Objective
The population-based incidence of early-onset (<34 weeks) and late-onset preeclampsia (≥34 weeks) has not been adequately studied. We examined the gestational age–specific incidence of preeclampsia onset and identified the associated risk factors and birth outcomes.
Study Design
All singleton deliveries in Washington State, 2003-2008 (n = 456,668), were included, and preeclampsia onset was determined from hospital records linked to birth certificates. Cox and logistic regression models were used to obtain adjusted hazard ratios and odds ratios (AORs) for risk factors and birth outcomes, respectively.
Results
The overall preeclampsia rate was 3.1% and the incidence increased sharply with gestation; early- and late-onset preeclampsia rates were 0.38% and 2.72%, respectively. Among women with early-onset preeclampsia, 12% delivered at a gestation of 34 weeks or longer. Risk/protective factors common to both diseases included older maternal age, Hispanic and Native-American race, smoking, unmarried status, and male fetus. African-American race, chronic hypertension, and congenital anomalies were more strongly associated with early-onset preeclampsia, whereas younger maternal age, nulliparity, and diabetes mellitus were more strongly associated with late-onset disease. Early- but not late-onset preeclampsia conferred a high risk of fetal death (AOR, 5.8; 95% confidence interval [CI], 4.0–8.3 vs AOR, 1.3; 95% CI, 0.8–2.0, respectively). The AOR for perinatal death/severe neonatal morbidity was 16.4 (95% CI, 14.5–18.6) in early-onset and 2.0 (95% CI, 1.8–2.3) in late-onset preeclampsia.
Conclusion
Early- and late-onset preeclampsia shares some etiological features, differ with regard to several risk factors, and lead to different outcomes. The 2 preeclampsia types should be treated as distinct entities from an etiological and prognostic standpoint.
Preeclampsia is characterized by elevated blood pressure and proteinuria or involvement of other organs in an exaggerated systemic inflamation. In industrialized countries, preeclampsia complicates approximately 3-5% of pregnancies and represents one of the most common causes of maternal mortality and severe maternal morbidity including eclampsia, placental abruption, pulmonary edema, and acute renal failure. Infants of mothers with preeclampsia are at approximately 2-fold higher risk of neonatal death and at increased risk of neonatal morbidity including low Apgar scores, seizures, neonatal encephalopathy, and neonatal intensive care admission. However, some previous studies have shown that infants born at very preterm gestation because of preeclampsia have a reduced risk of retinopathy of prematurity, cerebral palsy, and neonatal mortality compared with infants born very preterm for other reasons.
Preeclampsia is a heterogeneous disorder with 2 distinct subtypes that have been described based on the timing of disease onset: early-onset preeclampsia occurring before or at 33 weeks’ gestation and late-onset preeclampsia that occurs at 34 weeks’ gestation or later. Early-onset disease, in particular, confers a high risk of life-threatening maternal complications and fetal demise, and early delivery is the only treatment.
Although previous publications have described early-onset preeclampsia and associated neonatal outcomes, most reports have been based on small, hospital-based studies and clinical trials focusing on obstetric management. The gestational age-specific incidence of preeclampsia, based on the onset of symptoms and not gestational age at delivery, has not been documented to date at the population level. We therefore carried out a population-based study to describe the gestational age-specific incidence of preeclampsia onset among women with singleton pregnancies and to examine risk factors and birth outcomes associated with early-onset and late-onset disease.
Materials and Methods
We included all singleton deliveries in Washington State during the period from 2003 to 2008, utilizing information from 2 large population databases: (1) the Comprehensive Discharge Abstract Database (CHARS) which included all hospitalizations in Washington State, and (2) the Birth Events Record Database (BERD), which included birth records of all live born infants and fetal deaths. Women with a diagnosis of preeclampsia or eclampsia (henceforth referred to as preeclampsia), including preeclampsia superimposed on chronic hypertension were identified from the CHARS database ( International Classification of Diseases , ninth revision [ICD-9] diagnostic codes 642.4, 642.5, 642.6, and 642.7).
Hospitalization records with a diagnosis of preeclampsia were linked to birth records to obtain information about gestational age at delivery, maternal characteristics, clinical risk factors, and birth outcomes. The number of weeks between the hospitalization when the preeclampsia diagnosis was made and birth hospitalization was calculated based on the CHARS and BERD record linkage. Preeclampsia occurring at less than 34 weeks of gestation was identified as early-onset disease, whereas preeclampsia that occurred at 34 weeks or later was labeled late-onset disease, irrespective of the gestational week at delivery. Infant birth records were also linked to CHARS (infant) hospitalization records to identify cases of severe neonatal morbidity (see the following text).
There were 484,111 women who were residents of Washington State and who delivered a singleton stillbirth or live birth in a Washington State hospital between 2003 and 2008. Women with a missing estimate of gestation or gestational age at delivery less than 20 weeks and women without a linkage between the birth/fetal death certificate (BERD database) and maternal hospitalization data (CHARS database) were excluded (5.7%, n = 27,443).
Maternal characteristics and clinical risk factors examined for potential association with preeclampsia included maternal age (younger than 20 and 35 years old or older vs 20-34 years); parity (number of previous live births, none vs 1 or more); marital status (single/widowed/separated vs married/common law); education (less than high school vs high school education or greater); race (non-Hispanic white vs Hispanic, African-American, Native-American, and other); smoking during pregnancy (yes/no); infertility treatment (yes/no); diabetes mellitus (yes/no); chronic hypertension prior to pregnancy (yes/no); infant’s sex (male/female); and congenital anomalies (yes/no).
Fetal death was defined as in utero or intrapartum death of a fetus delivered at 20 weeks’ gestation or later, neonatal death was defined as a death of an infant within 28 days after birth, and perinatal death included fetal or neonatal death. Using birth hospitalization data for infants (obtained from the linked infants’ birth and hospitalization records), the following adverse birth outcomes were identified based on ICD-9 codes: bronchopulmonary dysplasia (BPD; code 770.7), intraventricular hemorrhage (IVH) grade III and IV (codes 772.13 and 772.14), periventricular leukomalacia (PVL; code 779.7), retinopathy of prematurity (ROP; code 362.2), necrotizing enterocolitis (NEC; code 777.5), and neonatal sepsis (code 771.81). Other neonatal outcomes were identified from birth records, namely, neonatal seizures, Apgar score at 5 minutes, and neonatal intensive care unit (NICU) admission.
Severe neonatal morbidity included any of the following: a 5-minute Apgar score of 3 or less, neonatal seizures, BPD, IVH grade III or IV, PVL, ROP, NEC, and neonatal sepsis. The composite outcome of neonatal mortality/morbidity included both neonatal death and severe neonatal morbidity, whereas perinatal mortality/morbidity included perinatal death and severe neonatal morbidity. Small-for-gestational-age (SGA) infants were defined as those weighing less than the 10th percentile of the sex- and gestational age–specific birthweight reference for the United States, whereas large-for-gestational age infants were those weighing over the 90th percentile. We used the clinical estimate of gestation provided in the data source because this is more accurate than gestational age estimated by the last menstrual period.
Gestational age–specific rates of preeclampsia were calculated using ongoing pregnancies as the denominator. χ 2 tests were used to assess the differences between rates of early-onset and late-onset preeclampsia across maternal and clinical characteristics. The Cox regression model, with preeclampsia onset as the outcome and gestational age as the time axis, was used to estimate adjusted hazard ratios (AHRs) and 95% confidence intervals (CIs). This enabled us to create the appropriate risk sets, with censoring of subjects who developed preeclampsia or who delivered at any particular gestation.
When the proportional hazards assumption was not satisfied, we examined the interaction term between the risk factor and gestational age at diagnosis categorized as less than 34 weeks and 34 weeks or longer and obtained AHRs for both early-onset and late-onset preeclampsia separately. The Wald statistic was used to assess statistical significance of the interaction terms. Shoenfeld residuals were used to evaluate the proportional hazards assumption of the final model.
Birth outcomes including fetal death, perinatal death, and severe neonatal morbidity were analyzed using the fetuses-at-risk approach. Under this formulation, all fetuses at a specific gestation were considered at risk for adverse outcomes at that gestation. Thus, for example, all fetuses at 28 weeks with early-onset preeclampsia were considered to be at risk of live birth, stillbirth, neonatal death, or severe neonatal morbidity at 28 weeks, irrespective of whether they actually delivered at 28 weeks or at a subsequent gestational week. Two fetuses-at-risk–based logistic regression models were used to estimate causal associations between early-onset and late-onset preeclampsia and birth outcomes.
These models were constructed using ongoing pregnancies (ie, fetuses at risk) as the denominator ; all ongoing pregnancies at 20 weeks’ gestation were included in models examining birth outcomes following early-onset preeclampsia, whereas all ongoing pregnancies at 34 +0 weeks’ gestation (among women without early-onset preeclampsia) were included in the denominator for birth outcomes following late-onset preeclampsia.
In addition, we compared neonatal outcomes between infants born to mothers with and without early-onset or late-onset preeclampsia, adjusting for gestational age at delivery (traditional analysis). This analysis used live births at a particular gestational age as the denominator and provided a predictive (noncausal) model comparing the odds of adverse neonatal outcomes among mothers with and without preeclampsia, conditional on delivery of a live-born infant at a specific gestational age. We further compared birth outcomes between mothers with early-onset preeclampsia who delivered at 34 weeks or longer, and mothers with late-onset preeclampsia (who, by definition, all delivered at ≥34 weeks).
We performed sensitivity analyses examining the effect of obesity on the risk of early-onset and late-onset preeclampsia and its association with birth outcomes. Obesity was defined as a body mass index (BMI) greater than 30 kg/m 2 . Missing values for BMI (27.7%, 14.6%, and 13.9% in the early-onset, late-onset, and no preeclampsia groups, respectively) were imputed using multiple imputation procedures (proc MI, SAS software, version 9.2; SAS Institute Inc, Cary, NC). In addition, we adjusted for time period (year of delivery) to address the potential effects of changes in obstetric and neonatal practices.
All analyses were performed on publicly accessible de-identified data. An exemption from ethics approval was granted by the Department of Social and Health Services, State of Washington. Analyses were carried out using SAS software, version 9.2 (SAS Institute Inc., Cary, NC). A 2-tailed P < .05 was considered significant.
Results
The study included 456,668 women who delivered a singleton live birth or stillbirth between 2003 and 2008. The rate of preeclampsia was 3.11 per 100 singleton deliveries (14,201 of 456,668), and the rate of eclampsia was 4.12 per 10,000 singleton deliveries (188 of 456,668). The frequency of early-onset preeclampsia was 0.38 per 100 deliveries, and the frequency of late-onset preeclampsia was 2.72 per 100 deliveries ( Table 1 ). The gestational age–specific incidence of preeclampsia increased with pregnancy duration, from 0.01 per 1000 ongoing pregnancies at 20 weeks’ gestation to 9.62 per 1000 ongoing pregnancies at 40 weeks’ gestation ( Figure ).
| Maternal characteristics/clinical factors | Ongoing pregnancies at 20 weeks | Early-onset preeclampsia | Ongoing pregnancies at 34 weeks a | Late-onset preeclampsia | ||
|---|---|---|---|---|---|---|
| (n = 456,668) | (n = 1752) | Rate per 1000 | (n = 447,822) | (n = 12,449) | Rate per 1000 | |
| Age, y | ||||||
| <20 | 39,584 | 160 | 4.0 | 38,561 | 1627 | 42.2 |
| 20-34 | 347,105 | 1256 | 3.6 | 340,850 | 9001 | 26.4 |
| ≥35 | 69,979 | 336 | 4.8 | 68,411 | 1821 | 26.6 |
| Race | ||||||
| Non-Hispanic white | 323,552 | 1154 | 3.6 | 317,775 | 8914 | 28.1 |
| African-American | 20,045 | 158 | 7.9 | 19,356 | 690 | 35.6 |
| Hispanic | 56,615 | 210 | 3.7 | 55,533 | 1555 | 28.0 |
| Native-American | 10,625 | 47 | 4.4 | 10,313 | 372 | 36.1 |
| Other | 44,032 | 164 | 3.7 | 43,192 | 851 | 19.7 |
| Education | ||||||
| Less than high school | 89,238 | 324 | 3.6 | 87,255 | 2465 | 28.3 |
| High school or more | 359,160 | 1357 | 3.8 | 352,795 | 9752 | 27.6 |
| Smoking during pregnancy | ||||||
| Yes | 46,936 | 162 | 3.5 | 45,737 | 1131 | 24.7 |
| No | 403,471 | 1527 | 3.8 | 396,224 | 11,185 | 28.2 |
| Marital status | ||||||
| Unmarried | 149,369 | 657 | 4.4 | 145,677 | 4740 | 32.5 |
| Married | 305,666 | 1076 | 3.5 | 300,669 | 7654 | 25.5 |
| Number of prior live births | ||||||
| 0 | 186,980 | 964 | 5.2 | 182,906 | 8131 | 44.5 |
| ≥1 | 258,135 | 673 | 2.6 | 254,003 | 4043 | 15.9 |
| Diabetes mellitus | ||||||
| Yes | 25,815 | 207 | 8.0 | 25,110 | 1353 | 53.9 |
| No | 430,853 | 1545 | 3.6 | 422,712 | 11,096 | 26.2 |
| Chronic hypertension | ||||||
| Yes | 5560 | 237 | 42.6 | 5163 | 699 | 135.4 |
| No | 451,108 | 1515 | 3.4 | 442,659 | 11,750 | 26.5 |
| Infertility treatment | ||||||
| Yes | 3455 | 26 | 7.5 | 3339 | 133 | 39.8 |
| No | 453,213 | 1726 | 3.8 | 444,483 | 12,316 | 27.7 |
| Infant sex | ||||||
| Male | 234,224 | 914 | 3.9 | 229,314 | 6647 | 29.0 |
| Female | 222,441 | 838 | 3.8 | 218,508 | 5802 | 26.6 |
| Congenital anomalies | ||||||
| Yes | 2249 | 23 | 10.2 | 1949 | 66 | 33.9 |
| No | 454,419 | 1729 | 3.8 | 445,873 | 12,383 | 27.8 |
Women who were at the extremes of maternal age (younger than 20 or 35 years old or older), African-American, unmarried, and nulliparous had higher rates of early-onset preeclampsia ( Table 1 ). Similarly, women who had diabetes mellitus or chronic hypertension, used infertility treatment to conceive, and had an infant with a congenital anomaly also had higher rates of early-onset preeclampsia. Women who were very young (younger than 20 years of age), unmarried, nulliparous, had diabetes mellitus or chronic hypertension, used infertility treatment to conceive, and were pregnant with a male fetus also had higher rates of late-onset preeclampsia. On the other hand women who smoked or belonged to the “other” race category (ie, other than non-Hispanic white, Hispanic, African-American and Native-American; Table 2 ) had lower rates of late-onset disease.
| Demographic/clinical factors | Preeclampsia, unadjusted analysis | Preeclampsia, adjusted analysis | ||||||
|---|---|---|---|---|---|---|---|---|
| Early onset | Late onset | Early onset | Late onset | |||||
| HR | 95% CI | HR | 95% CI | AHR | 95% CI | AHR | 95% CI | |
| Age, y | ||||||||
| <20 a | 1.12 | 0.95–1.32 | 1.56 | 1.48–1.65 | 0.84 | 0.70–1.00 | 1.07 | 1.01–1.14 |
| 20-34 | Ref | Ref | Ref | Ref | ||||
| ≥35 a | 1.33 | 1.18–1.50 | 1.04 | 0.99–1.10 | 1.15 | 1.10–1.21 | 1.15 | 1.10–1.21 |
| Race | ||||||||
| Non-Hispanic white | Ref | Ref | Ref | Ref | ||||
| African-American a | 2.21 | 1.88–2.61 | 1.26 | 1.16–1.36 | 1.75 | 1.45–2.12 | 1.20 | 1.11–1.31 |
| Hispanic | 1.01 | 0.96–1.06 | 1.01 | 0.96–1.06 | 1.07 | 1.01–1.13 | 1.07 | 1.01–1.13 |
| Native-American | 1.27 | 1.54–1.59 | 1.27 | 1.54–1.59 | 1.36 | 1.22–1.51 | 1.36 | 1.22–1.51 |
| Other a | 1.03 | 0.88–1.22 | 0.73 | 0.68–0.79 | 0.98 | 0.82–1.16 | 0.68 | 0.63–0.73 |
| Maternal education less than high school | 0.98 | 0.94–1.03 | 0.98 | 0.94–1.03 | 0.98 | 0.93–1.03 | 0.98 | 0.93–1.03 |
| Smoking during pregnancy | 0.91 | 0.86–0.96 | 0.91 | 0.86–0.96 | 0.87 | 0.82–0.93 | 0.87 | 0.82–0.93 |
| Unmarried | 1.27 | 1.23–1.32 | 1.27 | 1.23–1.32 | 1.14 | 1.10–1.19 | 1.14 | 1.10–1.19 |
| No prior live births a | 1.98 | 1.80–2.19 | 2.59 | 2.49–2.69 | 2.13 | 1.92–2.37 | 2.67 | 2.57–2.78 |
| Diabetes mellitus a | 2.45 | 2.32–2.58 | 2.45 | 2.32–2.58 | 1.87 | 1.60–2.18 | 2.46 | 2.32–2.61 |
| Chronic hypertension a | 13.06 | 11.39–14.97 | 6.72 | 6.22–7.25 | 11.72 | 10.11–13.59 | 5.83 | 5.39–6.32 |
| Infertility treatment | 1.60 | 1.37–1.87 | 1.60 | 1.37–1.87 | 1.17 | 0.99–1.37 | 1.17 | 0.99–1.37 |
| Infant sex (male) | 1.10 | 1.07–1.14 | 1.10 | 1.07–1.14 | 1.10 | 1.06–1.14 | 1.10 | 1.06–1.14 |
| Congenital anomalies a | 2.91 | 1.93–4.39 | 1.50 | 1.18–1.92 | 2.59 | 1.66–4.02 | 1.49 | 1.16–1.91 |
a The hazard ratios differed significantly between early- and late-onset preeclampsia. The Wald statistic was used to assess statistical significance of the interaction term between the risk factor and the gestational age at preeclampsia onset.
Several risk factors were associated with preeclampsia, without a significant difference in adjusted hazard ratios for early- and late-onset disease ( Table 2 ). These included smoking during pregnancy (AHR, 0.87; 95% CI, 0.82–0.93 for both early- and late-onset disease), unmarried status (AHR, 1.14; 95% CI, 1.10–1.19), older maternal age (AHR, 1.15; 95% CI, 1.10–1.21), and infant’s sex (AHR for male sex, 1.10; 95% CI, 1.06–1.14).
In contrast, several risk factors differed significantly in their association with early- vs late-onset preeclampsia. African-American race, chronic hypertension, and congenital anomalies were more strongly associated with early-onset disease, whereas young maternal age (younger than 20 vs 20-35 years), other race (not including African-American, Hispanic or Native-American vs non-Hispanic white), nulliparity, and diabetes mellitus were more strongly associated with late-onset disease. Other race had a protective effect on late-onset disease compared with non-Hispanic white race (AHR, 0.68; 95% CI, 0.63–0.73).
Women with chronic hypertension had the highest risk for preeclampsia, with more than a 10-fold higher rate of early-onset disease (AHR, 11.7; 95% CI, 10.1–13.6) and an approximately 5-fold higher rate of late-onset disease (AHR, 5.8; 95% CI, 5.4–6.3) as compared with women without chronic hypertension.
The rates of all adverse birth outcomes, except for large for gestational age (LGA), were significantly higher among women with early-onset preeclampsia compared with women without early-onset disease ( Table 3 ). Among women with early-onset preeclampsia, approximately 12%, delivered at 34 weeks’ gestation or later, and almost one half of births (49.5%) were very low birthweight (<1500 g). With the exception of neonatal death rates, the rates of all adverse birth outcomes were significantly higher among mothers with late-onset preeclampsia compared with those without preeclampsia ( Table 3 ).
| Birth outcomes | Early-onset preeclampsia | No early-onset preeclampsia | Late-onset preeclampsia | No late-onset preeclampsia | ||||
|---|---|---|---|---|---|---|---|---|
| n | Rate per 1000 FAR | n | Rate per 1000 FAR | n | Rate per 1000 FAR | n | Rate per 1000 FAR | |
| Ongoing pregnancies | 1752 | 454,916 | 12,449 | 435,373 | ||||
| Gestational age at delivery, wks | ||||||||
| 20-33 | 1539 | 878.4 | 7094 | 15.6 | n/a | n/a | ||
| 34-36 | 128 | 73.1 | 26,062 | 57.3 | 2911 | 233.8 | 23,151 | 53.2 |
| 37-43 | 85 | 48.5 | 421,760 | 927.1 | 9538 | 766.2 | 412,222 | 946.8 |
| Birthweight, g | ||||||||
| <1500 | 867 | 494.9 | 3208 | 7.1 | 42 | 3.4 | 149 | 0.3 |
| 1500-2499 | 641 | 365.9 | 18,020 | 39.6 | 2020 | 162.3 | 12,972 | 29.8 |
| 2500-4499 | 183 | 104.5 | 424,395 | 932.9 | 10,162 | 816.3 | 413,747 | 950.3 |
| ≥4500 | 0 | 0.0 | 7733 | 17.0 | 166 | 13.3 | 7567 | 17.4 |
| SGA (<10th percentile) | 563 | 321.3 | 29,439 | 64.7 | 2007 | 161.2 | 26,800 | 61.6 |
| LGA (>90th percentile) | 23 | 13.1 | 52,702 | 115.8 | 1171 | 94.1 | 51,372 | 118.0 |
| Apgar score at 5 min ≤3 | 78 | 44.5 | 2064 | 4.5 | 99 | 8.0 | 1328 | 3.1 |
| Neonatal seizures | 5 | 2.9 | 186 | 0.4 | 10 | 0.8 | 159 | 0.4 |
| Neonatal sepsis | 200 | 114.2 | 3468 | 7.6 | 170 | 13.7 | 2409 | 5.5 |
| NICU admission | 1202 | 686.1 | 22,434 | 49.3 | 1658 | 133.2 | 16,843 | 38.7 |
| Fetal death | 58 | 33.1 | 1648 | 3.6 | 28 | 2.2 | 631 | 1.4 |
| Neonatal death | 50 | 28.5 | 1071 | 2.4 | 15 | 1.2 | 402 | 0.9 |
| Severe neonatal morbidity a | 367 | 209.5 | 5748 | 12.6 | 274 | 22.0 | 3758 | 8.6 |
| Neonatal death/severe morbidity | 381 | 217.5 | 6198 | 13.6 | 283 | 22.7 | 4020 | 9.2 |
| Perinatal death | 108 | 61.6 | 2719 | 6.0 | 43 | 3.5 | 1033 | 2.4 |
| Perinatal death/severe morbidity | 439 | 250.6 | 7846 | 17.2 | 311 | 25.0 | 4651 | 10.7 |
a Includes any of the following: a 5 minute Apgar score of 3 or less, neonatal seizures, neonatal sepsis, bronchopulmonary dysplasia, necrotizing enterocolitis, intraventricular hemorrhage grades 3 and 4, periventricular leukomalacia, and retinopathy of prematurity.
The rate of adverse birth outcomes remained severalfold higher among mothers with early-onset preeclampsia as compared with mothers without early-onset disease after adjustment for risk factors ( Table 4 ). For example, the rate of fetal death was approximately 6 times higher (AOR, 5.8; 95% CI, 4.0–8.3), and the rate of perinatal death or serious neonatal morbidity was 16 times higher (AOR, 16.4; 95% CI, 14.5–18.6) among women with early-onset disease.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
