Low incidence of hypertensive disorders of pregnancy in women treated with spiramycin for toxoplasma infection

Abstract

Aims

Toxoplasma infection in pregnancy is usually treated with long-term administration of the macrolide spiramycin to prevent fetal malformations. We had empirically observed that treated patients seldom developed pregnancy-induced hypertension (PIH), a common and severe disorder of pregnancy whose aetiology and pathogenesis are still debated. Some clinical and experimental data suggest that infection could play a role in its development.

Methods

To test this hypothesis, we studied a cohort of 417 pregnant women treated with spiramycin because of seroconversion for Toxoplasma gondii and 353 low-risk women who did not take any antibiotic during pregnancy. PIH was defined as blood pressure >140/90 mmHg on two or more occasions, occurring after 20 weeks of gestational age.

Results

Seventeen (5.2%) women in the control group developed PIH compared with two (0.5%) in the case group. The odds of developing the disease were significantly lower in the treated subjects (odds ratio =0.092, 95% confidence interval 0.021, 0.399; P < 0.001).

Conclusions

Our results suggest that antibiotic treatment during pregnancy can reduce the incidence of PIH, thus opening new perspectives in its prevention and therapy.

Introduction

Toxoplasma infection is usually asymptomatic or induces mild, nonspecific symptoms. Primary infection during pregnancy may cause serious fetal effects, but it can be treated, thus reducing the risk of fetal malformations. The Italian National Health Service supports screening for toxoplasmosis in pregnancy. When seroconversion occurs (2–3/1000 pregnancies), the standard treatment is administration of spiramycin (9 × 10⁶ units day⁻¹) until delivery [1, 2]. Spiramycin has an intracellular toxoplasmocidal activity; moreover, like other macrolides, it inhibits protein synthesis by binding the large (50S) subunit of bacterial ribosomes, causing the growing polypeptide chain to dissociate from the ribosome. The antibacterial spectrum of spiramycin is quite broad: it is highly active against many pathogens such as Gram-negative and Gram-positive cocci, Parvobacteriaceae, Legionella spp., Chlamydia spp., Ureaplasma urealyticum, Mycoplasma pneumoniae, Listeria monocytogenes and Streptococcus pneumoniae[3].

We had empirically observed that patients treated throughout gestation for toxoplasma infection seldom developed pregnancy-induced hypertension (PIH), a relatively common disorder of pregnancy. Indeed, there is some evidence that infection can play a role in the pathogenesis of PIH [4–8].

Therefore, we designed a case–control study to evaluate the possible effect of spiramycin therapy during pregnancy on the incidence of PIH. The case group was a cohort of 417 pregnant women treated with spiramycin because of seroconversion for Toxoplasma gondii. The control group included 353 low-risk women who did not take any antibiotic during pregnancy.

Subjects

We analysed the data obtained from all consecutive pregnancies referred to three Italian Maternal-Fetal Medicine Units (University of Turin, University of Naples, University of Monza) because of seroconversion for T. gondii in the first and second trimester of pregnancy between 1990 and 2002. The infection was confirmed in 417 women by an enzyme-linked immunosorbent assay or an immunosorbent agglutination assay for toxoplasma-specific IgG and IgM antibodies. These women were prescribed spiramycin (9 × 10⁶ units day⁻¹) until delivery. Five had an allergic reaction and had to withdraw and nine were not compliant. Therefore, the study group consisted of 403 pregnant women. The mean gestational age at the beginning of treatment was 18.5 weeks (SD 8.1 weeks).

The control group included pregnant women cared for at the outpatient service of the Maternal-Fetal Medicine Unit of the University of Turin and pregnant women recruited within 18 weeks of gestational age at the blood test laboratory of Sant’Anna Hospital, Turin. These were low-risk pregnancies; in particular, no risk factor for preeclampsia (PE) was present [9]. Three hundred and fifty-three pregnancies were considered for the control group. However, 24 controls were excluded from the study: 10 women were lost to follow-up, three decided to terminate the pregnancy and 11 spontaneous abortions occurred. Therefore, the control group consisted of 329 pregnant women.

We collected data on the demographic characteristics and pregnancy outcome in both groups: development of hypertensive disorders, gestational age at birth, neonatal birth weight and sex, Apgar scores recorded 1 and 5 min after delivery, and admission to neonatal intensive care unit.

The study was approved by our Regional and Hospital Ethics Committee and informed consent was obtained from each pregnant woman.

Gestational age was calculated from the last menstrual period and confirmed by an ultrasonographic examination before 20 weeks of gestation. After delivery, the babies were weighed to the nearest 5 g using a digital scale.

PIH was defined as blood pressure >140/90 mmHg on two or more occasions, occurring after 20 weeks of gestational age. PE was defined as blood pressure >140/90 mmHg on two or more occasions, occurring after 20 weeks of gestation together with proteinuria (>0.3 g per 24 h or 3+ on dipstick testing where delivery precluded 24 h collection) [10].

Small for gestational age (SGA) was defined as neonatal weight below the 10th percentile according to Italian standards for birth weight and gestational age [11].

The gestational age and birth weight are presented as mean ± SD. Between-group differences were analysed by one-way anova. The prevalence of pathological outcomes in the groups was compared with the χ² test by means of a 2 × 2 contingency table. Fisher's exact test was used for small sample sizes. Odds ratios (OR) and 95% confidence intervals (95% CI) were used to assess the risk of pathological outcome in pregnancies treated with spiramycin. The impact of maternal age and parity on the risk for hypertensive disorders in pregnancy was obtained with a multivariate logistic regression. Results were considered significant for P < 0.05.

Results

The mean maternal age was 27.5 years (SD 4.6) in the study group and 29.5 years (SD 5.0) in the control group (P < 0.001). There were no other differences in demographic or obstetric characteristics between the two groups (Table 1).

Table 1.

Characteristics of pregnancy in the case and control groups

	Pregnancies with seroconversion (403) Mean (SD) or %	Control pregnancies (329) Mean (SD) or %	P-value
Maternal age at delivery (years)	27.5 (4.6)	29.5 (5.0)	<0.001
Gestational age at treatment (weeks)	18.5 (8.1)	NA	NA
Nulliparous	59.7%	62.3%	NS
Gestational age at delivery (weeks)	39.4 (1.5)	39.3 (1.5)	NS
Preterm delivery	2.7%	3.2%	NS
Multifetal gestation	0.5%	0.6%	NS
Infant birth weight (g)	3293.3 (466.1)	3238.3 (460.9)	NS
Small for gestational age	5.5%	8.4%	NS

NA, Not applicable; NS, not significant.

Two (0.5%) women in the study group developed PIH compared with 17 (5.2%) women in the control group. The odds of developing the disease were significantly lower in the treated subjects (OR = 0.092, 95% CI 0.021, 0.399; P < 0.001). In the study group, the disorder was always PE (2/2), while in the control group there were five cases of PE out of 17 PIH. When we limited the analysis to the PE subgroups, its incidence was lower in the treated subjects (0.5%) than in the controls (1.5%), although the difference was not statistically significant. The odds of developing PE in the treated group were 0.323 (95% CI 0.062, 1.677). Two women in the control group and none in the study group needed preterm delivery because of the disease (Table 2). At the last follow-up visit there was no PIH or PE in the women excluded.

Table 2.

Characteristics of preeclamptic women

	Delivery mode	Gestational age at delivery (weeks)	Fetal growth	Neonatal outcome
Controls
PE 1	Caesarean	35	AGA	Abnormal*
PE 2	Caesarean	39	AGA	Normal
PE 3	Caesarean	34	SGA	Abnormal*
PE 4	Caesarean	40	AGA	Normal
PE 5	Caesarean	41	AGA	Normal
Cases:
PE 6	Vaginal	38	SGA	Normal
PE 7	Caesarean	38	AGA	Normal

AGA, Appropriate for gestational age; SGA, small for gestational age.

^*Low Apgar scores (<7 at 5 min) and/or admission to the neonatal intensive care unit.

Age and parity adjustment did not alter these results (data not shown).

Discussion

Hypertensive disorders of pregnancy affect 12–22% of all pregnancies and remain a major cause of maternal and fetal mortality and morbidity. PE, a more severe form of PIH, has an incidence of 5–8%[10]. Maternal mortality attributable to PE has decreased in developed countries, where diagnosis and management of the disease are a major aim of prenatal care. However, perinatal mortality, perinatal and long-term morbidity and neurological sequelae due to fetal growth restriction and/or preterm delivery are still serious problems [12, 13]. There is no effective way to prevent PE and delivery is the only definitive treatment [14]. This is partly due to the fact that the aetiology and pathogenesis of the disease are poorly understood.

Generalized endothelial dysfunction can explain the symptoms of PIH and its complications (hypertension, proteinuria, liver abnormalities, etc.), which usually appear in the late second or third trimester [15]. However, the pathogenesis of the disease must be sought much earlier, when, at the beginning of pregnancy, an abnormal interaction between trophoblast and decidua occurs, contributing to the impairment of many physiological maternal adaptations to pregnancy [16], including trophoblastic invasion of the uterine spiral arteries. In normal pregnancies, the spiral arteries undergo important changes, which are completed by mid-pregnancy, to meet the needs of the developing placenta [17, 18]. Because of these changes, the utero–placental blood flow increases and the resistance to flow decreases, as can be assessed by ultrasound Doppler studies of the uterine arteries [19]. In women who later develop PIH, there is incomplete invasion of the spiral arteries [17, 20]. The failed trophoblastic invasion of the spiral arteries prevents the decrease in vascular resistance in the uterine vessels, which is reflected by abnormal uteroplacental Doppler findings. Most patients who later develop PE show abnormal uteroplacental Doppler findings at mid-trimester; however, a large proportion of women with abnormal Doppler at mid-trimester will not develop PE [19, 21]. We speculate that the abnormal interaction between trophoblast and decidua, leading to abnormal invasion of the spiral arteries, is necessary, but not sufficient, to explain the development of PE: something else must occur that triggers the cascade of events responsible for the generalized endothelial dysfunction. This could be a clinical or subclinical infection altering the balance between vasorelaxing agents, predominant in normal pregnancy, and vasoconstrictors, prevalent in PE [22, 23]. During normal pregnancy, there is an increase of endothelium-derived nitric oxide (NO) [24], while in PE there seems to be a failure of NO production [25, 26]. This failure could be due to different factors such as a decrease in endothelium-derived nitric oxide synthase (eNOS) expression by infection or/and an increase in NO inactivation by oxidative stress [27, 28]; furthermore, genetic factors could modulate the eNOS expression or activity, thus explaining the impact of family history of PE as a risk factor for PE [9]. We have previously demonstrated in an in vitro study that low levels of lipopolysaccharide and a slight increase of proinflammatory cytokines (mimicking an asymptomatic chronic infection) downregulate eNOS without increasing the inducible isoform (iNOS). This could be the mechanism leading to the lack of vasodilation observed in PE [4]. In contrast, acute infection, characterized by a high concentration of the bacterial endotoxin and very high levels of proinflammatory cytokines, causes increased NO production due to higher iNOS expression and activity [29], even if there is downregulation of eNOS.

The infection hypothesis is also supported by several epidemiological observations. Trogstad et al. showed that women who are seronegative for antibodies against a variety of viruses at the beginning of pregnancy, and therefore more susceptible to infections, are at increased risk of developing PE; however, in contrast to viral agents, they found no association between T. gondii antibody status and PE; their study design was obviously different from ours: the women of our study group were treated with spiramycin because of seroconversion, while the population studied by Trogstad and colleagues were characterized by the presence or absence of IgG antibodies against T. gondii[6]. von Dadelszen and Magee summarized the results of several studies showing some associations between PE and chronic or subclinical infections with viruses, parasites and bacteria [7]. Recently, Heine et al. found an increased IgG seroprevalence to Chlamydia pneumoniae in PE patients [8]; their data were not confirmed in another study [30].

Moreover, the animal model of PE is usually obtained by administration of low doses of endotoxin to pregnant rats [31, 32]. If this is the case, antibiotic treatment could prevent the development of PE.

We designed this study to test the hypothesis that prolonged antibiotic treatment reduces the incidence of PE by prevention of any long-lasting bacterial and/or protozoan infection. We found that treated patients had a lower risk of developing PIH (OR = 0.092) and, to a lesser extent, PE (OR = 0.323). In our study the treatment effect could be underestimated because the study group belongs to the general population, while the control group belongs to a low-risk population. This is underscored by the fact that the incidence of PIH and PE was 5.2% and 1.5%, respectively, in our control group, much lower than the incidence reported in the general population (12–22% and 5–8%, respectively) [10].

Paradoxically, our study group consisted of patients who contracted an infection; however, this occurred early in pregnancy (mean gestational age at diagnosis, 18 weeks), before the symptoms of PE usually develop, and was promptly treated with an effective antibiotic.

In conclusion, our results suggest that long-lasting spiramycin treatment during pregnancy lowers the risk of developing PIH, probably by preventing the onset of infections that could complicate pregnancy. These findings open new perspectives in the prevention of PIH. However, larger prospective studies are required to support the hypothesis that infection plays a role in the pathogenesis of some hypertensive disorders of pregnancy and, ultimately, a randomized trial is required to justify antibiotic treatment as a preventive strategy in clinical practice.