Seroprevalence, incidence of prenatal infections and reliability of maternal history of varicella zoster virus, cytomegalovirus, herpes simplex virus and parvovirus B19 infection in South-Western Finland
Dr A. Alanen, Department of Obstetrics and Gynaecology, University of Turku, Kiinamyllynkatu 4–8, FIN-20520 Turku, Finland.
Objective To study seroprevalence and incidence and fetal transmission of varicella zoster virus (VZV), cytomegalovirus (CMV), herpes simplex virus (HSV) types 1 and 2 and parvovirus B19 infections during pregnancy and to evaluate the reliability of maternal past history of VZV, HSV and parvovirus infections.
Design Prospective study of parturient women.
Setting South-Western Finland.
Participants Five hundred and fifty-eight parturient women.
Methods IgG and IgM antibodies against VZV, CMV, HSV-1 and -2, and parvovirus B19 were measured from maternal serum in the first trimester and at delivery and from cord serum, mother's own information of her past infections was compared with her serological status.
Main outcome measures Seroprevalence, seroconversions and fetal transmission of VZV, CMV, HSV and parvovirus B19, reliability of maternal history of VZV, HSV and parvovirus B19.
Results Seroprevalences were 96.2% for VZV, 56.3% for CMV, 54.3% for HSV, 46.8% for HSV-1, 9.3% for HSV-2 and 58.6% for parvovirus B19. Parity was associated with CMV seropositivity, maternal age differed only between HSV-2 seropositive and seronegative women, while area of residence (urban or rural) had no effect. Six seroconversions were observed: two VZV, one CMV and three parvovirus infections. No cases of primary HSV infections occurred. Fetal transmission was observed in two cases of parvovirus infection. No infants with anti-CMV IgM antibodies were born to CMV immunised women. False positive history of chickenpox was given only by 1.5% of the women, history of herpes infections was less reliable, and history of parvovirus infection was unreliable.
Conclusions Seroprevalence and the risk of viral infections during pregnancy cannot be extrapolated from one pregnant population to another.
Pregnancy induces a transient immunosuppression,1 which is thought to increase the vulnerability of pregnant women to viral infections. Since the abrogation of congenital rubella infections by vaccination, cytomegalovirus (CMV), varicella zoster virus (VZV) and parvovirus B19 are the most important viruses to cause clinically significant intrauterine fetal infections. Viruses of the herpesvirus group, CMV and VZV are potentially teratogenic,2 whereas parvovirus can cause fatal anaemia to the fetus.3 In case of VZV and parvovirus, the risk of a congenital infection is connected only with a maternal primary infection. Herpes simplex virus (HSV), although extremely rarely the cause of a congenital infection, can infect the newborn during vaginal delivery with poor outcome.4 Herpesviruses have the ability to reactivate from time to time. Recurrent or secondary infections during pregnancy have clinical relevance in case of HSV and CMV. Reactivation of HSV during pregnancy probably never leads to an intrauterine infection, but a recurrence of genital herpes infection creates a slight risk to the newborn if it occurs at the time of delivery.5 CMV reactivation, on the other hand, produces viraemia and can thus result in a congenital infection of the fetus. Recent evidence has shown that a part, maybe even a majority of these, are new CMV infections with a different strain.6 Thus, maternal seropositivity does not protect the fetus against intrauterine CMV infection as it does in case of VZV, parvovirus and HSV.
Congenital viral infections are not among the most usual complications affecting the outcome of pregnancy, but are among those that can be influenced by, for example, screening and vaccination programs. Prior to the evaluation of the effectiveness and cost of these interventions, the incidence of these infections as well as the proportion of pregnant women susceptible to them should be regionally known. In evaluating the risk of these potentially harmful viral infections during pregnancy, serological analysis determines the seroprevalence in the pregnant population revealing the amount of seronegative women susceptible to a primary infection. Several factors are known to influence seroprevalence, among them geography, climate, maternal socio-economic status, occupation, race, age and parity.7–9 VZV seroprevalence is mainly influenced by climate; in temperate climates the seroprevalence of VZV is uniformly high among native adult population irrespective of other factors. CMV and HSV seroprevalences, on the other hand, are significantly influenced by socio-economic factors and are thus highly dependent on which kind of subgroup has been studied. The seroprevalence of parvovirus seems to be fairly constant worldwide, although women working among children are more prone to be seropositive.10
Most studies on viral infections during pregnancy are conducted in big cities with a highly heterogenous population. This study was performed in a middle size Finnish city with surrounding rural areas. The study population is typical of Scandinavian or Northern European people living outside big metropolitan areas. We studied the seroprevalence and the incidence of seroconversion during pregnancy with VZV, CMV, parvovirus B19 and HSV types 1 and 2 (HSV-1 and HSV-2). Mother's own information of her earlier chickenpox, herpes simplex and parvovirus infection was compared with her serological status to evaluate the reliability of her past history.
This study was conducted with informed consent and with the approval of the Ethical Committee of the Hospital District of South-Western Finland.
Turku University Hospital serves as the only hospital for parturient women in an area with a population of 260,000 inhabitants including both urban and rural areas. The population is very homogenous concerning race and standard of living.
This study was undertaken between 15th June and 15th August 2000. Blood samples were taken from 580 women, who came to deliver at Turku University Hospital during this time. The women also filled a questionnaire, in which they were asked, whether they knew of having previously had chickenpox, orofacial or genital herpes or parvovirus infection. After the delivery, a cord blood sample was taken also from the newborn. Serum was separated from the blood samples and stored at −20°C. Serum samples taken during the first trimester of the same pregnancy for hepatitis B and HIV screening were obtained from National Health Laboratories in Oulu where they are stored.
Five hundred and fifty-eight complete sets of samples were obtained. These included a serum sample from the mother taken in the first trimester and at the time of delivery, serum sample from the cord blood of the newborn and questionnaire filled by the mother. Seven hundred and twenty-four women gave birth in the hospital during the study period, so the study population covered 77% of all parturient women. The main reason for some mothers not to be included in the study was the randomly occurring inability of the delivery room staff to take the samples (120 women), or a non-Finnish speaking mother unable to fill the questionnaire (38 women). All women in the study were either native Finnish or although born elsewhere, had lived in Finland for several years. Only eight parturients, who were asked to participate, refused. We are thus convinced that there was minimal selection bias.
The median age of the women was 30 years (range 16–45 years) and the median number of previous children was 1 (range 0–12). Two hundred and forty-seven women (44%) were primiparous.
IgG and IgM antibodies against VZV, CMV, herpes simplex virus types 1 (HSV-1) and 2 (HSV-2) and parvovirus B19 were measured by enzyme immunoassay from the maternal samples. The paired samples were always measured simultaneously in the same assay. Sample pairs with significant difference in the specific IgG antibody amount (seroconversion or at least a twofold increase) were measured in a repeated assay together with the cord serum sample from the infant for specific IgM analysis. In case of CMV, 300 cord serum samples from the 314 infants born to CMV seropositive mothers were studied for CMV-specific IgM.
Antibodies against VZV, CMV and HSV were determined by indirect enzyme immunoassay as described by Ziegler et al.11 Briefly, HSV F-strain-infected Vero cells, VZV OKA-strain- or CMV AD169 strain-infected human diploid fibroblast cells were used for preparation of antigens. The virus strains used were received from the American Type Culture Collection. Polystyrene microtitre strips were coated with 0.5 μg of antigen. Serum samples, diluted 1:100 in phosphate buffered saline, were tested as duplicates and incubated for 2 hours at 37°C, after which horseradish peroxidase (HRP)-conjugated anti-human-IgG or IgM (at dilutions of 1:12,000 and 1:3000, DAKO, Germany) were added. Orthophenylenediamine (OPD) was used as a colour substrate and the optical density was measured at 492 nm by a Multiscan Analyzer (Eflab OY, Helsinki, Finland). Results were expressed as ratios obtained by dividing the OD value of a serum sample by the mean OD value of the positive control and multiplied by 100.
Serum samples containing HSV IgG antibodies were retested with HSV-1 and HSV-2 type-specific commercial kits (MRL Diagnostics, Cypress, California, USA).
For IgG and IgM class antibodies against parvovirus B19, a commercial enzyme immunoassay kit (Biotrin, Sinsheim-Reihen, Germany) was used.
The differences in mean ages between virus-positive and virus-negative women were tested with two-sample t test. χ2 test was applied to investigate the difference between orofacial and genital herpes on seroprevalence of HSV viruses. The effects of parity and area of residence on seroprevalence of viruses were analysed using logistic regression analysis. Parity was divided into three groups (those with 0, 1 and 2 or more previous children) and area of residence into two groups (urban and rural area) for the logistic models. Results were quantified by calculating the odds ratios (OR) with 95% confidence intervals (95% CI). Seroconversions were excluded from statistical analysis. P values less than 0.05 were considered as statistically significant. Statistical computations were performed with SAS System for Windows, release 8.02.
The seroprevalence of VZV was 96.2%(Table 1), which was of expected range. Maternal age was not different between VZV seropositive (mean age [SD]: 29.6 [5.3] years) and seronegative (28.4 [6.6] years) (P= 0.32). There was a clear relationship of the rate of seropositivity with the increasing number of previous children (Table 2). Because of the very low number of the seronegative women compared with the seropositive ones, it did not reach statistical significance (women with one previous child vs primiparous: OR 1.12, 95% CI 0.42–3.01; women with two or more previous children vs primiparous: OR 2.44, 95% CI 0.53–11.30).
Table 1. Seroprevalence and rate of seroconversions and fetal transmission of VZV, CMV, HSV and parvovirus B19.
|Parvovirus B 19||327||58.6||231||41.4||3||0.54||1.3||2/3|
|HSV-total||303||54.3||255||45.7||0||0|| || |
|HSV-1||261||46.8||297||53.2||0||0|| || |
|HSV-2||52||9.3||506||90.7||0||0|| || |
|Double positive||10||1.8|| || || || || || |
Table 2. Effect of parity on maternal seroprevalence. Values are presented as n (%) or mean [range].
|0||247||44.3||27.5 [16 to 45]||236 (95.5)||149 (60.3)||123 (49.8)||126 (51)|
|1||194||34.7||30.2 [20 to 41]||186 (95.9)||111 (57.2)||117 (60.3)||103 (53.1)|
|2||88||15.8||32.1 [21 to 44]||86 (97.7)||48 (54.5)||54 (61.4)||56 (63.6)|
|>2||29||5.2||35.2 [25 to −44]||29 (100)||19 (65.5)||20 (70)||18 (62.1)|
The seroprevalence of CMV was 56.3%. Although there was no difference in maternal age between seropositive (29.7 [5.4] years) and seronegative mothers (29.5 [5.3] years) (P= 0.62), parity did significantly influence the seroprevalence (P= 0.02, women with one previous child vs primiparous: OR = 1.52, 95% CI 1.04–2.22; women with two or more previous children vs primiparous: OR = 1.72, 95% CI 1.10–2.70) (Table 2).
Seropositivity for HSV was 54.3%. Most were HSV-1 seropositive (86%) and 14% were seropositive for HSV-2. Ten women (1.8%) were positive for both HSV-1 and HSV-2 antibodies. Maternal age did significantly correlate with HSV-2 but not with HSV-1 seropositivity. The average age was 31.6 [5.2] years for HSV-2 positive and 29.4 [5.3] years for HSV-2 negative mothers (P= 0.005). There was a tendency for HSV seropositivity to increase with parity (Table 2) (P= 0.086). The difference was significant between women with two or more children and primiparous ones (OR = 1.65, 95% CI 1.05–2.60), although not between women with one previous child compared with primiparous (OR = 1.09, 95% CI 0.75–1.58).
Fifty-eight percent of the pregnant women were seropositive for parvovirus B19. Neither maternal age nor parity had an effect on the rate of seropositivity. Mean age was 29.7 [5.4] years for seropositive women and 29.3 [5.3] years for seronegative women (P= 0.35) and the number of previous children was not associated with maternal seropositivity (P= 0.78).
The effect of area of residence, urban or rural, on seroprevalence was also studied (Table 3). Although there was significantly more primiparous women living in urban versus rural area (P= 0.004), the area of residence had no significant effect on seroprevalence in case of any of the viruses studied.
Table 3. Effect of living area on seroprevalence.
|OR|| || ||1.05||1.36||1.30||0.83|
|95% CI|| || ||0.37–2.95||0.92–2.0||0.88–1.89||0.57–1.21|
Six seroconversions were detected in this study: two cases of VZV, one CMV and three parvovirus infections (Table 1).
Twenty-one women were negative for VZV antibodies at the beginning of the pregnancy and two of them seroconverted during pregnancy. One had chickenpox in the 19th week of gestation and was treated with acyclovir. Amniocentesis was performed two weeks after the disease with negative results on VZV isolation and VZV-specific PCR. The other woman with VZV seroconversion was not aware of the disease during pregnancy. According to the questionnaire and to a later interview, she thought that she had the disease during childhood. Of the 21 non-immune women for VZV infection, there was a 9.5% risk of contracting this infection during pregnancy. Although exposure to VZV during pregnancy is fairly frequent, this study is too small to give an accurate estimate of this risk. The incidence of congenital VZV is very low 1–2% in the presence of maternal infection during pregnancy. Both the infants of the seroconverted mothers were healthy and IgM negative.
One seroconversion of CMV during pregnancy occurred in this study. Her newborn was healthy and IgM negative. The incidence of primary CMV infection during pregnancy was 0.4% for seronegative women.
Parvovirus infection was the most common infection in this study. Three mothers seroconverted during pregnancy. The incidence of parvovirus infection was 1.3% for seronegative women during pregnancy. Two of the affected mothers had worked in a daycare centre during pregnancy. One of them had fever, but no rash or arthralgia, and in none of the three cases was parvovirus infection suspected. Two infants were positive for parvovirus-specific IgM antibody. One of these two fetuses had several antenatal ultrasound scans because of intrauterine growth restriction. No hydrops was observed and parvovirus infection was not suspected. An increased risk of newborns small for gestational age has been reported in cases of maternal parvovirus infection during pregnancy.12
There were no cases of primary HSV infection. None of the 10 mothers who were positive to both HSV-1 and HSV-2 at delivery seroconverted to either serotype during pregnancy, indicating that there were no first episode non-primary HSV infections either (first HSV-1 infection in a HSV-2 positive person or vice versa).
Gestational age did not affect specific antibody levels. The median ratios between the simultaneously measured samples (late pregnancy/early pregnancy) were 0.83 for VZV (range 0.19–5.13), 0.93 for HSV (range 0.11–3.85) and 1.0 for CMV (range 0.08–3.67). Antibody ratios of more than 4 indicating a reactivation were excluded in case of HSV and CMV infections.
Reactivation of CMV infection causes an increase in the specific IgG levels. Even a twofold increase is considered as a probable recurrence. On the other hand, estimation of the rate of recurrence based on serology alone without evidence of viral excretion is not accurate. A fourfold rise in CMV IgG levels without an increase in the levels of other viral antibodies was observed in seven cases (2.2% of seropositive mothers), a threefold increase in 18 cases (5.7%) and a twofold increase in 42 cases (13%). The rate of reactivations is, however, less important than the amount of infected children. The cord blood of 300 newborns of 314 CMV seropositive mothers was studied including all those samples with any increase in maternal IgG levels. All cord blood samples were negative for specific anti-CMV IgM antibodies, indicating that in spite of one seroconversion and several probable recurrences, there was no congenital CMV infections based on cord blood CMV-specific IgM antibodies.
The parturient mothers were asked whether they knew of previously having had chickenpox, parvovirus infection, orofacial or genital herpes infection.
With VZV, 467 out of 558 (84%) women gave a positive history of previous infection (Table 4). The majority (98%) was seropositive with less than 2% being seronegative. Of those who thought that they had never had chicken pox infection, 87% were seropositive.
Table 4. Reliability of maternal history of VZV and parvovirus infection. Values are presented as n (%).
|Positive history||467||460 (98.5)||7 (1.5)|
|Negative history||91||79 (86.8)||12 (13.2)|
|Positive history||9||7 (77.8)||2 (22.2)|
|Negative history||549||323 (58.8)||226 (41.2)|
Parvovirus B19 infection was included in the questionnaire but was presumed to be relatively unfamiliar among the general population. Human parvovirus infection was only discovered in 1983,13 and this cohort of pregnant women cannot be expected to have knowledge of previous childhood parvovirus infection. Only nine women thought they had a previous parvovirus infection and two of them were seronegative.
The reliability of the information concerning herpes infection is difficult to evaluate because both serotypes, HSV-1 and HSV-2, can cause either orofacial or genital herpes. The results are presented in Table 5. One-third of the women (190/558) gave a positive history of either HSV-1 or HSV-2 infection. One hundred and seventy-five (31%) women gave a previous history of orofacial HSV and 4.1% thought they previously had genital herpes. Eight women (14%) had previous orofacial and genital herpes. Information on genital herpes seemed to be less reliable than that of orofacial herpes. False positive information, that is, positive history given by a HSV seronegative woman, was greater for genital compared with orofacial herpes (13%vs 6%, respectively). The rate of false negative cannot be evaluated based on serology alone but there was significantly less HSV seropositive women with negative history of orofacial herpes (36%) compared with that of genital herpes (53%). One hundred and sixty-six out of 261 HSV-1 positive women (64%) gave a history of either herpes infection, while only 24 out of 52 HSV-2 positive women did so (46%).
Table 5. Maternal history of orofacial and genital herpes simplex infection and HSV seropositivity.
|Positive history||190||176 (92.6)||159 (83.7)||22 (11.6)|
|Negative history||368||127 (34.5)||102 (27.7)||30 (8.2)|
|Positive history||175||164 (93.7)||155 (88.6)||14 (8.0)|
|Negative history||383||139 (36.3)||106 (27.7)||38 (9.9)|
|Positive history||23||20 (87.0)||11 (47.8)||10 (43.5)|
|Negative history||535||283 (5.9)||250 (46.7)||42 (7.9)|
The seroprevalence of VZV was high (96.2%), as was expected. In temperate climates where VZV epidemics among children are common, most adults are immune. There has been some reports from Britain and USA suggesting that the amount of susceptible adults could be increasing.14,15 The low number of seronegative pregnant Finnish women (3.8%) in this study indicates that such an increase is not occurring in Finland, at least not among the native citizens. The proportion of individuals from tropical countries is known to affect the seroprevalence of VZV. The seronegativity rate in a high risk obstetric population in the USA with a high degree of immigrants has been reported to be as high as 11.6%.16 Seroprevalence of CMV was 56.3%, which is within the range usually reported for Europe (50–80%).17,18 In a previous study of pregnant women conducted in Helsinki, Finland, the seroprevalence was found to be 70.7% (range 60.9–76.4%).19 The difference between these seroprevalence rates for CMV demonstrates that marked variations can exist between quite close regions within the same country. The seroprevalence of parvovirus B19 varies much less compared with that of VZV and CMV. Except for the Far East where the reported seroprevalence rate is lower (20–30%), 50–70% of adult population elsewhere are usually seropositive.10,20 The rate of immunity to parvovirus B19 in this study (57%) is therefore within this expected range.
The number of previous children was related to the rate of VZV seroprevalence. Half of seronegative women were primiparous and all women with more than two children were immune. In previous studies, the number of siblings has been shown to correlate with VZV immunity among adolescents and young adults, emphasising the significance of household exposure to VZV.21,22 Likewise, the seropositivity to CMV correlated with the number of children. Breastfeeding is one important route of CMV transmission. Stagno and Cloud23 have shown that most seropositive women occasionally secrete CMV into breastmilk. Infected symptomless infants excrete the virus for prolonged periods and transmit CMV to other children they come in contact with, for instance, in daycare centres. The transmission from children to their family members may explain the relationship between VZV and CMV seropositivity and parity. Interestingly, the number of previous children did not affect the immunity to parvovirus infection. Similar results have been reported previously24; a large Danish study reported the presence of such a relationship.10 Parvovirus is, however, much less contagious than VZV.
Positive maternal history of past VZV is usually considered quite reliable22,25 and the rate of false positive history in this study was low (1.5%). Exposure to VZV is not rare during pregnancy. In such cases, a reliable positive maternal history may affect the need to test the mother for susceptibility as prophylactic post-exposure treatment with acyclovir is likely to become a recommended option to prevent clinical chickenpox among pregnant women.
In this study, the most frequent infection detected by seroconversion during pregnancy was parvovirus B19 infection. Parvovirus infections have, however, a seasonal variation peaking in spring and early summer. As the parturient women participating in this study delivered in the summer, the interval between the two samplings included the peak seasons. Parvovirus epidemics occur every three to five years with a considerably increased rate of seroconversion also during pregnancy.10 There were no epidemics of parvovirus infection during the study period. The rate of seroconversion among susceptible women was 1.3%, which is in accordance with those of previous reports. Parvovirus was transmitted into the fetus in two out of three cases.
The rate of primary CMV infection was surprisingly low. Only one woman out of 244 originally seronegative women seroconverted (0.4%). The frequently cited average risk of primary CMV infection is in the range of 2.0–2.5%,8,26 although incidence of less than 1% of seronegative women have been reported from Europe.27,28 CMV is usually acquired by sexual transmission or from small children at home. Two hundred and forty-seven (44%) women in this study were primiparous. In addition, there were no primary or first episode non-primary (e.g. first HSV-2 infection in a HSV-1 positive person) HSV infections among the women, suggesting that the acquisition of a sexually transmitted infection was not common, hence, the probable risk of CMV exposure was low.
CMV infection is considered to be the most frequent congenital infection of the fetus with reported incidence varying between 0.2% and 2% of live births in Europe and in the USA.18,29,30 The lower the rate of pre-immunised women, the higher is the proportion of infants infected due to a primary versus secondary CMV infection. Stagno et al.31 have estimated that in a population with 55% seropositivity, corresponding to ours, congenital infections are transmitted in equal numbers from primary and secondary infections. In the only case of a primary CMV infection, the newborn was healthy and IgM negative. Likewise, no IgM positive infants were born to the pre-immunised women. As only about 70% of congenitally infected children have IgM antibodies in the cord blood after birth,30 we cannot rule out the possibility of a congenital CMV infection.
HSV seropositivity in this study was 54.3%, which is low compared with reports from Europe and the USA.9 It is also lower than in the Helsinki city area where 70% of pregnant women were HSV-1 seropositive and 16% were HSV-2 seropositive.32 The respective rates of seropositivity in our study were 46.8% for HSV-1 and 9.3% for HSV-2. Although HSV-2 seropositivity has been connected with genital herpes infection and correlates with sexual behaviour,33 recent reports indicate that the incidence of genital infection due to HSV-1 is increasing especially among young people.34 In the clinic of sexually transmitted diseases in Turku, HSV-1 accounts for 40.9% of the HSV isolates from the genital area. Genital herpes is very often symptomless. Even primary infections give recognisable symptoms leading to diagnosis in only one-third of the cases during pregnancy.33,35 Accordingly, mother's own information proved not to be very reliable as has been shown with patients attending clinics for sexually transmitted diseases.36 There was equal amount of HSV-1 and HSV-2 seropositive women with positive history of genital herpes infection.
The unexpected feature of this study in addition to the low seroprevalences of HSV and CMV was the low incidence of primary CMV infections and the lack of primary HSV infections. This indicates that most parturient women in this study lived in stable relationships and that infections transmitted sexually were less frequent than previous studies have lead us to believe, thereby emphasising the difficulty of predicting rates of seroprevalence or incidence of infection based on studies done elsewhere.