Dec 1, 1999 - Seroprevalence to human cytomegalovirus (HCMV) varies in different populations and in different age groups.1â3 During pregnancy, rates of ...
© International Epidemiological Association 2000
International Journal of Epidemiology 2000;29:587–591
Printed in Great Britain
Human cytomegalovirus seroprevalence in three socioeconomically different urban areas during the first trimester: a population-based cohort study Pia Mustakangas,a Seppo Sarna,b Pirkko Ämmälä,c Marketta Muttilainen,d Pentti Koskelae and Marjaleena Koskiniemia
Background To re-evaluate the impact of socioeconomic status and human cytomegalovirus (HCMV) seroprevalence during pregnancy, we carried out a population-based cohort study. Methods
IgG and IgM antibodies to HCMV and IgG avidity were studied by enzyme-linked immunosorbent assay (ELISA) in three different socioeconomic areas (SEA) in the 9–12th week of pregnancy of 1088 consecutive mothers.
Results
The overall IgG seropositivity was 70.7%, ranging from 60.9 to 76.4% in ‘upper’ to ‘lower’ SEA (P = 0.0004). The HCMV IgM seropositivity was 4.0%, ranging from 3.8% in the ‘upper’ and ‘intermediate’ SEA to 4.6% in the ‘lower’ SEA. Serologically acute cases, defined by low avidity of IgG, represented 1.7% of the pregnancies in the ‘upper’ SEA compared with 1.0 and 1.1% in the other two areas. In the ‘lower’ SEA there were twice as many recurrent infections as in the others, 3.6 versus 1.7%. The low impact of age did not increase after elimination of the effects of SEA and parity. Miscarriages were associated neither with IgG nor with IgM positivity, although the percentage of ù2 miscarriages was 8.8% in seronegative women compared with 11.2% and 13.6% in IgG- and IgM-positive women.
Conclusions Social environment seems to be the most powerful factor, predicting both IgG seroprevalence and recurrences during pregnancy. Keywords
Human cytomegalovirus, cohort, pregnancy, seroprevalence, socioeconomic
Accepted
1 December 1999
Seroprevalence to human cytomegalovirus (HCMV) varies in different populations and in different age groups.1–3 During pregnancy, rates of 25–100% have been reported. The incidence of primary infections varies from 0 to over 10%.3,4 Primary infection is transmitted to the fetus in about 40% of the cases, but only rarely in chronic infection.2,5–8 Primary infections are rarer in low-income groups because of their high HCMV seropositivity. Yet the total number of congenital infections, and particularly of those that are symptomatic, is higher in the lowincome group9 and, moreover, immune women can give birth
a Department of Virology, Haartman Institute, b Public Health Sciences, c Obstetrics and Gynaecology, University of Helsinki, Helsinki, Finland. d Jorvi Hospital, Espoo, Finland. e National Public Health Institute, Oulu, Finland.
Reprint requests to: Pia Mustakangas, Haartman Institute, Department of Virology, PO Box 21, FIN-00014 University of Helsinki, Finland.
to babies with symptomatic infection.3,10 In spite of efforts for several decades, very little progress has been made regarding congenital HCMV infection.11,12 Vaccination programmes have been planned, but owing to inadequate basic data on transmission, not implemented. In order to re-evaluate the impact of socioeconomic factors, we carried out a population-based cohort study on HCMV seroprevalence during early pregnancy in three socioecomically differing areas and confirmed the decisive role of the social environment—there being a consistently lower and stable seroprevalence in the ‘higher’ socioeconomic area (SEA) and a concentration of recurrent infections in the ‘low’ SEA, but no correlation with age or parity.
Patients We studied 1088 consecutive pregnant women in the 9–12th week of pregnancy for HCMV antibodies. They were recruited from three different Maternity Health Care Centres in the
587
588
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Helsinki region, situated in three different socioeconomic areas, from March 1992 to June 1994. The participation rate was 98% of all pregnancies. The serum samples were stored at –20°C until assayed. Demographic data of the three population groups are presented in Table 1 (Finnish Official Statistics, 1993).
Methods Antibodies to HCMV were studied by an indirect enzyme-linked immunosorbent assay (ELISA) using commercial kits for specific IgG (Cytomegalovirus IgG EIA kit®, Labsystems, Helsinki, Finland) and IgM (CMV IgM EIA kit®, Labsystems) with alkaline phosphatase-conjugated mouse anti-human IgG and IgM as conjugates. The HCMV IgM-positive sera were retested for possible non-specific reactions after IgG inactivation (Gullsorp absorption). All HCMV IgM-positive samples were tested for IgG avidity,13,14 and the avidity index (AI) was calculated.15 For avidity, the standard HCMV IgG test was used, with serial dilutions of sera that were studied in parallel. The other series were rinsed with 8 M urea. For each serum, the AI was calculated by dividing the mean optic density (OD) of two sequential dilutions of the urea-washed sera by the mean OD of the reference wells not exposed to urea and multiplying by 100. An AI ,50% was regarded as low, suggesting an acute primary infection, 50–60% as borderline and .60% as high, indicating a chronic, and in association with IgM, a recurrent infection.16 A serological test for syphilis serology (the TPHA test, Fujirebio, Inc., Tokyo, Japan) was performed routinely at the Public Health Laboratory; all the woman were negative. Commercial kits were used for rubella virus (Rubaset® EIA-G and EIA-M, Orion Diagnostica, Helsinki, Finland) and Toxoplasma gondii (T. gondii IgG and IgM EIA kit; Labsystems).
Statistical methods Analysis of variance was used to compare the group means of continuous variables. The c2 test and log-linear models were used to test the associations for the categorical data. BMDP, dynamic 7.0 and StatXact 4.0 statistical software were used.17,18 All P-values given are two-sided.
Table 1 Demographic data of the three population groups, with a description of socioeconomic levels
Figure 1 Distribution of pregnancies (pregnancy rates in the respective age groups of the female population; distances between grouped values are smoothed) in three socioeconomically different areas (SEA). _____ ‘upper’ SEA, _ _ _ _ ‘intermediate’ SEA, -------- ‘lower’ SEA
Results Demographic data in Table 1 Academic degrees in education were obtained by 30% of the population in the ‘upper’ SEA compared to 2% in the ‘lower’ SEA (Table 1). Over 20% of individuals in the ‘upper’ SEA had a high income level (.36 000 $/year) compared with 6% and 1% in the ‘intermediate’ and ‘lower’ SEA. Sixty-three per cent of those in the upper SEA owned the flats they occupied compared with 32% in the lower SEA. More than 25% of the people in the lower SEA were unemployed, compared with 11% in the upper SEA (at that time, unemployment in Finland was about 20%). There were twice as many school-age and teenage children in the upper SEA as in the others and the numbers of family members were also higher. The pregnancy rate was lowest in the upper SEA and highest in the intermediate SEA, 5.7 and 15.3% (Figure 1) (P = 0.001). The difference was evident in all age groups. In upper SEA there was only one (0.1%) pregnancy in a woman ø19 years of age compared with 2.0% and 2.8% in the other SEA. More than 20% of pregnancies occurred in women ø24 years of age in the lower SEA compared with 8.9% and 11.7% in same other age groups in the upper and intermediate SEA. The overall difference in age between the pregnant women in the three SEA was highly significant (P = 0.001).
Human cytomegalovirus IgG seroprevalence
Intermediate %
Academic degree
30
14
2
Basic education only
22
43
62
Table 2 Prevalence of IgG antibodies to human cytomegalovirus in different age and socioeconomic groups Upper Intermediate Lower Mean Socioeconomic group n = 235 n = 573 n = 280 n = 1088
Socioeconomic group
Lower %
The overall HCMV IgG seroprevalence was 70.7%, ranging from 60.9% in the upper to 76.4% in the lower SEA (P = 0.0004) (Table 2). It was highest in the women ù35 years of age in the
Upper %
Education/profession
Income group ($/year) High ù36 000
21.9
5.9
1.4
Moderate 25 500–35 999
13.2
12.2
8.1
Medium 15 000–25 499
24.7
36.0
37.8
Low ,15 000
15.8
26.4
32.3
% of 15–24
61.9
68.2
75.9
70.3
No income tax
24.4
19.4
20.4
% of 25–34
60.8
71.5
76.3
70.3*
Self-employed
63
53
32
% of ù35
60.5
75.4
78.1
72.5
Unemployed
11
16
27
% of all
60.9
71.9
76.4
70.7**
No. of family members
2.7
2.1
1.9
Age group (years)
* P = 0.0042, ** P = 0.0004, c 2 test.
HUMAN CYTOMEGALOVIRUS SEROPREVALENCE IN THE FIRST TRIMESTER
lower SEA and lowest in the same age group in the upper SEA (78.1% versus 60.5%). In the women ø24 years of age, the seroprevalence was 70.3%, ranging from 61.9% in the upper to 75.9% in the lower SEA. In the intermediate SEA the seroprevalence tended to increase with age (NS).
Human cytomegalovirus IgM seropositivity and IgG avidity The overall HCMV IgM seropositivity was 4.0% (44 of 1088), ranging from 3.8% in the upper and intermediate SEA to 4.6% in the lower, suggesting primary or recurrent infections (Figure 2). In 41 cases IgM was associated with IgG seropositivity. Ten women showed a low avidity of IgG which with three IgM positive and IgG negative cases, suggested acute primary infections. The acute cases represented 1.7% of the pregnancies in the upper SEA compared with 1.0 and 1.1% in the others. The number of recurrent infections, defined as a presence of IgG, IgM and high or borderline avidity, was 3.6% in the lower SEA, more than twofold greater than in the other SEA (odds ratio [OR] = 2.08, 95% CI : 0.64–6.7, NS). Considering the low sensitivity of the IgM test in recurrent infections,6 the true number of recurrences may have been 36 and 17% (OR = 2.6, 95% CI : 1.7–4.0, P = 0.0001).
Determinants of seroprevalence The difference in seroprevalence between the upper and the lower SEA was highly significant (P = 0.0004) and was also apparent in the major age subgroup 25–34 years (P = 0.0042). With a log linear model we showed that seroprevalence was significantly associated with SEA even after eliminating the
589
influence of age. The impact of age did not increase on eliminating the impact of SEA and was suggestive only in the intermediate SEA. Neither parity nor number of gestations was associated with HCMV seropositivity. A three-point discrepancy between gestation and parity, indicative of ù2 miscarriages, appeared in 114 of the 1088 women (in 8.8% of the seronegative, 11.2% of the IgG-positive and 13.6% of IgM-positive women). In the upper SEA the number of ù2 miscarriages was 7.7%, and in the other SEA it was 11.2% and 11.4%. Thus, social environment seems to be the most powerful factor in predicting both IgG seroprevalence and recurrent HCMV infections, and with increasing age the seroprevalence remains at the same level, primary infections being few and distributed evenly among the different ages and SEA.
Discussion Our study indicates the basic role of the social environment as a predictor of both HCMV IgG seroprevalence and recurrent infections. Epidemiological studies made in different geographical and socioeconomic areas19–22 have shown great differences between different areas and emphasize the impact of race, place of birth, education, age and parity.23,24 In our population-based cohort study, the impact of the social environment was clearly shown by comparing three socioeconomically different areas, age and parity being without effect. In contrast with previous studies, the number of children in the family was not associated with seropositivity. Although the Nordic countries, including Finland, have a high socioeconomic standard,
Figure 2 The human cytomegalovirus IgM seropositivity and IgG avidity in the different socioeconomic areas (A) upper, (B) intermediate, (C) lower, (D) all. low avidity indicative of an acute primary infection, borderline, high indicative of recurrent infection.
590
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
people of different income levels concentrate in different areas, and thus we were able to select areas with unlike socioeconomic backgrounds. This, together with the nearly 100% participation, made it possible to separate the impacts of various factors. It must be realized, however, that even a lower socioeconomic level in Finland is high compared with the comparable level in poor and developing countries. Chronic maternal infection is defined by the presence of HCMV-specific IgG antibodies without HCMV-specific IgM antibodies during the first 12 weeks of gestation. In our series, chronic infection was significantly higher among women in the lower SEA than in the other two groups, and showed no correlation with age, gestation or parity. The HCMV-specific IgM was also most frequent, 4.6%, in the lower SEA. Whether an IgM case is associated with primary infection or reactivation can be determined with an IgG avidity test, although the correlation is far from complete.13–14 The presence of IgM-positive women with high or borderline avidity indicated that, compared with the other two groups, more than twice as many women in the lower SEA had apparent re-activations. Remarkably, the most frequent cause of reactivation of infection and intrauterine transmission of HCMV in immune women seems to be endogenous infection.25 In an earlier report, avidity ,50% reflected an acute primary infection within 3 months, avidity .65% was always associated with an older infection,15,16 and only 20% of all IgM were associated with acute infections. This resembles our percentage of primary infections among IgM positive cases. Primary infections were distributed in every age group and in all the SEA. Clearly, the sensitivity of IgM in recurrences is low in HCMV infections, perhaps only 10%.6 Thus 90% of recurrences may remain unidentified. A high proportion of seronegativity, and thus a high socioeconomic level, is regarded as a risk factor for severe congenital disease. However, families with a congenitally HCMV infected child have not been characterized. According to Stagno, if HCMV IgG seropositivity is 55%, congenital infections are transmitted half from primary and half from recurrent infections.26 In the low income group, seropositivity is 82% and congenital infections are transmitted from primary infections in 25% and from recurrent infections in 75% of cases. The children who are symptomatic at birth will be damaged. However, asymptomatic children may develop sequellae insidiously over time.27 Young, low socioeconomic group women excrete HCMV in 11–28% of cases and in even greater numbers in late pregnancy. Excretion is rarer in early pregnancy but increases towards the end, 25%, or even .50%.28 Maternal antibodies afford substantial protection to the fetus but the protection is imperfect. According to a British study,10 most congenital symptomatic infections arise from recurrent maternal infections. Although this may be disputed and it contrasts with previous reports26 it, however, emphasizes the possible role of chronic infection. The conclusion is often drawn that congenital HCMV would be more frequent and more severe in the upper socioeconomic groups, but this may not be true.10 Several reports indicate the importance of age and parity,24 which we were unable to confirm. Decisions about vaccination are not straightforward decisions.29 Pregnant women may not be protected from recurrences by vaccination and the seropositivity status is clearly associated with the socioeconomic status. The recurrence rate is
much higher in the lower SEA with inherently high seropositivity. The same or some other CMV strain may be responsible for repeated infections.8 Notably, CMV seroprevalence in general seems to have been decreasing since 1980.30 Simple hygienic measures such as hand-washing may be the reason.8,31 This is consistent with the improvement in socioeconomic factors in accordance with our study.
Acknowledgements This study was in part supported by the Helsinki University Central Hospital Research Funds. The valuable time and assistance of the three Maternity Health Care Centres and of the Departments of Obstetrics and Gynaecology, Helsinki University Central Hospital and Jorvi Hospital is appreciated.
References 1 Griffiths PD, Stagno S, Pass RF et al. Infection with cytomegalovirus
during pregnancy: Specific IgM antibodies as a marker of recent primary infection. J Infect Dis 1982;145:647–53. 2 Griffiths P, Baboonian C, Ashby D. The demographic characteristics of
pregnant women infected with cytomegalovirus. Int J Epidemiol 1985; 14:447–52. 3 Stagno S, Pass RF, Cloud G et al. Primary cytomegalovirus infection in
pregnancy. JAMA 1986;256:1904–08. 4 Stern H, Tucker SM. Prospective study of cytomegalovirus infection
in pregnancy. Br Med J 1973;2:268–70. 5 Kumar ML, Gold E, Jacobs IB et al. Primary cytomegalovirus infection
in adolescent pregnancy. Pediatrics 1984;74:493–500. 6 Stagno S, Tinker MK, Elrod C et al. Immunoglobulin M antibodies
detected by enzyme-linked immunosorbent assay and radioimmunoassay in the diagnosis of cytomegalovirus infections in pregnant women and newborn infants. J Clin Microbiol 1985;21:930–35. 7 Fowler KB, Stagno S, Pass RF et al. The outcome of congenital cyto-
megalovirus infection in relation to maternal antibody status. N Engl J Med 1992;326:663–67. 8 Nelson CT, Demmler GJ. Cytomegalovirus infection in the pregnant
mother, fetus, and newborn infant. Clin Perinatol 1997;24:151–60. 9 Fowler KB, Stagno S, Pass RF. Maternal age and congenital cyto-
megalovirus infection: Screening of two diverse newborn populations, 1980–1990. J Infect Dis 1993;168:552–56. 10 Griffiths PD, Baboonian C, Rutter D et al. Congenital and maternal
cytomegalovirus infections in a London population. Br J Obstet Gynecol 1991;98:135–40. 11 Yow MD. Congenital cytomegalovirus disease: A NOW problem. J Infect
Dis 1989;159:163–67. 12 Hagay ZJ, Biran G, Ornoy A et al. Congenital cytomegalovirus infec-
tion: A long-standing problem still seeking a solution. Am J Obstet Gynecol 1996;174:241–45. 13 Ruellan-Eugene G, Barjot P et al. Evaluation of virological procedures
to detect fetal human cytomegalovirus infection: Avidity of IgG antibodies, virus detection in amniotic fluid and maternal serum. J Med Virol 1996;50:9–15. 14 Hedman K, Rousseau SA. Measurement of avidity of specific IgG for
verification of recent rubella. J Med Virol 1989;27:288–91. 15 Grangeot-Keros L, Mayaux MJ, Lebon P et al. Value of cytomegalo-
virus (CMV) IgG avidity index for the diagnosis of primary CMV infection in pregnant women. J Infect Dis 1997;175:944–46. 16 Bodéus M, Feyder S, Goubau P. Avidity of IgG antibodies dis-
tinguishes primary from non-primary cytomegalovirus infection in pregnant women. Clin Diagn Virol 1998;9:9–16.
HUMAN CYTOMEGALOVIRUS SEROPREVALENCE IN THE FIRST TRIMESTER
17 BMDF Dynamic 7.0, BMDP Statistical Software Inc., Cork Ireland, 1993. 18 StatXact 4.0, Cytel Software Corporation, MA, USA, 1998. 19 Gambarotto K, Ranger-Rogez S, Aubard Y et al. Primary cytomegalo-
virus infection and pregnant women: epidemiological study on 1100 women at Limoges. Path Biol 1997;45:453–61. 20 Luchsinger V, Suarez M, Schultz R et al. Incidence of cytomegalovirus
congenital infections in newborns of different socioeconomic strata. Revista Med Chile 1996;124:403–08. 21 Natali A, Valcavi P, Medici MC et al. Cytomegalovirus infection in an
Italian population: antibody prevalence, virus excretion and maternal transmission. New Microbiol 1997; 20:123–33. 22 Stein O, Sheinberg B, Schiff E et al. Prevalence of antibodies to cyto-
591
25 Huang E-S, Alford CA, Reynolds DW et al. Molecular epidemiology of
cytomegalovirus infection in women and their infants. N Engl J Med 1980;303:958–62. 26 Stagno S, Pass RF, Cloud G et al. Congenital cytomegalovirus infec-
tion. The relative importance of primary and recurrent maternal infection. N Engl J Med 1982;306:945–49. 27 Reynolds DW, Stagno S, Stubbs G et al. Inapparent congenital cyto-
megalovirus infection with elevated cord IgM levels. N Engl J Med 1974;290:291–96. 28 Montgomery R, Youngblood L, Medearis DN Jr. Recovery of cyto-
megalovirus from the cervix in pregnancy. Pediatrics 1972;49: 524–31.
megalovirus in a parturient population in Israel. Isr J Med Sc 1997; 33:53–58.
29 Pass RF. Immunization strategy for prevention of congenital cyto-
23 Walmus BF, Yow MD, Lester JW et al. Factors predictive of cyto-
30 Hoshiba T, Asamoto A, Yabuki Y. Decreasing seropositivity of cyto-
megalovirus immune status in pregnant women. J Infect Dis 1988; 157:172–77.
megalovirus of pregnant women in Japan. Jpn J Clin Med 1988;56: 193–96.
24 Gratacap-Cavallier B, Bosson JL, Morand P et al. Cytomegalovirus
31 Murph JR, Baron JC, Brown CK et al. Occupational risk of cyto-
seroprevalence in French pregnant women: parity and place of birth as major predictive factors. Eur J Epidemiol 1998;14:147–52.
megalovirus infection among day-care providers. JAMA 1991;265: 603–05.
megalovirus infection. Infect Agents Dis 1996;5:240–44.
