded lyophilized sera were first titrated for antibody by complement fixation using the locally adapted techniques, as described in the preliminary 4094 -605- 606 U. KRECH & J. TOBIN Table 1. Cytomegalovirus complement-fixing antibodies in cord blood and children collected in different regions of the world Age of children (months) Region in which Cord Blood blood was 4-6 7-12 13-24 25-36 37-48 4-48 Total) collecteda No. posi- No. posi- No. posi- No. posi- No. posi- No. posi- No. posi- tive/No. % tive/No. % tive/No. % tive/No. % tive/No. % tive/No. % tive/No. % tested positive tested positive tested positive tested positive tested positive tested positive tested positive Oxford A 47/107 43.9 0/22 0.0 1/21 4.8 0/33 0.0 0/19 0.0 3/46 6.5 4/141 2.8 Albany A 53/120 44.2 7/50 14.0 8/50 16.0 8/77 10.4 missing 6/58 10.3 29/235 12.3 B 48/105 45.7 8/50 16.0 9/46 19.5 7/61 11.4 8/50 16.0 32/207 15.4 Manchester A 67/109 61.5 8/32 25.0 1/29 3.4 3/33 9.1 4/35 11.4 8/32 25.0 24/161 14.9 St Gallen A/B 58/100 58.0 10/60 16.6 10/64 15.6 5/51 9.8 16/51 31.3 12/48 25.0 53/274 19.3 Lyons A 51/100 51.0 7/56 12.5 9/58 15.5 13/60 21.6 16/81 19.7 21/79 26.5 66/334 19.7 B 60/ 98 61.2 8/55 14.5 8/56 14.2 15/59 25.4 20/80 25.0 22/72 30.5 73/322 22.6 Los Angeles A 71/ 99 71.7 19/99 19.2 12/99 12.2 12/98 12.2 13/81 16.0 12/96 12.5 68/473 14.3 B 74/ 97 76.2 22/86 25.5 14/86 16.2 14/80 17.5 16/81 19.7 11/91 12.1 77/424 18.2 Freiburg A 72/ 94 76.6 34/67 50.7 7/20 35.0 10/26 38.5 3/ 8 37.5 5/16 31.2 59/137 43.0 (in Breisgaul B 66/ 90 73.3 13/47 27.7 2/20 10.0 5/26 19.2 1/ 8 12.5 2/16 12.5 23/117 19.7 Anchorage B 103/125 82.4 3/21 14.3 5/20 25.0 9/34 26.5 27/43 62.8 25/40 62.5 69/158 43.7 Trinidad A 100/101 99.0 3/13 23.1 16/28 57.1 20/44 45.5 16/27 59.2 9/19 47.3 64/131 48.9 B 80/ 84 95.2 3/13 23.1 11/24 45.8 14/36 38.8 8/19 42.1 9/17 52.9 45/109 41.3 Buenos Aires A 104/105 99.0 11/32 34.4 23/46 50.0 18/39 46.1 12/23 52.5 2/ 3 66.6 06/143 46.2 B 102/107 95.3 10/31 32.2 18/46 39.1 17/37 45.9 11/22 50.0 2/ 3 66.6 58/137 42.3 Bratislava A 91/ 98 92.8 5/21 23.8 19/39 48.7 11/27 40.7 10/19 52.6 7/10 70.0 52/116 44.8 B 93/105 88.6 7/19 36.8 20/38 52.6 11/26 42.3 10/18 55.5 8/11 72.7 56/112 50.0 Rorne A 11/17 64.7 8/33 24.2 20/39 51.3 29/48 63.0 16/26 61.5 84/161 52.2 B 92/ 94 97.8 13/24 54.2 13/37 35.1 26/49 53.1 27/43 62.8 23/41 56.1 102/194 52.6 Hong Kong A 96/100 96.0 21/49 42.8 24/52 46.2 27/50 54.0 27/49 56.1 31/50 62.0 129/250 51.6 B 98/100 98.0 17/44 38.6 22/50 44.0 28/50 56.0 25/44 56.8 36/48 78.3 128/234 54.7 Mostar B 118/129 91.5 6/16 37.5 24/41 58.5 26/48 54.2 19/30 63.3 17/20 85.0 92/155 59.4 Sendai A 93/100 93.0 26/44 59.1 26/39 86.6 21/33 63.6 14/24 568.3 30/55 54.5 117/196 60.0 B 91/ 96 94.8 25/38 65.8 22/35 62.8 22/28 78.6 12/20 60.0 34/53 64.1 115/174 86.1 Fiji A 93/ 93 100.0 14/23 60.8 24/35 68.5 45/59 76.3 15/21 71.4 13/17 76.5 111/145 76.6 B 67/ 68 98.5 14/22 63.6 20/32 82.5 36/45 80.0 11/17 64.7 8/13 61.5 89/129 69.0 Vellore A 99/100 99.0 28/45 62.2 51/65 78.4 49/61 80.3 36/48 75.0 50/63 79.3 214/282 75.9 B 96/ 97 98.9 25/40 62.5 38/58 65.5 41/53 77.3 31/43 72.1 41/53 77.3 176/247 71.3 Ibodan B 111/115 96.5 70/75 93.3b 29/41 70.7 51/60 85.0 27/31 87.1 24/26 92.3 131/158 82.9 Entebbe B 121/121 100.0 24/26 92.3 28/28 100.0 24/27 88.8 52/53 98.1 128/134 95.5 a A- reaultB obtained locally: B - results obtained in St Galen. b Data given for age group 0-4 years; data for infants 4-6 months of age are not available. CMV IN MOTHERS AND YOUNG CHILDREN Serum A 512 - 256 - 128 - 641- 32 -- 16 -~ 8 - 4 - Serum B 0 @0 0 Serum C 0 0 0 0 0 0 0 0 0 0 0 0 0 0 < 4 -- O O 00 0 _o Fig. 1. Cytomegalovirus complement-fixing antibody titres in coded serum samples (A, B, and C). One result was not reported for serum A. study (2). The results are summarized in Fig. 1. All the laboratories found the negative serumA to be without antibodies, and the titre for serum B to be lower than that for serum C. Although the actual values varied, the ratio of the titres for sera B and C was consistent in all laboratories. The results demonstrate a high degree of uniformity; for example, the standard deviation of the titres recorded for serum B was so small that the threshold titre value of 4 was more than two standard deviations from the data range. Therefore, it was regarded as unlikely that any of these laboratories would record false negative results (if their methods remained constant). Comparison ofantibody titres obtained by the local and reference laboratories The data obtained by the local and reference labora- tories, on the presence ofCMV antibodies in cord and children's blood, are given in Table 1. For 5 regions (including that of the reference laboratory), data were available only from the reference laboratory. A good correlation was obtained between the results of the local laboratories and those of the reference labora- tory, with the exception of Freiburg. The results from this laboratory were all approximately 20% higher than the corresponding figures recorded in St Gallen. In the case of Albany, the local results were all lower than those from the reference laboratory, but they were so similar that the statistical probability that both laboratories would produce the same result was 94%. In the other regions the largest discrepancy, 49% as against 41% positive, was recorded in the data from Trinidad. Overall, there was good agreement between the results of the reference and local laboratories, the proportions found positive, of the total number of sera tested, being 40.2% and 40.7%, respectively. Changes in thefrequency of antibody-positive children with age In all areas the percentage of subjects positive for antibody was smaller in the first year of life than among the mothers (Fig. 2). The difference was significant in most cases. a I w. M a~ lo UoU M *W -1 Caz~ ~aA WOm U 4 a a Children Cord Blood Fig. 2. Cytomegalovirus complement-fixing antibody titres in serum from cord blood and from infants and children aged 4-48 months. In some regions, the percentage of antibody posi- tive children stayed fairly constant during their first four years of life, while in other areas the frequency increased during infancy and early childhood. Thus, the centres can be grouped into three cate- gories (Table 2): group A, those centres in which the prevalence of positivity was low in children aged 4-6 months and did not increase significantly in children up to 4 years of age; group B, those centres in which the level of positivity in children aged 4-6 months was high, but nevertheless did not increase significantly in Ia . 607 : U. KRECH & J. TOBIN Table 2. Prevalence of CMV complement-fixing antibody carriers in different areas, grouped according to the level of incidence at birth and its variation within the first four years Approximate percentage of Group Region antibody carriers in age groups 4-48 months A. Low antibody pre- Oxford 3 valence after birth Albany 15 with no significant Los Angeles 15 increase in the first Freiburg 20 four years of life (in Breisgau) B. High antibody pre- valence after birth Fiji 70 with no significant Sendai 70 change during the Vellore 70 first four years Ibadan 90 of life Entebbe 95 C. Low antibody pre- Manchester 5-25 valence after birth St Gallen" 10-30 with different degrees Lyons 10-30 of increase during the Anchorage 15-60 first four years of life TrinidadO 20-60 (approximate range Rome 25-65 given) Buenos Aires 30-65 Bratislava 30-70 Hong Kong 40-80 Mostar 40-85 Regression analysis indicated that all the areas in group C showed a statistically significant trend (P < 0.05) except for St Gallen (P = 0.2) and Trinidad (P = 0.5). b See Table 1 for details. children up to 4 years old; and group C, those areas where the proportion of antibody-positive children did increase during the first four years of life, thus implying exposure to CMV infections between 6 and 48 months of age. These increases in the antibody positive frequency in group C were statistically significant (P< 0.05) for seven of the regions (Table 2). Relation between CMVantibodies in mothers (cord blood) and samplesfrom children Results obtained from cord blood and children between 6 months and 4 years ofage are given in Table I and Fig. 2. The prevalence of antibody-positive mothers varied from 44/o in Oxford to 100/o in Entebbe and Fiji. However, the prevalence of anti- body-positive children between 6 months and 4 years of age varied from 3% in Oxford to 95% in Entebbe. A prevalence of positivity among mothers of approxi- mately 60% or less was associated with a low infection rate in their offspring, whereas a prevalence of over 85% was associated with a high prevalence of early childhood infection, ranging from 43% to 95% . How- ever, in two areas high frequencies of76-77% of anti- body-positive mothers were associated with a low prevalence of childhood infection, as judged by the percentage of antibody-positive children in these areas. DISCUSSION The data collected, on the incidence of CMV anti- bodies in mothers and in children from 6 months to 4 years of age, indicated that CMV infections occur frequently in all the areas tested. In some European and North American centres, approximately half to two-thirds of the mothers had been infected at some stage, but in other areas the frequency approached 100%. The development ofCMV antibodies in response to CMV infections after birth, showed different patterns in the regions investigated. The prevalence of infec- tion in infancy was low ((26%) in areas where the level of maternal infection was low or moderate ((60%). In places where the prevalence in adults approached 100%, that in infants varied from over 40% to 70%, except in the 2 African areas where it was between 87 and 95%. Two areas, Los Angeles and Freiburg, had a low level of infection in children although the frequency of maternal infection was over 75%. However, there was no region in which a high prevalence of infection was observed in young chil- dren whose mothers showed only a low or moderate rate of antibody-positivity. In the majority of areas (groups A and B) there was little evidence of transmission of infection from child to child during the first four years of life; this supports the hypothesis that most CMV infections in the first year of life are transmitted from mother to infant (3). However, in seven regions there was an increase in the frequency of infection of young children with age (Table 2) that was presumably due to child to child transmission. The different patterns may be explained by the differ- ences in child/mother contact, such as a longer period of breast feeding, that may increase the frequency of mother to child transmission in early infancy. Additional evidence indicating that mothers are the main source of infection in infancy, comes from observations in St Gallen (4), Manchester, and Singa- pore (5) that there is very little virus circulation after the breast-feeding period until the children reach puberty. In these areas, very few cases of CMV infec- tion were found in the age groups between 2 and 15 years. The incidence ofCMV infections in laboratory personnel handling, or exposed to, contaminated material does not differ significantly from that in other hospital personnel (6). This also supports the hypothesis that transmission ofCMV is more likely to occur by close personal contact than through contact 608 CMV IN MOTHERS AND YOUNG CHILDREN with contaminated specimens. Transmission among primary school children seems to be low (6). Numazaki et al. (3) related the high incidence of mother to child transmission in Japan to an increase in the cervical excretion of the virus during the latter stages of pregnancy. This conclusion is supported by the work of Alexander (7) in Taiwan. Stagno et al. (8),working in southern USA, considered the increase in cervical excretion in late pregnancy to be the result of a suppression of the normal level of excretion in early pregnancy. However, the occurrence of cervical excretion in the United Kingdom appears to be much less (12). It has been suggested that reactivation of viraemia might be related to infection at an early age and be less common after primary infection in adult life (9). This observation could explain the results from Los Angeles and Freiburg where a low level of infection in the first years of life was accompanied by a high preva- lence of antibody in the child-bearing population. Therefore, children in Western Europe and parts of the USA would be expected to have relatively little infection in the early years of life; in contrast, in Eastern Europe, Africa, Asia, and South America high rates of cervical excretion may cause early infection in succeeding generations. Intrauterine infection rates in developed countries have been estimated to be between 0.25 and 1.0%, but in areas where 90-100% of adult females have CMV antibodies these rates may exceed 2% (10). Secondly, in a recent study, which was part of Phase III of the WHO CMV project, the rate of intrauterine CMV- infection was found to be about 1.4% in populations in which the prevalence of CMV antibodies was almost 100%. These results were obtained in Abidjan and the Ivory Coast, and suggest that this increase in the prevalence of intrauterine infection in antibody carriers is due to reinfection or reactivation of a previous infection (11). ACKNOWLEDGEMENTS The statistical evaluation was performed by U. Gessner, in St Gallen, Switzerland, and by R. Mayon-White, in Oxford, England. R1iSUMtI ETUDE COLLECTIVE SUR LES ANTICORPS ANTI-CYTOMPGALOVIRUS CHEZ LES MARES ET LES JEUNES ENFANTS DANS 19 PAYS On a ttudit dans 19 regions du monde les variations regio- nales de l'incidence des infections A cytomegalovirus (CMV) chez des meres et de jeunes enfants en recherchant la pre- sence d'anticorps anti-CMV dans le sang du cordon ombi- lical et dans des tchantillons de serums preleves sur des nourrissons et de jeunes enfants Ages de moins de 4 ans. Les tchantillons ont ett soumis A des tests dans les laboratoires virologiques locaux et dans le laboratoire de rtftrence A St-Gall (Suisse), qui ont utilist le meme lot d'un antigtne CMV fixant le compltment et des techniques dont on a ttabli la valeur en comparant les titres relevts dans les tchantillons de strum humain codts qui avaient ett envoyts A chaque laboratoire. Les donntes recueillies indiquent que les infections A CMV se produisent frtquemment dans toutes les rtgions ttudites. Dans certains centres d'Europe et d'Amtrique du Nord, environ la moitit et parfois meme les deux tiers des meres ont ett infectees a un moment donne, mais ailleurs il arrive que la frtquence approche de 100%. La production d'anticorps anti-CMV aprts une infection aCMV postnatale a suivi difftrents schtmas dans les rtgions ttudites. La prtvalence de l'infection chez les nourrissons est faible (< 26%) dans les rtgions ou le niveau de l'infection maternelle est bas ou modtrt (< 60%). Dans les endroits oCu la prtvalence approche de 100% chez les adultes, chez les nourrissons elle varie de plus de 400/ A 70%, sauf dans les deux zones africaines o6 elle se situe entre 87% et 95%. Dans deux zones, Los Angeles et Fribourg, il existe un faible niveau d'infection chez les enfants, bien que la frtquence de l'infection maternelle dtpasse 75%. Cependant, dans aucune rtgion on n'a observt une forte prtvalence de l'infec- tion chez de jeunes enfants dont les mtres prtsentent une stropositivitt seulement faible ou modtrte. Dans la plupart des zones (groupes A et B) il n'est pas tvident qu'il y ait une transmission de l'infection d'enfant A enfant pendant les quatre premitres anntes de vie; cette observation semble confirmer l'hypothtse que la plupart des infections A CMV dans la premitre annte de vie sont transmises par la mtre au nourrisson. Toutefois, dans sept rtgions la frtquence de l'infection augmente avec l'Age chez les jeunes enfants, ce qui est vraisemblablement dO A une transmission d'enfant A enfant. Les differents schtmas peuvent s'expliquer par des contacts mtre/enfant difftrents, par exemple une longue ptriode d'allaitement maternel qui peut augmenter la frtquence de la transmission de mtre A enfant dans la premitre enfance. La signification de ces observations est examinee. 609 610 U. KRECH & J. TOBIN REFERENCES 1. Diagnostic proceduresfor viral and rickettsial diseases, 4th ed., New York, American Public Health Associ- ation, 1969, pp. 724-725. 2. KRECH, U. Bulletin of the World Health Organization, 49: 103 (1973). 3. NUMAZAKI, Y. ET AL. American journal of epidemio- logy, 91: 410 (1970). 4. KRECH, U. ET AL. Zeitschriftfalr Jmmunitatsforschung, 141: 411 (1971). 5. TOBIN, J. O'H. Annals of the Singapore Academy of Medicine, 8: 98-101 (1979). 6. HANSHAW, J. B. In: Remington, J. S. & Klein, J. O., ed., Infectious diseases ofthefetus and newborn infant, Philadelphia, W. B. Saunders, 1976, p. 118. 7. ALEXANDER, E. R. Pediatric research, 1: 210 (1967). 8. STAGNO, S. ET AL. Journal of infectious diseases, 131: 522 (1975). 9. COLLABORATIVE STUDY Archives of disease in child- hood, 45: 513 (1970). 10. STAGNO, S. ET AL. New England journal of medicine, 296: 1254 (1977). 11. SCHOPFER, K. ET AL. Archives of disease in childhood, 53: 536 (1978). 12. TOBIN, J. O'H. In: Elliot, K. & Knight, J., ed., Intra- uterine infections, Ciba Foundation Symposium, Amsterdam, Associated Scientific Publishers, 1973. (Ciba Foundation Symposium 10, new series). Annex I LIST OF PARTICIPATING LABORATORIES M. Aymard and Agn&s Calvet, Laboratoire de Viro- logie, Universite Claude Bernard, F-69373 Lyons, France Maria Isabel Berria, National Microbiology Institute, Buenos Aires, Argentina W. K. Chang, Virus Unit, Medical and Health Department, Queen Mary Hospital, Hong Kong Rudolf Deibel, Virus Laboratories, New York, Department of Health, Albany, New York 12201, USA J. H. S. Gear, The South African Institute for Medical Research, Johannesburg, South Africa Barbara Hull, Caribbean Epidemiology Centre, Port of Spain, Trinidad, West Indies Malati Jadhav, Enterovirus Laboratory, Christian Medical College, Vellore-4, India U. Krech, Institut fUr medizinische Mikrobiologie des Kantons St Gallen, CH-9000 St Gallen, Switzer- land Pauli Leinikki, Department of Virology, University of Helsinki, Helsinki 29, Finland J. A. R. Miles, University of Otago, Dunedin, New Zealand (results from Suva, Fiji) Michael R. Moser, Alaska Activities, Bureau of Epi- demiology, Anchorage, Alaska 99501, USA Yoshio Numazaki, Virus Laboratory, Sendai National Hospital, Sendai 983, Japan A. 0. Osoba, Department of Medical Microbiology, University College Hospital, Ibadan, Nigeria H. Schmitz, Hygiene-Institut der Universitat, D-78 Freiburg/Breisgau, Germany V. Schwanzer, Research Institute of Epidemiology and Microbiology, Bratislava, Czechoslovakia A. G. Senyakazana, East African Virus Research Institute, Entebbe, Uganda J. O'Hara Tobin, Public Health Laboratory, Rad- cliffe Infirmary, Oxford, England (results from Oxford and Manchester) Harry T. Wright, Jr., University of Southern Cali- fornia, School of Medicine, Los Angeles, Cali- fornia 90054, USA M. Zurovec, Regional Public Health Services, Mostar, Yugoslavia