Acta Obstet Gynecol Scand 2001; 80: 1036–1038 Copyright C Acta Obstet Gynecol Scand 2001

Printed in Denmark ¡ All rights reservedActa Obstetricia et

Gynecologica ScandinavicaISSN 0001-6349

ORIGINAL ARTICLE

Is preeclampsia an infectious disease?LILL I. S. TROGSTAD1,2, ANNE ESKILD1,2, ANNE-LISE BRUU3, STIG JEANSSON4 AND PÅL A. JENUM5

From the Departments of 1Population Health, 3Virology, and 5Bacteriology Sciences, National Institute of Public Health, Osloand the Departments of 2Obstetrics and Gynecology, and 4Microbiology, Ullevaal University Hospital, Oslo, Norway

Acta Obstet Gynecol Scand 2001; 80: 1036–1038. C Acta Obstet Gynecol Scand 2001

Background. Studies have suggested a strong paternal factor in the etiology of preeclampsia.If preeclampsia is caused by an infectious agent transmitted by the woman’s partner, seronega-tive women who may experience primary infection in pregnancy should be at increased riskof preeclampsia as compared to previously infected women. The aim of this study was toassess the impact of being seronegative for some viruses transmitted by close contact on therisk of developing preeclampsia.Methods. Nine hundred and seventy-eight women were randomly drawn from a basic studypopulation of 35,940 pregnant women in Norway. A serum sample drawn at the first antenatalvisit was analyzed for specific IgG antibodies against herpes simplex virus type-2, cytomegalo-virus and Epstein-Barr virus. For comparison, antibody status against Toxoplasma gondiiwas also assessed. Information on preeclampsia in pregnancy was obtained through linkageto the Medical Birth Registry of Norway.Results. Thirty-three (3%) women developed preeclampsia. The risk of developing preeclamp-sia seemed to be increased for women who were seronegative for the viruses studied. Seronega-tivity for Toxoplasma gondii did not show such a pattern.Interpretation. Women who are seronegative for antibodies against viral agents transmittedthrough close contact seem more likely to develop preeclampsia. This finding indicates thatwomen who are seronegative to such agents may acquire primary infection in pregnancy, andsubsequently be at increased risk of preeclampsia. This hypothesis could represent a newapproach to the causes of preeclampsia, and encourage search for yet unidentified microbesas a possible causal factor.

Key words: antibodies; infection; preeclampsia

Submitted 30 March, 2001Accepted 6 July, 2001

New attention is brought to infectious agents aspossible causes of diseases that earlier were con-sidered to be caused by genetic disposition or life-style factors. Examples are human papillomavirusin the development of cervical carcinoma, Helico-bacter pylori in gastric ulcer, and Chlamydia pneu-moniae and cytomegalovirus in coronary heart dis-ease and atherosclerosis. The mechanisms de-scribed in the development of atherosclerosis andin acute rejection following organ transplantationsresemble the pathophysiological mechanisms de-

Abbreviations:HSV-2: herpes simplex virus type-2; EBV: Epstein-Barr virus;CMV: cytomegalovirus; EIA: enzyme immunoassay.

C Acta Obstet Gynecol Scand 80 (2001)

scribed in preeclampsia, and have been associatedwith infectious agents (1, 2).

Also, inflammatory changes similar to those ofsepsis have been described in normal pregnancies,and preeclamptic pregnancies seem to be associ-ated with an exaggeration of such changes (3). Theadditive effect of other conditions producing in-flammatory responses in pregnancy, for instancecertain autoimmune conditions and infections,have been suggested to precipitate or exacerbatethe signs of preeclampsia (4, 5).

Some indicators may support a hypothesis ofpreeclampsia being caused by an infectious agenttransmitted by the woman’s partner. If a womanbecomes pregnant by a man who has already

Antibody status (IgG) and preeclampsia risk 1037

Table I. Presence of IgG antibodies in the first trimester in 978† pregnant women according to preeclampsia status at delivery and odds ratio for developingpreeclampsia

Preeclampsia

Yes (%) No (%) Crude odds ratio Adjusted* odds ratioPresence of IgG antibodies nΩ33 nΩ945† (95% CI) (95% CI)

Herpes simplex virus type 2No 27 (82) 683 (73) 1.7 (0.7–4.1) 1.7 (0.7–4.2)Yes 6 (18) 253 (27) 1.0 1.0

CytomegalovirusNo 14 (42) 282 (30) 1.7 (0.8–3.4) 1.6 (0.8–3.2)Yes 19 (58) 650 (70) 1.0 1.0

Epstein-Barr virusNo 4 (12) 36 (4) 3.4 (1.1–10.1) 3.5 (1.1–10.6)Yes 29 (88) 879 (96) 1.0 1.0

Toxoplasma gondiiNo 30 (91) 854 (90) 1.1 (0.3–3.6) 1.0 (0.3–3.5)Yes 3 (9) 91 (10) 1.0 1.0

* Adjusted for age and parity. † Varying numbers of women in the non-preeclampsia group are due to insufficient serum volume from some women.

fathered a pregnancy with preeclampsia in a differ-ent woman, her risk of developing preeclampsia isincreased (6). A similar observation was made con-cerning the partner’s impact on the risk of cervicalcarcinoma (7). The risk of developing preeclampsiadecreases according to the duration of sexual co-habitation before conception (8, 9), and womenusing barrier contraceptives prior to conceptionhave more than twice the risk of developing preec-lampsia as compared to women using non-barriercontraceptives (9).

We hypothesize that preeclampsia is caused by atransmittible agent, and that women who developpreeclampsia more often lack immunity againstthis agent early in pregnancy. The objective of thisprospective study of 978 pregnant women wastherefore to study the impact of seronegativity inearly pregnancy for antibodies (IgG) against vi-ruses transmitted through close contact on the riskof developing preeclampsia; herpes simplex virustype-2 (HSV-2), cytomegalovirus (CMV) andEpstein-Barr virus (EBV). For comparison, pres-ence of IgG antibodies against Toxoplasma gondii,an intracellular parasite transmitted mainly byingestion of contaminated food, was also studied.

Patients and methods

The source study population comprised 35,940pregnant women in Norway 1992–94 (10). Almost100% of the pregnant women in 11 counties inNorway were included, representing 60% of allpregnant women in the country during the studyinclusion period. A serum sample was drawn at the10th week of pregnancy. Through linkage to the

C Acta Obstet Gynecol Scand 80 (2001)

Medical Birth Registry of Norway information onpreeclampsia was obtained (6). After linkage a ran-dom study sample of 978 women was drawn.

The women’s mean age was 28.7 years (median28 years, range 16–47 years). Four hundred andseventeen women (43%) had no previous births re-ported to the Medical Birth Registry. Three hun-dred and sixty-five women (37%) had one, and 196women (20%) had two or more previous births re-ported. All women with a specified diagnosis ofpreeclampsia, or pregnancy induced hypertension(blood pressure reaching Ø140/90 mmHg after 20weeks gestation) and co-existing proteinuria (Ø0.3g/24hours or Øπ1 by dipstick) as reported to theMedical Birth Registry, were classified as havingpreeclampsia (6).

The first trimester serum samples were analyzedfor specific IgG antibodies against HSV-2 (11),CMV (CMV-IgG ELA-Test, Medac, Hamburg,Germany), EBV (anti EBV recombinant EA IgGELISA and anti EBV recombinant EBNA IgGELISA, Biotest Diagnostics, Dreieich, Germany),and Toxoplasma gondii (Platelia Toxo-IgG, SanofiDiagnostics Pasteur, Marnes la Coquette, France).

The risk of developing preeclampsia accordingto antibody status was estimated as odds ratios inunivariate and multivariate logistic regressionmodels. In the multivariate model the women’s ageand parity were included as potential confoundingvariables.

Results

Thirty-three (3%) of the 978 women developed pre-eclampsia. The first trimester prevalence of IgG

1038 L. I. S. Trogstad et al.

antibodies was 10% for antibodies against Toxo-plasma gondii, 27% against HSV-2, 69% againstCMV and 96% for antibodies against EBV. Theseroprevalences of virus antibodies were lower inthe women who developed preeclampsia as com-pared to women who did not develop preeclamp-sia, whereas the prevalence of antibodies againstToxoplasma gondii showed no such difference(Table I). The adjusted odds ratios of preeclampsiawere above 1.0 for being seronegative to HSV-2,CMV and EBV, but only significantly increased forEBV (Table I). Being seronegative for Toxoplasmagondii was not associated with an increased risk ofdeveloping preeclampsia.

Discussion

The results in our study suggest that women whoare seronegative for antibodies against virusestransmitted by close contact are at increased riskof developing preeclampsia during pregnancy.Seronegative women are susceptible to acquire pri-mary infections, our observation may thereforesupport a hypothesis of some transmittible agentplaying a role in the development of preeclampsia.The viral agents included in this study may serve asindicators of one or more other infectious agents,spreading in the same manner which may contrib-ute to the development of preeclampsia. The lackof association between Toxoplasma gondii anti-body status and preeclampsia supports such hy-pothesis.

Our study included relatively few women. Thusthe confidence intervals around the risk estimatesare wide, and the statistical power is limited. Webelieve, however, that our observations are worth-while reporting. Our preliminary findings couldrepresent a new approach to the causes of preec-lampsia, and encourage search for yet unidentifiedmicrobes as a possible causal factor.

Acknowledgments

We are grateful to the National Institute of Public Health andthe National Board of Health, Norway for funding this re-search.

The efforts of the staff at the Medical Birth Registry of Nor-way are gratefully acknowledged.

C Acta Obstet Gynecol Scand 80 (2001)

References

1. Sabate M, Manito N, Cequier A, Roca J, Mauri J, Gomez-Hospital JA et al. Acute rejection, cytomegalovirus infec-tion and endothelial dysfunction early after heart trans-plantation. Transplant Proc 1995; 27: 2346–8.

2. Anderson JL, Muhlestein JB, Carlquist J, Allen A, TrehanS, Nielson C et al. Randomized secondary prevention trialof azithromycin in patients with coronary artery diseaseand serological evidence for chlamydiae pneumoniae infec-tion: the azithromycin in coronary artery disease: elimin-ation of myocardial infection with chlamydia (ACA-DEMIC) study. Circulation 1999; 99 (12): 1540–7.

3. Sacks GP, Studena K, Sargent IL, Redman CWG. Normalpregnancy and preeclampsia both produce inflammatorychanges in peripheral blood leukocytes akin to those of sep-sis. Am J Obstet Gynecol 1998; 179: 80–6.

4. Redman CWG, Sacks GP, Gavin P, Sargent IL. Preeclamp-sia: An excessive maternal inflammatory response to preg-nancy. Am J Obstet Gynecol 1999; 180: 499–506.

5. Faas MM, Schuiling GA, Baller JFW, Visscher CA, BakkerWW. A new animal model for human preeclampsia: ultra-low-dose endotoxin infusion in pregnant rats. Am J ObstetGynecol 1994; 171: 158–64.

6. Lie RT, Rasmussen S, Brunborg H, Gjessing HK, Lie-Niel-sen E, Irgens LM. Fetal and maternal contributions to therisk of pre-eclampsia: population based study. BMJ 1998;316: 1343–7.

7. Kessler II. Human cervical cancer as a venereal disease.Cancer Res 1976; 36: 783–91.

8. Robillard P-Y, Hulsey TC, Perianin J, Janky E, Miri EH,Papiernik E. Association of pregnancy-induced hyperten-sion with duration of sexual cohabitation before concep-tion. Lancet 1994; 344: 973–5.

9. Klonoff-Cohen HS, Savitz DA, Cefalo RC, McCann MF.An epidemiologic study of contraception and preeclampsia.JAMA 1989; 262: 3143–7.

10. Jenum PA, Stray-Pedersen B, Melby K, Kapperud G,Whitelaw A, Eskild A et al. Incidence of toxoplasma gondiiinfection in 35,940 pregnant women in Norway and preg-nancy outcome for infected women. J Clin Microbiol 1998;36: 2900–6.

11. Ho DWT, Field PR, Sjøgren-Jansson E, Jeansson S, Cun-ningham AL. Indirect ELISA for the detection of HSV 2specific IgG and IgM antibodies with glycoprotein G (gG-2). J Virol Methods 1992; 36: 249–64.

Address for correspondence:

Lill I. S. Trogstad, M.D.National Institute of Public HealthSection of EpidemiologyDepartment of Population Health SciencesP.O.Box 4404 Nydalen0403 OsloNorwaye-mail: lill.trogstad/folkehelsa.no