Hypertensive disorders in pregnancy are common complications that affect 5% to 10% of all pregnancies in the United States.1 The contribution of hypertensive disorders in pregnancy to maternal mortality is well documented.1–3 In recent years, some developed countries have started to implement surveillance of severe obstetric complications as a valuable adjunct to maternal death audit.4–7 Data from these large national surveillance studies have confirmed that severe obstetric morbidity occurs at rates at least 10 times higher than maternal mortality. However, how hypertensive disorders contribute to the indicators of severe obstetric morbidity has not been quantified at the population level.
The classification scheme of hypertensive disorders in pregnancy proposed by the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy includes eclampsia/preeclampsia, preeclampsia superimposed on chronic hypertension, gestational hypertension, and chronic hypertension.8 The American College of Obstetricians and Gynecologists also proposed to classify preeclampsia into mild and severe forms based on the degree of hypertension, proteinuria, and multisystem involvment.9 Severe preeclampsia is defined as a presence of one or more of the following criteria: severe hypertension (sustained systolic blood pressure higher than 160 mm Hg or diastolic blood pressure higher than 110 mm Hg), severe proteinuria (more than 5 g/24 h), severe oliguria (less than 500 mL/24 h), cerebral or visual disturbances, epigastric or right upper-quadrant pain, pulmonary edema or cyanosis, impaired liver function, thrombocytopenia (less than 100,000/mm3), and fetal growth restriction (less than the 10th percentile).9 Our current understanding of the association of hypertensive disorders in pregnancy with severe obstetric complications relies on a few population-based studies that focused on eclampsia/severe preeclampsia10 and chronic hypertension.11 The relationships of mild preeclampsia and gestational hypertension, the most prevalent hypertensive disorders in pregnancy, with severe obstetric complications remain to be clarified.
The purpose of this study was to examine 1) national trends in rates of hypertensive disorders in pregnancy, 2) associations among major types of hypertensive disorders in pregnancy and severe obstetric complications, and 3) the contribution of hypertensive disorders in pregnancy to the rate of severe obstetric complications during delivery hospitalization using data from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project.
MATERIALS AND METHODS
We used 1998–2006 data from the Nationwide Inpatient Sample of the Healthcare Cost and Utilization Project, a federal-state-industry partnership sponsored by the Agency for Healthcare Research and Quality.12 The Nationwide Inpatient Sample is the largest all-payer inpatient care database in the United States and uses discharges from the sampled hospitals to produce nationwide estimates. During annual data collection by the Healthcare Cost and Utilization Project, all of the nonfederal community hospitals from the participating states are stratified by rural/urban location, number of beds, region of the country, teaching status, and ownership. Within each stratum, a systematic random 20% sample of hospitals is drawn. The sample size of this large, cross-sectional study was predetermined by sampling methodology, a full description of which can be found on the Healthcare Cost and Utilization Project Web site.12 Because the Nationwide Inpatient Sample excludes data elements that could directly or indirectly identify individuals, this research was approved as exempt from the institutional review board of the Centers for Disease Control and Prevention. The Nationwide Inpatient Sample includes International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes for primary and secondary diagnoses, procedures, total charges, length of stay, hospital characteristics, and limited patient demographics. However, the Nationwide Inpatient Sample does not have detailed clinical information, such as gestational age, on maternal discharge records.
Our analysis included all 1998–2006 delivery hospitalizations. To identify delivery hospitalizations, we used a combination of delivery-related Diagnosis-Related Group and ICD-9-CM codes13 as documented in detail elsewhere.14 Hospitalization records were excluded if the discharge summary contained ICD-9-CM codes for nondelivery pregnancy outcomes. It is possible that ICD-9-CM codes at discharge might be used to indicate an historical condition or a “ruled out” condition. Such hospitalizations might be indicated by implausibly short lengths of stay given the recorded diagnosis. As such, after identifying hospitalizations with severe obstetric complications using condition-specific ICD-9-CM codes, we used mortality, transfer from or to another health care facility, and length of stay (defined as hospitalizations with lengths of stay more than 90th percentiles calculated separately for vaginal and primary and repeat cesarean deliveries) as criteria to distinguish between hospitalizations with and without severe complications. The detailed description of the methodology with the specific ICD-9-CM codes for severe obstetric complications can be found elsewhere.6 In-hospital mortality was identified using the variable “died during hospitalization” in the Healthcare Cost and Utilization Project data system.
Because some hospital records had codes for multiple types of hypertensive disorders, hypertensive disorders in pregnancy were classified in hierarchical order into four mutually exclusive groups using ICD-9-CM codes: eclampsia/severe preeclampsia (642.5x and 642.6x), mild preeclampsia (642.4x), chronic hypertension (642.0x, 642.1x, 642.2x, 642.7x, 642.9x, and 401.x-405.xx), and gestational hypertension. The ICD-9-CM code 642.3x, “transient hypertension in pregnancy,” was used to denote the clinical entity of gestational hypertension.13 The ICD-9-CM code 642.7, for preeclampsia or eclampsia superimposed on preexisting hypertension, does not allow distinguishing between severe and mild preeclampsia. Because only a small number of records (0.3%) had this code, we could not analyze this group separately. Hence, we chose to classify these records as chronic hypertension. In addition to maternal age and mode of delivery, we also investigated multiple births, diabetes, presence of at least one chronic condition, and hospital characteristics as potential predictors of obstetric complications. To identify preexisting conditions, we modified the Healthcare Cost and Utilization Project Comorbidity Software (3.2)15 to include ICD-9-CM pregnancy-specific codes for preexisting conditions: 646.2x (unspecified renal disease in pregnancy), 646.4x (peripheral neuritis in pregnancy), 648.1x (thyroid dysfunction), 648.4x (mental disorders), 648.5x (congenital cardiovascular disorders), 648.6x (other cardiovascular diseases), and 648.8x (abnormal glucose tolerance). The complete list of preexisting comorbidities and detailed descriptions of hospital-characteristic variables that were predefined in the Healthcare Cost and Utilization Project data system are available on the Healthcare Cost and Utilization Project Web site.15 The list of hospital characteristics included U.S. region (Northwest, Midwest, South, West), location type (rural/urban), teaching status (yes/no), and bed size (small, medium, large). Because some states do not report race/ethnicity information to the Healthcare Cost and Utilization Project, we did not include this variable in the analysis.12
The unit of analysis was a hospitalization, not an individual. We adjusted rates of hypertensive disorders to the 1998 distribution of delivery hospitalizations by mother’s age in our study sample and reported age-adjusted rates of hypertensive disorders per 1,000 delivery hospitalizations by year. Overall temporal change in rates of each type of hypertensive disorder was tested by fitting 1-year intervals as a continuous variable in the logistic regression model. In addition, the rates of hypertensive disorders in pregnancy by selected patient characteristics were assessed. Finally, we calculated the rates of severe obstetric complications by type of hypertensive disorder in the 1998–2006 aggregate sample and compared these rates with rates of severe complications among hospitalizations without hypertensive disorders. All rate comparisons were performed using χ2 tests with adjustment of P-values for multiple comparisons using the Bonferroni method.
Odds ratios and their 95% confidence intervals (CIs), obtained from logistic regression, are good estimates of relative risks (RR) when outcomes are rare events.16 Thus, we used odds ratios as estimates of RRs to examine the associations of each outcome with hypertensive disorders in pregnancy. We calculated population-attributable fraction accounting for adjustment of RR and confidence limits using methods described elsewhere.17,18 We used SAS software (SAS Institute, Inc., Cary, NC) to manage data and SAS-callable SUDAAN software (RTI International, Research Triangle, NC) to account for the multistage probability sampling design. Thus, all results are based on the weighted estimates of delivery hospitalizations in the United States during the study period.
Using the enhanced method, we identified 36,598,421 discharges that occurred during the period from 1998–2006. After excluding 43,208 (0.14 %) with nondelivery pregnancy outcomes, such as hydatidiform mole, other abnormal product of conception, ectopic pregnancy, and abortion, and 18,152 (0.05 %) records with missing information on age, we were left with 36,537,061 delivery discharges available for analysis.
The overall prevalence of hypertensive disorders among delivery hospitalizations increased from 67.2 per 1,000 deliveries in 1998 to 83.4 per 1,000 deliveries in 2006. The linear increasing trends were significant for all types of hypertensive disorders (P<.001) except for mild preeclampsia (Fig. 1). The highest increase (from 11.0 to 16.9 per 1,000 deliveries or by about 50%) was observed for chronic hypertension. The prevalence of eclampsia/severe preeclampsia increased moderately (from 9.4 to 12.4 per 1,000 deliveries or by 30%). The overall prevalence of hypertensive disorders by mode of delivery ranged from 42.2 per 1,000 deliveries for vaginal after cesarean to 154.1 per 1,000 deliveries for primary cesarean delivery (Table 1). Regardless of the type of hypertensive disorders in pregnancy, higher rates were observed among delivery hospitalizations with multiple births (178.4 per 1,000 deliveries), gestational diabetes (159.1 per 1,000 deliveries), and chronic conditions (116.2 per 1,000 deliveries) compared with hospitalizations without these conditions (about 70.0 per 1,000 deliveries). Although hospitalizations covered by private payer had a higher overall prevalence of hypertensive disorders compared with public payer (75.8 compared with 72.5 per 1,000 deliveries), hospitalizations covered by public payer had a significantly higher prevalence of eclampsia/severe preeclampsia and mild preeclampsia compared with private payer.
The rate of severe obstetric complications among delivery hospitalizations without hypertensive disorders ranged from 1.0 per 10,000 deliveries for renal failure to 9.7 per 10,000 deliveries for heart failure (Table 2). Rates for heart failure, pulmonary embolism (1.3 per 10,000 deliveries), and severe complications of anesthesia (1.1 per 10,000 deliveries) were calculated, but they are not shown in the table. For nearly every severe complication category, the complication rate was higher for each category of hypertensive disorder compared with delivery hospitalizations without hypertension. The highest rate of complications was observed among hospitalizations with eclampsia/severe preeclampsia, ranging from 12.7 to 184.5 per 10,000 deliveries for pulmonary embolism and disseminated intravascular coagulation syndrome, respectively. Although the rates of severe obstetric complications were the lowest among hospitalizations with gestational hypertension compared with hospitalizations with other hypertensive disorders, the rates of all complications, with the exception of pulmonary embolism and in-hospital mortality, were significantly higher compared with hospitalizations without hypertensive disorders.
After adjusting for confounding factors, the risk of severe complications was increased for almost every complication category for hospitalizations with hypertensive disorders compared with hospitalizations without hypertension (Table 3). The significant RRs ranged from 1.5 (95% CI 1.1–2.1) for pulmonary edema among hospitalizations with gestational hypertension to 34.8 (95% CI 30.4–39.9) for acute renal failure among hospitalizations with eclampsia/severe preeclampsia. Notably, compared with hospitalizations without any hypertensive disorders, hospitalizations with severe preeclampsia/eclampsia were 16.6, 14.8, and 11.0, times more likely to have ICD-9-CM codes for puerperal cerebrovascular disorders, disseminated intravascular coagulation syndrome, and ventilation, respectively. The RRs for pulmonary embolism were 3.3 (95% CI 2.6–4.2), 1.7 (95% CI 1.3–2.2), and 2.0 (95% CI 1.5–2.7) for eclampsia/severe preeclampsia, mild preeclampsia, and chronic hypertension, respectively (data not shown). Gestational hypertension was moderately associated with 6 out of 10 severe complications that were examined in the study. The relationship of gestational hypertension with pulmonary embolism, adult respiratory distress syndrome, ventilation, and in-hospital mortality was statistically nonsignificant.
The estimated population-attributable fractions for each type of hypertensive disorder and all hypertensive disorders combined and their 95% CIs are shown in Table 4. As expected, the largest attributable fractions were observed for eclampsia/severe preeclampsia and the smallest for gestational hypertension. Hospitalizations with eclampsia/severe preeclampsia were associated with 38% of hospitalizations with acute renal failure and 19% or more of hospitalizations with ventilation, disseminated intravascular coagulation syndrome, pulmonary edema, puerperal cerebrovascular disorders, and respiratory distress syndrome. Overall population-attributable fractions of all hypertensive disorders combined were about 57% for acute renal failure, 27% for disseminated intravascular coagulation syndrome, and above 30% for pulmonary edema, puerperal cerebrovascular disorders, adult respiratory syndrome, and ventilation. They were 11% for pulmonary embolism, 18% for heart failure, and 17% for severe complications of anesthesia (data not shown).
The results of our large, nationwide study show that the prevalence of hypertensive disorders has been increasing among delivery hospitalizations, especially since 2001, with the largest proportional increase observed for chronic hypertension and the largest absolute increase observed for gestational hypertension. In our study, all types of hypertensive disorders during pregnancy were associated with severe obstetric complications, with eclampsia/severe preeclampsia and chronic hypertension being major contributors to a burden of severe obstetric morbidity during delivery hospitalizations.
The interpretation of our results on prevalence trends of hypertensive disorders should take into account the acceptance of stricter diagnostic criteria for hypertensive disorders in pregnancy beginning in 2000.8 However, our findings, confirmed by the previous reports from Los Angeles County19 and the National Hospital Discharge Survey,20 show increasing trends in prevalence of overall hypertensive disorders and in each subtype of hypertensive disorder in pregnancy except for mild preeclampsia. The abrupt increase in the prevalence of gestational hypertension after 1999 suggests the possibility of a clinical reclassification of hospitalizations with mild preeclampsia to hospitalizations with gestational hypertension owing to the elimination of edema as a criterion for diagnosing preeclampsia.8 The extent to which the definition put forth by the National High Blood Pressure Education Program influences the diagnosis and coding of gestational hypertension is unclear.8 Gestational hypertension, as proposed by the National High Blood Pressure Education Program, does not have a specific label in ICD-9-CM.13 The International Codes for Diagnosis continue to reflect the outdated classification “transient hypertension” and include the term “gestational hypertension” only in the description of ICD-9-CM code 642.3. Unfortunately, gestational hypertension may be transient or it may represent early (before proteinuria) preeclampsia or chronic hypertension. For example, use of an ICD-9-CM code to differentiate between gestational and chronic hypertension during delivery hospitalization may be problematic because an accurate diagnosis depends on normalization of blood pressure as late as 12 weeks postpartum.8
Comparable with the results from the study based on the 1988–1997 national hospital discharge data,10 eclampsia/severe preeclampsia was a strong predictor of severe obstetric complications in our study. Our results are also in a line with the results from a study that examined the relationship between chronic hypertension and severe complications using California hospital discharge data.11 However, the information on mild preeclampsia and gestational hypertension is scarce. The modest increased risks for severe morbidity conditions for gestational hypertension in our study are reassuring given the relatively high prevalence of this condition.
We estimated the population-attributable fraction of hypertensive disorders in pregnancy for selected severe obstetric complications to better understand how hypertension contributes to the overall burden of severe pregnancy complications. In our study, delivery hospitalizations with hypertensive disorders contributed to a relatively large proportion of hospitalizations with severe obstetric complications. Assuming that exposure is causal and removable, population-attributable fractions often are used to estimate the effectiveness of interventions in reducing the prevalence of disease.18 However, in our cross-sectional study, causality cannot be established, and it is unrealistic to assume that all hypertensive disorders in pregnancy can be prevented. Thus, our findings can be interpreted only as a proportion of hospitalizations with particular severe complications associated with overall or a particular type of hypertensive disorder in pregnancy.
Similar to maternal mortality, severe complications associated with hypertensive disorders in pregnancy are probably among the most difficult to predict and prevent.3 Despite extensive research during the past decade, reliable biomarkers for prediction and detection of preeclampsia have not been identified.21 Based on the results of several randomized trials and systematic reviews published during the past 2 years, no specific medication or supplement can be recommended for prevention of preeclampsia.21 Finally, the only known definitive cure for preeclampsia/eclampsia is early delivery, which may put the fetus at risk.21 The decision between delivery and continued pregnancy remains a challenge for clinicians.
Our study has several limitations that should be taken into account when interpreting the results. First, our identification of severe complications is based solely on ICD-9-CM codes and data-driven criteria such as mortality, transfer from or to another health care facility, and length of stay. Second, ICD-9-CM codes for hypertensive disorders in pregnancy vary greatly in their accuracy; several studies have reported high to moderate specificity but low sensitivity, especially when only pregnancy-specific ICD-9-CM codes (642.xx) are used. Thus, our results may reflect underreporting of the prevalence of hypertensive disorders in pregnancy and misclassification of exposure or outcomes or both. However, the effects of misclassification on measures of association may be small for rare outcomes.22 Third, although marked racial disparities in prevalences of maternal morbidity, mortality, and hypertensive disorders in pregnancy have been reported in previous studies,10,23 the unavailability of information on race precluded us from adjusting for this confounder. Similarly, we had no information about body mass index. Finally, because our data did not allow us to account for multiple delivery hospitalizations of the same women during the study period, the estimated standard error may be underestimated and, thus, may lead to unwarranted statistical significance in hypothesis testing. However, because the majority of our results were highly significant (P<.001) and the year-to-year sampling of hospitals is random, taking into account the clustering effect is unlikely to change our conclusions. Despite these limitations, the present study is the only recent, large, population-based study that reported trends in hypertensive disorders in pregnancy, including chronic hypertension, and examined associations between four major types of hypertensive disorders and severe obstetric complications.
The number of delivery hospitalizations in the United States with hypertensive disorders in pregnancy is increasing, and these hospitalizations are more likely to have severe complications compared with hospitalizations without hypertensive disorders. A substantial fraction of severe morbidity in the obstetric population could be avoided if these conditions, particularly severe preeclampsia, could be prevented. The increasing proportion of births to women at the older end of the reproductive spectrum24 and the increase in obesity among pregnant women may explain some fraction of this trend.25 However, future studies are needed to examine factors contributing to these increasing trends. More importantly, we need a better biological understanding of the etiologies and natural histories of the hypertensive disorders of pregnancy if we are to make progress in preventing them and their attendant morbidity and mortality.
1. Wagner SJ, Barac S, Garovic VD. Hypertensive pregnancy disorders: current concepts. J Clin Hypertens (Greenwich) 2007;9:560–6.
2. Chang J, Elam-Evans LD, Berg CJ, Herndon J, Flowers L, Seed KA, et al. Pregnancy-related mortality surveillance-United States, 1991–1999. MMWR Surveill Summ 2003 Feb 21;52:1–8.
3. Fiander A. Maternal mortality associated with hypertensive disorders of pregnancy in Africa, Asia, Latin America and the Caribbean. Br J Obstet Gynecol 1992;99:937–8.
4. Callaghan WM, Mackay AP, Berg CJ. Identification of severe maternal morbidity during delivery hospitalizations, United States, 1991–2003. Am J Obstet Gynecol 2008;199:133.e1–8.
5. Knight M, UKOSS. Peripartum hysterectomy in the UK: management and outcomes of the associated haemorrhage. BJOG 2007;114:1380–7.
6. Kuklina EV, Meikle SF, Jamieson DJ, Whiteman MK, Barfield WD, Hillis SD, et al. Severe obstetric morbidity in the United States: 1998–2005. Obstet Gynecol 2009 Feb;113:293–9.
7. Wen SW, Huang L, Liston R, Heaman M, Baskett T, Rusen ID, et al. Severe maternal morbidity in Canada, 1991–2001. CMAJ 2005;173:759–64.
8. Report of the National High Blood Pressure Education Program Working Group on High Blood Pressure in Pregnancy. Am J Obstet Gynecol 2000;183:S1–S22.
9. Diagnosis and management of preeclampsia and eclampsia. ACOG Practice Bulletin No. 33. American College of Obstetricians and Gynecologists. Obstet Gynecol 2002;99:159–67.
10. Zhang J, Meikle S, Trumble A. Severe maternal morbidity associated with hypertensive disorders in pregnancy in the United States. Hypertens Pregnancy 2003;22:203–12.
11. Gilbert WM, Young AL, Danielsen B. Pregnancy outcomes in women with chronic hypertension: a population-based study. J Reprod Med 2007;52:1046–51.
12. Healthcare Cost and Utilization Project (HCUP). Overview of the Nationwide Inpatient Sample (NIS). Available at: www.hcup-us.ahrq.gov/databases.jsp
. Retrieved December 23, 2008.
13. The National Center for Health Statistics (NCHS), the Centers for Medicare and Medicaid Services (CMMS). International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). Available at: http://www.cdc.gov/nchs/about/otheract/icd9/abticd9.htm
. Retrieved December 23, 2008.
14. Kuklina EV, Whiteman MK, Hillis SD, Jamieson DJ, Meikle SF, Posner SF, et al. An enhanced method for identifying obstetric deliveries: implications for estimating maternal morbidity. Matern Child Health J 2008;12:469–77.
16. Kelsey J, Whittemore A, Evans A, Thompson W. Methods in observational epidemiology. New York (NY): Oxford University Press; 1996.
17. Natarajan S, Lipsitz SR, Rimm E. A simple method of determining confidence intervals for population attributable risk from complex surveys. Stat Med 2007;26:3229–39.
18. Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions [published erratum appears in Am J Public Health 2008;98:2119]. Am J Public Health 1998;88:15–9.
19. Baraban E, McCoy L, Simon P. Increasing prevalence of gestational diabetes and pregnancy-related hypertension in Los Angeles County, California, 1991–2003. Prev Chronic Dis 2008;5:A77.
20. Wallis AB, Saftlas AF, Hsia J, Atrash HK. Secular trends in the rates of preeclampsia, eclampsia, and gestational hypertension, United States, 1987–2004. Am J Hypertens 2008;21:521–6.
21. Sibai BM. Hypertensive disorders of pregnancy: the United States perspective. Curr Opin Obstet Gynecol 2008;20:102–6.
22. Callaghan WM. Invited commentary: identifying women with hypertension during pregnancy–is high specificity sufficient? Am J Epidemiol 2007;166:125–7.
23. Tucker MJ, Berg CJ, Callaghan WM, Hsia J. The Black-White disparity in pregnancy-related mortality from 5 conditions: differences in prevalence and case-fatality rates. Am J Public Health 2007;97:247–51.
24. Hamilton BE, Miniño AM, Martin JA, Kochanek KD, Strobino DM, Guyer B. Annual summary of vital statistics: 2005. Pediatrics 2007;119:345–60.
25. Kim SY, Dietz PM, England L, Morrow B, Callaghan WM. Trends in pre-pregnancy obesity in nine states, 1993–2003. Obesity (Silver Spring) 2007;15:986–93.
© 2009 by The American College of Obstetricians and Gynecologists.
This article has been cited