Preeclampsia complicates approximately 5% of all pregnancies.1 The associated hypertension and proteinuria usually resolve after delivery. However, recent studies show that preeclampsia is associated with an increased risk of remote cardiovascular disease2 and end-stage renal disease.3 More knowledge on postpartum resolution of hypertension and proteinuria will help to properly identify chronic hypertension and proteinuria. This can help to justify further invasive diagnostic tests for underlying disease.
The time required for resolution, however, is not clear. International guidelines state that hypertension should resolve within 3 months after a preeclamptic pregnancy.4,5 Otherwise, chronic hypertension is present. Previous publications indeed show rapid postpartum decrease of blood pressure.6–9 However, they were not designed to estimate resolution after 3 months postpartum, having a follow-up period of 7 days or excluding patients with persistent hypertension at 3 months postpartum. If data of resolution of hypertension beyond 3 months postpartum were available, women with chronic hypertension could be better identified. For proteinuria, no statement exists regarding when postpartum persistence should be considered chronic. Previous publications suggest a high risk of renal injury if proteinuria is persistent after 6 or 12 weeks postpartum. However, publications are case reports10 or consist of retrospectively collected data on renal biopsies.11,12 If prospectively collected data of postpartum resolution of proteinuria were available, women with persistent proteinuria could also be better identified. In our own experience, hypertension and proteinuria could resolve beyond 3 months after a preeclamptic pregnancy.
In this historic prospective cohort study, we address the following questions: 1) How long can it take for hypertension and proteinuria to resolve after a preeclamptic pregnancy, and 2) how do prolonging pregnancy after the development of preeclampsia and the levels of blood pressure and proteinuria during preeclampsia correlate with postpartum time to resolution and to what extent?
MATERIALS AND METHODS
This historic prospective cohort study was performed in preeclamptic women who were admitted to the Department of Obstetrics and Gynecology of the Erasmus MC, Rotterdam, The Netherlands, and delivered between January 1, 1990, and December 31, 1992. Preeclampsia was defined as the occurrence after 20 weeks of gestation of a blood pressure of 140/90 mm Hg or more and proteinuria of 0.3 g/d or more. The hemolysis, elevated liver enzymes, low platelets syndrome was defined as the simultaneous occurrence of a platelet count of less than 100×109/L and serum aspartate aminotransferase and serum alanine aminotransferase concentrations greater than 30 units/L (2 standard deviations above the mean in our hospital). Women with a record of chronic hypertension or kidney disease before pregnancy or before 20 weeks of gestation were excluded. Severity of the disease was retrospectively defined according to the American College of Obstetricians and Gynecologists criteria.5
During the study period, a consultant of internal and obstetric medicine (author W.V.) used a protocol for the follow-up. All women who were admitted with preeclampsia were offered visits at 1.5, 3, 6, 12, 18, and 24 months after delivery. If a woman did not present for the checkup, she was sent a reminder once. However, not every woman attended every visit. The hospital files were searched to collect data on blood pressure, use of antihypertensive drugs, urinary protein excretion, and serum creatinine levels at every attended visit.
Hypertension was defined as the use of antihypertensive drugs or a blood pressure 140/90 mm Hg or greater. Proteinuria was defined as 0.3 g/d or greater urinary protein loss. For both conditions separately, the first visit of measured resolution was identified. Because resolution obviously occurred somewhere between visits, the actual time of resolution was analyzed as interval censored. If no visit of resolution could be identified and follow-up data were incomplete, women were considered lost to follow-up. The Turnbull extension of the Kaplan-Meier procedure to interval-censored data13 was then used to calculate resolution curves with 95% confidence bands (SAS 9.1.3; SAS Institute Inc., Cary, NC). The women who were lost to follow-up were right censored in the analysis under the assumption of independent censoring from the time of resolution. This means that after the last known visit, the proportion of right-censored women who showed resolution was equal to the proportion of women with complete follow-up who showed resolution at a specific time.
Hospital files were searched to collect clinical features of preeclampsia. They included maximal systolic and diastolic blood pressure during admission, maximal level of proteinuria during admission, gestational age at onset of preeclampsia, and the time interval between diagnosis and delivery. To analyze the correlation between these covariates and the rate of resolution, we used an accelerated failure time model with flexible error distribution allowing for interval censoring as described elsewhere14 (smoothSurv 0.3-12 R 2.7.1, Komárek). This model is not restricted to assume a specific distribution for the error like gaussian or extreme value distribution, but allows a smooth error distribution resulting from the penalized normal mixture.
This study was approved by the Institutional Research Board of the Erasmus MC Medical Center (Medisch Ethische Toestingscommissie EMC, trial number MEC-2009-169).
A total of 242 women were admitted with preeclampsia during the study period. Of these women, 37 were excluded: 30 women had been diagnosed with hypertension before pregnancy, 2 had been diagnosed with proteinuria before pregnancy, and 5 had been diagnosed with both. A total of 205 women were included for analysis. Complete blood pressure data were available for 116 women, and complete proteinuria data were available for 121 women. General characteristics are summarized in Table 1.
For 116 women, complete blood pressure data were available: For 112 women, resolution of hypertension was identified; 4 women still had (known) hypertension 2 years postpartum. The remaining 89 women were lost to follow-up before normalization of blood pressure. Table 2 shows the comparison between these two groups, showing the latter group having had worse clinical features of preeclampsia.
Figure 1 shows the resolution curve for hypertension with 95% confidence interval (CI). At discharge after delivery, 78% still had hypertension. At 6 weeks postpartum in 54% and at 3 months postpartum in 39% of the women, hypertension was still present. Finally, at 2 years postpartum, 18% had persistent hypertension. No difference was observed in normalization between primiparous and multiparous women. In the accelerated failure time model, resolution time increased by 60% (95% CI 32-95%, P<.001) for every 10-mm Hg increase in maximal systolic blood pressure. Resolution time increased by 40% (95% CI 0.5-96%, P=.044) for every 10-mm Hg increase in maximal diastolic blood pressure. Resolution time increased by 3.6% (95% CI 1.4-5.8%, P=.001) for every addition of 1 day to the diagnosis-to-delivery interval. Figures 2–4 show the effect of these variables on the resolution curve. The baseline in each figure is defined as maximal systolic blood pressure of 140 mm Hg, maximal diastolic blood pressure of 90 mm Hg, and time interval between diagnosis and delivery of 0 days. The curves show the effect on the rate of resolution of hypertension when the specific variable changes, leaving the other variables at baseline. The coefficients for the corresponding accelerated failure time model are presented in Table 3. Resolution time of hypertension was not significantly correlated with maximum level of proteinuria or gestational age at onset of preeclampsia.
For 121 women, complete proteinuria data were available: For all those women, resolution of proteinuria was identified. The remaining 84 women were lost to follow-up before resolution of proteinuria. Table 4 shows the comparison between these two groups, showing the latter group having had higher levels of maximal proteinuria during preeclampsia.
Figure 5 shows the resolution curve for proteinuria with 95% CI. At discharge after delivery, 65% of women still had proteinuria. At 6 weeks postpartum in 21% and at 3 months postpartum in 14% of women, proteinuria was present. Finally, at 2 years postpartum, 2% had persistent proteinuria. No difference was observed in normalization between primiparous and multiparous women. In the accelerated failure time model, resolution time increased by 16% (95% CI 12-21%, P=.001) for every 1-g/d increase in maximal proteinuria. Figure 6 shows the effect on the rate of resolution of proteinuria when this variable changes. The coefficients for the corresponding accelerated failure time model are presented in Table 3. Resolution time of proteinuria was not significantly correlated with maximal systolic and diastolic blood pressure during preeclampsia, gestational age at onset of preeclampsia, or time interval between diagnosis and delivery.
Our results show that hypertension and proteinuria can resolve during the first 2 years after a preeclamptic pregnancy. Of all women who had hypertension at 3 months postpartum, 50% resolved within 2 years postpartum. Therefore, these women could not have had true chronic hypertension and would have been misclassified by the current definition. Furthermore, a surprising 14% still had proteinuria at 3 months postpartum. Of these women, 85% resolved within 2 years postpartum. Clinical features that were associated with a longer time to resolution were higher maximal blood pressure and higher maximal level of proteinuria during pregnancy and a longer time interval between diagnosis and delivery.
Previous longitudinal studies of resolution of hypertension are limited in follow-up time and mainly focus on the differences of resolution between the types of hypertensive disorder in pregnancy.6–9 The longest previous longitudinal study on resolution of hypertension had a follow-up time of 50 days after delivery.6 However, women with persistent hypertension at 50 days postpartum were excluded for analysis, thereby precluding any estimate of later resolution. Suzuki et al9 described a postpartum observational study of 52 women with gestational hypertension. Of the 52 women, nine (17%) still had hypertension 2 years postpartum. However, the rate of resolution between delivery and 2 years postpartum was not analyzed.
We found an incidence of proteinuria of 14% at 3 months postpartum. This is higher than in previous studies. In 1992, Chua and Redman15 described a longitudinal study on the resolution of proteinuria after preeclampsia. They showed persistent proteinuria in 2% at 3 months postpartum. However, proteinuria was defined as 0.5 g or more of protein a day. That might explain the difference with our findings.
Other studies have looked at renal lesions after preeclampsia.11,12 Two types of renal lesions, characteristic of preeclampsia, can be discriminated: those that disappear within days after delivery (glomerular deposits of various hemostatic factors) and those that resolve gradually until as long as 18 months after delivery (endothelial swelling and subendothelial enlargement). The latter supports our conclusion that proteinuria at 3 months postpartum does not imply chronic disease, but might still be a transient effect of endothelial damage during preeclampsia. Moreover, this finding suggests that a renal biopsy for prolonged proteinuria within 2 years postpartum may be postponed, because the majority of lesions will disappear.
Sometimes, resolution of proteinuria might be explained by a decrease in glomerular filtration rate or a decrease in blood pressure. Unfortunately, serum creatinine levels were not routinely measured during follow-up, because all but five women had normal serum creatinine levels within 3 months postpartum. Therefore, we think it is unlikely that a decrease in glomerular filtration rate would have contributed to resolution of proteinuria. At every postpartum visit, blood pressure levels were comparable between women with and without proteinuria. Thus, it is unlikely that the level of blood pressure has interfered with the resolution of blood pressure.
This study of resolution of hypertension and proteinuria could well reflect endothelial recovery after preeclampsia, as previously suggested.6 Endothelial dysfunction plays a central role in the pathogenesis of preeclampsia and might account for most of the pathologic changes.16,17 Our findings suggest that the level of endothelial cell injury during preeclampsia, reflected by highest blood pressure and level of proteinuria, is correlated with the time to postpartum recovery. Whether prolonged resolution is reassuring remains unclear. It might very well reflect worse subclinical cardiovascular health. This should be an aim for further research. If true, it can help to identify women at risk of future cardiovascular disease.
Moreover, our findings also suggest that the duration of exposure to endothelial cell injury, as reflected by diagnosis-to-delivery interval, is correlated with the time to postpartum resolution of hypertension. This association suggests that preeclampsia itself affects remote cardiovascular health. This implies that temporizing treatment may increase remote cardiovascular risk. Whether this is true should be part of further study.
The main limitation of this study is the number of women lost to follow-up. With regard to the Turnbull analysis, they are considered right censored. This means that after the last known visit, they were expected to behave like the women with complete follow-up. Because the women lost to follow-up had more severe clinical features of preeclampsia, our method probably resulted in a best-case scenario. However, within the accelerated failure time model, the clinical features of preeclampsia have been adjusted for. These clinical features could be related to the probability of censoring; therefore, given the clinical features, the censoring mechanism could have been independent from the time of resolution. In this case, because the resolution curves obtained from the accelerated failure time models are conditional on the clinical features, they would provide a valid estimate of survival probability for the group with given values of clinical features.14
The second limitation of this study is that the follow-up period is 2 years. Hereby we preclude any estimate of later resolution. However, we believe it is unlikely that resolution would occur beyond 2 years postpartum because we found an incidence of hypertension of 18% at 2 years postpartum. This is comparable with other studies that showed similar proportions of women with hypertension until 7 years after preeclampsia.9,18–20
The third limitation of this study is that because our hospital is a referral center, a large proportion of women have had severe preeclampsia (89%). This may limit the ability to generalize our findings to women with mild preeclampsia.
The period of recruitment of this study was 1990-1992 for two reasons. First, the data of this cohort had already been processed to minimize missing data. However, proper analysis had never been conducted. Recent development of the accelerated failure time model14 made it possible to analyze the data set as presented and quantify the effect of blood pressure, proteinuria, and temporizing treatment of preeclampsia on postpartum resolution. Second, temporizing treatment was more common during the study period, compared with current practice. This allowed a better analysis of effect of temporizing treatment on postpartum recovery than if we had used more recent data.
Because women with a preeclamptic pregnancy are at risk of cardiovascular disease and end-stage renal disease, proper follow-up after delivery is necessary. However, this is not always the case: In 2004, Samwiil et al21 reported that of 257 former preeclamptic women, at 6 weeks postpartum 6% had not had their blood pressure measured and 68% had not had their urine tested for proteinuria. Hypertension, if present, must be treated. Persistent proteinuria at 3 months postpartum must make the physician aware of a possible underlying renal disease and necessitates evaluation by a nephrologist. However, if the level of proteinuria stays below 3 g/d, further invasive diagnostic tests for underlying renal disease may be postponed until hypertension or proteinuria persists after 2 years postpartum. Further studies must focus on the correlation between time to resolution of hypertension and proteinuria and remote cardiovascular risk.
1. Wallenburg HCS, Visser W. Pregnancy-induced hypertensive disorders. Curr Opin Obstet Gynecol 1994;6:19–29.
2. Bellamy L, Casas JP, Hingorani AD, Williams DJ. Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ 2007;335:974–85.
3. Vikse BE, Irgens LM, Leivestad T, Skjaerven R, Iversen BM. Preeclampsia and the risk of end-stage renal disease. N Engl J Med 2008;359:800–9.
4. Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol 2000;183:S1–22.
5. Diagnosis and management of preeclampsia and eclampsia. ACOG Practice Bulletin No. 33. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet 2002;77:67–75.
6. Ferrazzani S, De Carolis S, Pomini F, Testa AC, Mastromarino C, Caruso A. The duration of hypertension in the puerperium of preeclamptic women: relationship with renal impairment and week of delivery. Am J Obstet Gynecol 1994;171:506–12.
7. Makkonen N, Harju M, Kirkinen P. Postpartum recovery after severe pre-eclampsia and HELLP-syndrome. J Perinat Med 1996;24:641–9.
8. Stepan H, Nordmeyer AK, Faber R. Proteinuria in hypertensive pregnancy diseases is associated with a longer persistence of hypertension postpartum. J Hum Hypertens 2006;20:125–8.
9. Suzuki H, Watanabe Y, Arima H, Kobayashi K, Ohno Y, Kanno Y. Short- and long-term prognosis of blood pressure and kidney disease in women with a past history of preeclampsia. Clin Exp Nephrol 2008;12:102–9.
10. Durham JH, Desnick RJ, Imbriano L, Wasserstein M, D’Agati VD, Markowitz GS. Prolonged postpartum proteinuria after early preeclampsia. Am J Kidney Dis 2004;43:186–91.
11. Kincaid-Smith P. The renal lesion of preeclampsia revisited. Am J Kidney Dis 1991;17:144–8.
12. Suzuki S, Gejyo F, Ogino S, Maruyama Y, Ueno M, Nishi S, et al. Postpartum renal lesions in women with pre-eclampsia. Nephrol Dial Transplant 1997;12:2488–93.
13. Turnbull BW. The empirical distribution function with arbitrarily grouped, censored and truncated data. J R Stat Soc Ser B 1976;38:290–5.
14. Komárek A, Lesaffre E, Hilton JF. Accelerated failure time model for arbitrarily censored data with smoothed error distribution. J Comput Graph Stat 2005;14:726–45.
15. Chua S, Redman CW. Prognosis for pre-eclampsia complicated by 5 g or more of proteinuria in 24 hours. Eur J Obstet Gynecol Reprod Biol 1992;43:9–12.
16. Lain KY, Roberts JM. Contemporary concepts of the pathogenesis and management of preeclampsia. JAMA 2002;287:3183–6.
17. Ness RB, Sibai BM. Shared and disparate components of the pathophysiologies of fetal growth restriction and preeclampsia. Am J Obstet Gynecol 2006;195:40–9.
18. Edlow AG, Srinivas SK, Elovitz MA. Investigating the risk of hypertension shortly after pregnancies complicated by preeclampsia. Am J Obstet Gynecol 2009;200:e60–2.
19. Manten GT, Sikkema MJ, Voorbij HA, Visser GH, Bruinse HW, Franx A. Risk factors for cardiovascular disease in women with a history of pregnancy complicated by preeclampsia or intrauterine growth restriction. Hypertens Pregnancy 2007;26:39–50.
20. Berends AL, de Groot CJ, Sijbrands EJ, Sie MP, Benneheij SH, Pal R, et al. Shared constitutional risks for maternal vascular-related pregnancy complications and future cardiovascular disease. Hypertension 2008;51:1034–41.
© 2009 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
21. Samwiil L, Mercer C, Jarrett P, O’Malley S; Genetics of Pre-Eclampsia Collaborative Study (GOPEC) Research Midwives. Blood pressure and urinalysis are often omitted in women who have suffered pre-eclampsia at their six-week postnatal check. BJOG 2004;111:623–5.