Apheresis treatment of cryoglobulinemic vasculitis: A multicentre cohort study of 159 patients.

OBJECTIVE: To assess the effectiveness of apheresis therapy (AT) in treating the clinical manifestations of patients with complicated cryoglobulinemic vasculitis (CV). METHODS: A retrospective cohort study of 159 CV patients attending 22 Italian Centers who underwent at least one AT session between 2005 and 2015. The response to AT was evaluated on the basis of a defined grading system. RESULTS: Peripheral neuropathy was the most frequent clinical condition leading to AT. Therapeutic plasma exchange was used in 70.4% of cases. The outcome of AT was rated very good in 19 cases, good in 64, partial/transient in 40, and absent/not assessable in 36. Life-threatening CV-related emergencies and renal impairment independently correlated with failure to respond to AT. The independent variables associated with an increased risk of death were age at the time of the first AT session, multi-organ life-threatening CV, the presence of renal impairment and failure to respond to AT. The time-dependent probability of surviving until CV-related death in the second year was 84%, with an AHR in patients with absent/not assessable response to AT of 11.25. CONCLUSION: In this study AT is confirmed to be a safe procedure in patients with CV. Early AT should be considered in patients with severe CV, especially in cases with impending renal involvement, in order to prevent irreversible kidney damage. Although its efficacy in patients with multi-organ failure is limited, AT is the only treatment that can rapidly remove circulating cryoglobulins, and should be considered an emergency treatment.

Prenatal diagnosis of congenital anomalies.

Up till the early 1970s, prenatal diagnosis of congenital anomalies was primarily aimed at detecting chromosomal abnormalities by amniocentesis.1. Over the last two decades, prenatal diagnosis has greatly benefited from advances in ultrasound technology and in our ability to detect microscopic and submicroscopic chromosome abnormalities as well as single gene disorders, leading to substantive improvements in detection of such congenital anomalies.2 At present, invasive prenatal diagnosis continues to be the gold standard for pregnancies at increased risk for chromosomal anomaly or other genetic disease, with chorionic villus sampling being the procedure of choice for the first trimester,3 whereas mid-trimester amniocentesis continues to be the most common form of invasive procedure for prenatal diagnosis.4 Still, invasive techniques are restricted to subgroups at risk for anomalies, for whom such time-consuming procedures are believed to be cost-effective, also accounting for procedure-related abortive risks. In the low-risk population prenatal diagnosis generally consists of screening procedures by means of ultrasound and maternal serum biochemistry.

Bone mineral changes during and after lactation.

OBJECTIVE: To investigate variations in bone mineral density during lactation and throughout the 12 months after scheduled cessation of lactation in relation to the resumption of ovarian function. METHODS: Three hundred eight mothers who decided to lactate were scheduled to fully breast-feed for 6 months, followed by a 1-month weaning period, and then suppress lactation with cabergoline. Their bone mineral density variations were compared with those of a control group of nonlactating mothers during the first 18 months postpartum. Half the lactating women were given daily oral calcium supplements of 1 g in an open design. RESULTS: There was a significant progressive decrease in bone mineral density in lactating women over the first 6 months, followed by recovery of bone mass up to levels that at 18 months were higher than baseline. In nonlactating women, bone mineral density increased progressively after delivery, and at 18 months postpartum had increased by 1.1-1.9% compared with baseline. Compared with lactating women who resumed menstruation within 5 months of delivery, breast-feeding mothers with longer amenorrhea initially lost more bone, but they also gained significantly more bone after resumption of menses, so there were no differences at 18 months postpartum. Oral calcium supplementation decreased bone loss, but had only a transient effect. CONCLUSION: A scheduled lactation period of 6 months, followed by a 1-month weaning period, allowed bone mineral density to reach higher values compared with early postpartum, regardless of calcium supplementation and duration of postpartum amenorrhea.

Long-term sequential treatment with combined estradiol valerate and cyproterone acetate in early postmenopause.

BACKGROUND: We sought to evaluate the long-term safety and efficacy of cyclic combined estradiol valerate and cyproterone acetate. METHODS: A six-year, single center, open study was performed of forty-eight recently postmenopausal women. We evaluated biochemical markers of bone turnover, bone mineral content, bone mineral density, serum lipids, climacteric symptoms, and bleeding throughout the six-year period. Ultrasound evaluation of the endometrium was performed yearly for the first two years. RESULTS: Markers of bone turnover were decreased over the study period, while bone mineral content and density remained unchanged. High density lipoprotein cholesterol levels were increased with unchanged blood pressure over the six years. Menopausal symptoms showed significant improvements within 6 months. Endometrial scan results never required hysteroscopic evaluation. CONCLUSION: Sequentially combined estradiol valerate and cyproterone acetate can be used for an indefinite number of years in women with recent natural menopause. Cyproterone acetate provides adequate endometrial protection and cycle control while allowing estradiol valerate to exert its positive effects on lipid profile, bone metabolism, and climacteric symptoms.

Effect of antibiotic use on the prevalence of symptomatic vulvovaginal candidiasis.

OBJECTIVE: The purpose of the study was to evaluate the influence of antibiotic use on the prevalence of symptomatic vulvovaginal candidiasis. STUDY DESIGN: This is a case-control study of 684 women with symptomatic vulvovaginal candidiasis who were enrolled at a vaginitis clinic and 901 control subjects who attended a cytologic screening service. RESULTS: The prevalence of antibiotic use in the month preceding the visit was 19.3% (132/684) among patients with Candida infection compared with 11.9% (107/901) among control subjects (P <.001). After adjustment by logistic regression for potential confounders (age, marital status, and contraceptive method used), the odds ratio for vulvovaginal candidiasis among patients who reported antibiotic use in the month preceding the visit was 1. 75 (95% confidence interval 1.33 to 2.32). No differences in risk were found by type of antibiotic used. The likelihood of vulvovaginal candidiasis was directly related to the duration of antibiotic use and was higher in patients who had a history of repeated episodes of Candida vaginal infection. CONCLUSIONS: Antibiotic use is a short-term risk factor for symptomatic vulvovaginal candidiasis, either as a first episode or in the form of recurrence. Increasing duration of antibiotic use is directly related with an increased prevalence of Candida vaginal infection.

Obstetric risk factors for periventricular leukomalacia among preterm infants.

OBJECTIVE: To evaluate the obstetric antecedents of cystic periventricular leukomalacia and transient echodense periventricular lesions among preterm infants. DESIGN: A cohort study of preterm singleton infants born between 25 and 33 weeks gestation. SETTING: Pavia, Italy. POPULATION: Three hundred and forty-nine infants admitted to a Division of Neonatal Intensive Care who were screened for periventricular leukomalacia. METHOD: The obstetric factors in infants with either cystic periventricular leukomalacia or transient echodense periventricular lesions were compared to those in infants with negative cranial ultrasonographic findings. Stepwise multiple logistic regression analysis was used to evaluate the association between risk factors and outcomes adjusting for confounders. RESULTS: The prevalence of cystic periventricular leukomalacia and transient echodense lesions was 5.7% (20/349) and 14% (49/349), respectively. The main risk factors for cystic leukomalacia were first trimester haemorrhage (OR 4.49; 95% CI 1.63-12.39), maternal urinary tract infection on admission (OR 5.71; 95% CI 1.91-17.07), and neonatal acidosis (pH < 7.2) at birth (OR 5.97; 95% CI 1.93-18.52). Meconium-stained amniotic fluid (OR 3.95; 95% CI 1.42-10.98) and long term (> 72 hours) ritodrine tocolysis (OR 2.54; 95% CI 1.28-5.05) were associated with an increased risk of echodense lesions. The likelihood of overall leukomalacia (cystic plus echodense periventricular lesions) was increased among cases with meconium-stained amniotic fluid (OR 4.06; 95% CI 1.65-10.0), long-term ritodrine tocolysis (OR 2.56; 95% CI 1.38-4.72), maternal infection (OR 1.73; 95% CI 1.0-3.0), and acidosis at birth (OR 1.98; 95% CI 1.0-3.98). CONCLUSIONS: This study confirms that maternal infection, acidosis at birth, and meconium-stained amniotic fluid increase the risk of periventricular leukomalacia in preterm infants. Long-term ritodrine use seems to increase the risk for transient echodense lesions.

Adequacy of screening cervical cytology among human immunodeficiency virus-seropositive women.

OBJECTIVE: To evaluate the adequacy of cytology in detecting cervical intraepithelial neoplasia (CIN) among human immunodeficiency virus (HIV)-seropositive women compared to controls. METHODS: A cross-sectional study was carried out evaluating 241 HIV-seropositive women and 991 controls (404 HIV seronegative and 587 of unknown HIV status) at risk for CIN attending a vaginitis clinic. All patients had a Pap smear and a standard colposcopic examination of the lower genital tract. Cervical biopsies were taken as indicated by colposcopy. Cytology and histology slides were read by a cytopathologist blinded to patients' serostatus. False-negative cytologic cases were reviewed by three independent cytopathologists to estimate sampling and screening error rates. Sensitivity, specificity, and false-negative rate of cytologic smears were compared between HIV seropositives and controls. We estimated the sampling and screening error rates among cases with false-negative cytology. RESULTS: Among seropositives, the sensitivity, specificity, and the false-negative smear rate for CIN were 73.4% (47/64), 97.1% (134/138), and 26.6% (17/64), respectively. The corresponding figures in controls were 83.8% (83/99), 99.04% (825/833), and 16.2% (16/99), respectively, and did not differ significantly from those of seropositives. The negative predictive value of cytology was lower among seropositives (134/151) than in controls (825/841, chi2 = 34.8, P < .001). The agreement between cytologic readings and combined colposcopy and histology was stronger among controls (kappa = 0.789, 95% CI 0.723 to 0.856) than among seropositives (kappa = 0. 593, 95% CI 0.475 to 0.712). Three independent cytopathologists were unable to detect atypical cells in 52.9% (9/17) of false-negative smears taken from seropositive women as opposed to 37.5% (6/16) of controls. CONCLUSIONS: The sensitivity, specificity, and false negative rate of screening cytology for CIN among HIV seropositive women are comparable with those in the general population. Since almost 50% of false-negative results could be attributed to sampling errors, more frequent cytological screening may prove to be beneficial to this high-risk group.

Preeclampsia, preterm delivery and infant cerebral palsy.

OBJECTIVE: To evaluate the association between preeclampsia and cerebral palsy among preterm infants. STUDY DESIGN: A cohort study of 345 singleton preterm (24 to 33 weeks gestation) infants delivered at an institution where no mothers received magnesium sulphate. We investigated the relationship of preeclampsia to the development of infant cerebral palsy (spastic quadriplegia, hemiplegia or diplegia) at two years' corrected age. Potential confounders were controlled for by politomous logistic regression analysis. RESULTS: The prevalence of cerebral palsy in surviving infants from normotensive and preeclamptic pregnancies was 13.4% (38/284) and 3.3% (2/61), respectively (P=0.026 by Fisher exact test). After control for potential confounders (gestational age, infant gender, birthweight standard deviation score, mode of delivery), the likelihood of infant cerebral palsy was reduced by preeclampsia (odds ratio 0.16, 95% confidence interval=0.04-0.74 for infants from preeclamptic versus normotensive pregnancies being palsied). CONCLUSION: Among preterm infants, the protective effect of preeclampsia against cerebral palsy is manifest in the absence of magnesium sulphate utilization.

Risk for spontaneous preterm delivery by combined body mass index and gestational weight gain patterns.

BACKGROUND: To investigate whether maternal anthropometric factors interact with one another or with other risk factors, thus modifying the risk of spontaneous preterm delivery. METHODS: We carried out a case-control study of 230 spontaneous preterm births with intact membranes between 24 and 35 weeks gestation and 460 control term births. All the patients had prenatal care at the same institution. Logistic regression analysis was used to test for possible interactions adjusting for potential confounders. RESULTS: A pre-pregnancy body mass index < or = 19.5 Kg/m2 (odds ratio (OR) = 1.68, 95% confidence interval (CI) = 1.20-2.38) and a rate of weight gain < or = 0.37 Kg/week during the second and third trimesters (OR = 2.4, 95% CI = 1.69-3.42) were associated with an increased risk of spontaneous preterm delivery. The risk of spontaneous preterm delivery associated with a low second/third trimester weight gain was greater among patients with a body mass index < or = 19.5 (OR = 5.63, 95% CI = 2.35-13.8) compared to those with a body mass index > 19.5 (OR = 2.45, 95% CI = 1.60-3.75, adjusted p value for interaction = 0.05). The risk of spontaneous premature delivery associated with a maternal pre-pregnancy weight < or = 48 Kg was higher among smokers (OR = 5.81, 95% CI = 1.60-22.9) than among non-smokers (OR = 2.4, 95% CI = 1.53-3.74, adjusted p value for interaction = 0.05). CONCLUSIONS: The risk of spontaneous preterm delivery associated with a low pre-pregnancy body mass index is greater among patients with low rate of gestational weight gain during the second and third trimesters compared to those with a higher rate. The results of this study support the recommendation for increased rates of weight gain to patients with low body mass index compared to those with a higher body mass index.

Meconium-stained amniotic fluid and risk for cerebral palsy in preterm infants.

OBJECTIVE: To estimate the risk for cerebral palsy in preterm infants in relation to the presence of meconium in the amniotic fluid (AF). METHODS: A cohort study was conducted of 404 consecutive preterm infants delivered between 24 and 33 weeks' gestation at a single institution. Sociodemographic and clinical data were collected at birth. The diagnosis of cerebral palsy was made at 2 years' corrected age. Politomous logistic regression models were used to evaluate the odds for cerebral palsy while adjusting for potential confounders. RESULTS: The overall prevalence of cerebral palsy among survivors was 11.6% (40/345). The cerebral palsy rate was 41.2% (7/17) among infants who were meconium-stained at birth and 10% (33/328) among those who were not (P = .006 by Fisher exact test). After adjustment for potential confounders (gestational age and fetal gender), the odds ratio of cerebral palsy among infants delivered to women with meconium-stained AF was 6.9 (95% confidence interval 2.32, 20.81, P = .001) relative to those delivered to women with clear AF. CONCLUSION: The results of the present study support the view that the presence of meconium in the AF is a gestational age-independent risk factor for cerebral palsy among preterm infants.

Significance of low birthweight for gestational age among very preterm infants.

OBJECTIVE: To estimate the risk of specific adverse neonatal events resulting from the combined effects of prematurity and low birthweight in very preterm infants (delivered at 24-31 weeks of gestation). DESIGN: A cohort study of specific adverse neonatal events in preterm infants born at between 24 and 31 weeks of gestation. SETTING: Pavia, Italy. POPULATION: Two hundred and thirty singleton infants with sonographically confirmed gestational age, delivered at 24 to 31 weeks of gestation. METHODS: To evaluate the impact of a lower than expected birthweight on selected neonatal events independently of gestational age, we calculated birthweight standard deviation scores (differences between actual birthweight and fitted birthweight divided by fitted standard deviation) for each week of gestation. RESULTS: After adjustment for gestational age and other confounders, there was a significant linear trend relating a decreasing birthweight SDS to an increased likelihood of neonatal death, intraventricular haemorrhage, severe respiratory distress syndrome, and acidosis. Compared with infants with SDS > or = 0 (> or = 50th centile of birthweight), infants with birthweight SDS < -1 (< 16th centile) had increased odds for neonatal death [odds ratio (OR) 3.7, 95% confidence interval (CI) 1.42-9.6], grade III-IV intraventricular haemorrhage (OR 17.5, 95% CI 4.04-75.9), and neonatal acidosis (OR 3.22, 95% CI 1.41-7.4). The significance of birthweight SDS as a predictor of neonatal outcome, however, was lower than that of gestational age. CONCLUSIONS: A lower than expected birthweight affects the likelihood of several adverse neonatal events in very preterm infants. However, a decreasing birthweight SDS affects neonatal outcome less than decreasing gestation does.

Antenatal and delivery risk factors simultaneously associated with neonatal death and cerebral palsy in preterm infants.

To evaluate the simultaneous effects of antenatal and delivery risk factors on neonatal death and cerebral palsy in preterm infants, we conducted a cohort study of 363 singleton pregnancies delivered between 24 and 33 weeks gestation. Neurodevelopmental outcome of the infants was evaluated at 2 years of corrected age. Risk factors associated with death or cerebral palsy were analysed by politomous logistic regression. Overall, the mortality rate was 14.6% (53/363) and the prevalence of cerebral palsy among surviving infants was 12.3% (38/310). Decreasing gestation and meconium-stained amniotic fluid were the only antenatal factors associated with increased odds for both death and cerebral palsy. The effect magnitude and the predictive value of gestational age were greater for death than for cerebral palsy. After adjustment for confounders, prolonged (> or = 48 h) rupture of membranes (odds ratio 2.98, 95% confidence interval 1.12-7.96) and male sex of the infant (odds ratio 3.01, 95% confidence interval 1.32-6.71) were significantly associated only with cerebral palsy. We conclude that neonatal death and cerebral palsy share few common antenatal risk factors. The characteristics of antenatal risk factors for cerebral palsy suggest that bacterial infestation of the amniotic cavity may be implicated in the etiology of the cerebral impairment.

Prevalence of and risk factors for fungal vaginitis caused by non-albicans species.

OBJECTIVE: Our purpose was to evaluate the prevalence of symptomatic yeast vaginitis caused by non-albicans species among patients attending a vaginitis clinic over an 8-year period. STUDY DESIGN: A retrospective study of 1263 patients with symptomatic yeast vaginitis confirmed by culture techniques was performed. RESULTS: The prevalence of symptomatic fungal vaginitis caused by non-albicans species increased from 9.9% (10/101) in 1988 to 17.2% (36/209) in 1995 (chi 2 for trend = 9.33, p = 0.002). Non-albicans species were found more frequently in known human immunodeficiency virus-seropositive patients (23/102 vs 143/1161, odds ratio 2.07, 95% confidence interval 1.2 to 3.46) than in seronegative subjects or subjects of unknown status for the virus. Recurrent vaginal candidiasis was an additional risk factor for vaginitis caused by non-albicans species (odds ratio 2.47, 95% confidence interval 1.72 to 3.52). The increase in non-albicans isolates during the study period was confirmed in stratified analysis and in the subgroup of self-referred patients with no history of either human immunodeficiency virus infection or recurrent vaginal candidiasis. CONCLUSION: The prevalence of fungal vaginitis caused by non-albicans species has increased sharply in the setting of a vaginitis clinic. The characteristics of risk factors suggest that fungal cultures should be done routinely in human immunodeficiency virus-seropositive subjects with suspected vaginal candidiasis and in patients with recurrent vaginal infection.

The effect of work activity in pregnancy on the risk of fetal growth retardation.

BACKGROUND: The relationship between physical activity at work and risk of fetal growth restriction is controversial. For the most part, previous studies investigated the effect of work activity on birthweight alone. We evaluated the impact of type of occupation and physical effort at work on the risk of ultrasonographically confirmed fetal growth retardation among nulliparous women. METHODS: We compared the characteristics of work and the intensity of occupational fatigue (work posture, weekly working hours, physical effort at work) in 349 patients with ultrasonographically confirmed fetal growth retardation and 698 control pregnancies with appropriate fetal growth. Physical demands at work were evaluated by interview at birth. Logistic regression analysis was used to evaluate the association of employment status, type of occupation, and intensity of occupational fatigue with the risk of fetal growth retardation, correcting for potential confounders (maternal age, pre-pregnancy body mass index, mean weight gain in pregnancy, education, partner's social status, smoking in pregnancy, alcohol use, illicit drug use, time of stopping work, and hypertension). RESULTS: After adjusting for confounding, the risk of fetal growth retardation was similar between unemployed and formally employed women at the beginning of pregnancy (OR = 1.26; 95% confidence interval = 0.86 - 1.83). However, manual workers were at slightly higher risk of IUGR than not formally employed women (OR = 1.81; 95% CI = 1.15 - 2.85). Among formally employed women, standing or walking at work, and working > or = 30 hours a week were not significantly associated with IUGR. Finally, the risk of IUGR was significantly higher (OR = 2.46; 95% CI = 1.36 - 4.21) among women reporting moderate-to-heavy as compared to light physical effort at work. CONCLUSION: Formal employment at the beginning of pregnancy is not associated with an increased risk of IUGR. However, moderate-to-heavy physical effort at work seems to increase the risk of sonographically confirmed fetal growth retardation.

Effect measures for behavioral factors adversely affecting fetal growth.

The purpose of this study was to estimate the proportion of ultrasonically diagnosed fetal growth retardation that may be attributable to potentially remediable factors (for example, cigarette smoking, alcohol use, and drug addition). We computed multivariate odds ratios for fetal growth retardation associated with cigarette smoking and alcohol and illicit drug consumption in 350 singleton pregnancies complicated by fetal growth retardation and 700 controls. The odds ratio of fetal growth retardation among women who smoked throughout pregnancy was 2.61 (95% confidence interval, 1.38 to 4.93) compared to women who claimed to have stopped smoking by the 18th week of gestation. The summary attributable risk of fetal growth retardation associated with behavioral variables measured during the first trimester and thereafter ranged from 18 to 21%. By implication, the maximum proportion of fetal growth retardation that could theoretically be prevented by eliminating these risk factors accounts for about 1% of all births. However, allowance for the limited modifiability of preventable factors and the consistent overlap between them would probably reduce such estimate to less than 1%. We conclude that only interventions to reduce the prevalence of strong risk factors (for example, preeclampsia, sociodemographic variables) may have a substantial effect on the incidence of impaired fetal growth and subsequent infant morbidity.

[Genital infections as a cause of abortion in the first trimester of pregnancy. Review of the literature]. / Le infezioni genitali quali causa di aborto nel primo trimestre di gravidanza. Revisione della letterature.

The rate of spontaneous abortion from embryo-fetal infection is believed to range from 10 to 15%. This figure, however, is likely to be underestimated because of subclinical abortions occurring at the earliest stages of gestation that go undiagnosed. The actual frequency of infections resulting in 1st trimester spontaneous abortion is therefore unknown, whereas the impact of infectious agents in the TORCH group on the pathogenesis of such event is well known, although the pathogenetic mechanisms are not well understood. A wide range of microorganisms have been isolated by several investigators from the lower genital tract of women with 1st trimester spontaneous abortion. No causative relationship of HIV-1 infection to spontaneous abortion has been established yet. The diagnosis of abortion from infection can only be made retrospectively based on histologic examination of fetal and placental tissue and isolation on culture of the infectious agent assumed to have caused infection.

Effect of preterm premature rupture of membranes on neurodevelopmental outcome: follow up at two years of age.

OBJECTIVE: To evaluate the impact of preterm premature rupture of membranes on the neurodevelopmental outcome of infants, assessed at two years of age. DESIGN: A prospective observational study of surviving preterm infants born after premature rupture of membranes and of infants born after spontaneous preterm labour with intact membranes. The study was carried out in the period 1986 to 1991. SETTING: Pavia, Italy. SUBJECTS: One hundred and forty singleton infants born prematurely after premature rupture of membranes between 24 and 34 weeks of gestation and 120 controls of similar gestational age born after spontaneous preterm labour with intact membranes. MAIN OUTCOME MEASURES: Infant neurodevelopmental outcome at two-year follow up. RESULTS: After adjustment, by logistic analysis for the effect of gestational age and birthweight, infants born after premature rupture of membranes were more likely to have severe neurodevelopmental impairment (spastic tetraplegia and/or Bayley mental developmental index < 71) than controls (adjusted OR 5.75, 95% CI 1.22-27.18). Multivariate analysis of linear trend showed a statistically significant relation of duration of membrane rupture to occurrence of severe intraventricular haemorrhage, cystic periventricular leucomalacia and moderate to severe infant neurodevelopmental impairment. CONCLUSION: Infants born after prolonged premature rupture of membranes are at higher risk of subsequent moderate to severe neurodevelopmental impairment than those born after spontaneous labour with intact membranes.

The effect of work activity in pregnancy on the risk of severe preeclampsia.

The aim of this study was to evaluate the impact of type of employment and level of physical activity at work on the risk of severe preeclampsia. For this purpose, we carried out a case control study of 160 nulliparous pregnant women with severe preeclampsia and 320 normotensive nulliparous controls who received prenatal care from members of one medical staff. The type of employment and the level of physical activity sustained at work were ascertained at birth through an interview based on a standard questionnaire. The degree of physical activity at work was assessed by a 4-level activity score based on type of work, physical intensity, posture at work, and weekly working hours. In logistic regression analysis, after adjustment for potential confounders (maternal age, time of stopping work, prepregnancy body mass index, social status of the partner, history of previous abortion, and prepregnancy smoking status), clerical workers had a significantly lower risk of severe preeclampsia than women who were unemployed at the beginning of pregnancy (OR 0.53, 95% CI 0.30, 0.96). In multivariate analysis, there was a significant linear trend relating the degree of physical activity at work, to the risk of preeclampsia (likelihood chi-square = 9.38, 3 df, p = 0.002). We then restricted the analysis to women who had ever worked in pregnancy (n = 339) also adjusting for confounders, and found that clerical workers were still at significantly lower risk of severe preeclampsia than women not formally employed (OR 0.2, 95% CI 0.08, 0.49). In addition, moderate/high physical activity at work was associated with a 2-fold increase in the risk of severe preeclampsia compared to mild activity (OR 2.08, 95% CI 1.11, 3.88). We conclude that moderate to high physical activity at work seems to increase the risk of severe preeclampsia.