- Email: [email protected]
Sickle Cell Disease and Pregnancy Outcomes
Sickle Cell Disease and Pregnancy Outcomes Women of African Descent Wanda D. Barfield, MD, MPH, Danielle T. Barradas, PhD, Susan E. Manning, MD, MPH, Milton Kotelchuck, PhD, MPH, Carrie K. Shapiro-Mendoza, PhD, MPH Background: Sickle cell disease (SCD) is a severe hematologic condition that presents unique complications among affected pregnant women. Many studies of adverse perinatal outcomes associated with SCD are limited by small samples or fail to consider important risk factors.
Purpose: This study compared perinatal outcomes among women of African ancestry with and without SCD in a large, population-based sample. Methods: Data from the Massachusetts Pregnancy to Early Life Longitudinal (PELL) Data System were analyzed during June–August 2009 to identify in-state deliveries to resident women of African descent. Logistic regression analyses compared perinatal outcomes for deliveries among women with and without SCD, adjusted for maternal age, education, parity, plurality, insurance status, adequacy of prenatal care, smoking during pregnancy, and infant gender.
Results: During 1998 –2006, there were 116,076 deliveries to 84,561 women; SCD prevalence was 0.6%. Adjusted odds of fetal death among deliveries to women with SCD were 2.2 times those among women without SCD (95% CI⫽1.2, 4.2). Compared to women without SCD, the odds of preterm delivery, low birth weight, and having babies small for gestational age (SGA) among women with SCD were 1.5 (95% CI⫽1.2, 1.8); 1.7 (95% CI⫽1.1, 2.6); and 1.3 (95% CI⫽1.0, 1.7), respectively. Sickle cell disease was positively associated with cesarean delivery and inductions. Conclusions: Population-based linked data systems are useful for assessing risks of adverse health outcomes among women with specifıc medical conditions, such as SCD. Women with SCD should seek preconception care to identify and modify risk behaviors and receive counseling regarding potential adverse sequelae associated with pregnancy-related morbidity and preterm delivery. (Am J Prev Med 2010;38(4S):S542–S549) Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine
Background
S
ickle cell disease (SCD), an inherited hematologic disorder, affects approximately 50,000 African Americans.1 The disease causes anemia and acute vaso-occlusion affecting the major organ systems, including the long bones, heart, lungs, kidneys, intestinal tract, and brain. Individuals affected with SCD may require
From the Division of Reproductive Health (Barfıeld, Barradas, Manning, Shapiro-Mendoza), National Center for Chronic Disease Prevention and Health Promotion, CDC, Atlanta, Georgia; Massachusetts Department of Health (Manning); and Boston University School of Public Health (Kotelchuck), Boston, Massachusetts Address correspondence and reprint requests to: Wanda D. Barfıeld, MD, MPH, Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, CDC, 4770 Buford Highway NE, MS-K22, Atlanta GA 30341. E-mail: wbarfı[email protected]. 0749-3797/00/$17.00 doi: 10.1016/j.amepre.2009.12.020
S542 Am J Prev Med 2010;38(4S):S542–S549
prolonged and recurrent hospitalizations for disorders such as acute pain crises, infections, cardiac problems, renal failure, and acute chest syndrome.2,3 Pregnant women with SCD experience these medical risks, as well as vascular effects to the gravid uterus and placenta, which create additional risk for both women and their fetuses.4 Yet data on pregnancy outcomes for U.S. women with SCD and their offspring are limited.5 An important contribution to our understanding of pregnancy outcomes among women with SCD involves a nationwide sample of inpatient delivery hospitalizations during 2000 –2003.6 This study indicates that women with SCD are at signifıcantly increased risk of mortality and morbidity, including cerebral and deep venous thrombosis, postpartum infection, sepsis, and pneumonia. However, this cross-sectional study of individual birth hospitalizations does not account for correlations
Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
among individual women who may have more than one delivery over time. This and other hospital discharge studies compare deliveries of pregnant women with SCD to all other deliveries with limited information about maternal race or ancestry;7 however, women of African descent have an increased risk of adverse pregnancy outcomes compared to European women.8 –10 Most studies of pregnancy-associated complications of SCD among African-American, Caribbean, or African women are not population-based, but limited to individual hospitals or multiple referral centers.11–13 This study describes maternal characteristics and delivery outcomes among a population of women of African descent who delivered in Massachusetts with SCDrelated hospitalizations compared to women without SCD, identifıed by using longitudinally linked antenatal, delivery, and postpartum hospitalizations (emergency room, observational stay, delivery, and inpatient hospitalizations). The aim of examining the contribution of SCD to maternal and perinatal outcomes among women of African ancestry is to understand the unique reproductive health burden of SCD on maternal and infant health.
Methods Data Source Data for this analysis were derived from the Massachusetts Pregnancy to Early Life Longitudinal (PELL) Data System, a partnership between the Massachusetts Department of Public Health (MDPH); Boston University (BU) School of Public Health; and the Centers for Disease Control and Prevention (CDC). PELL provides a unique, population-based reproductive data system, with multiple linked data sets that can be used for cross-sectional and longitudinal analyses. The PELL core consists of the annual linkage of all Massachusetts’ birth certifıcates, fetal death reports, and deliveryrelated hospital discharge records for mothers and infants. This core is longitudinally linked to statewide pre- and postdelivery hospital utilization records (inpatient visits, observational stays, and emergency room records) for both mother and child. PELL contains data from 1998 to 2006 on 736,987 live births, 3,896 fetal deaths, and 505,703 unique women who delivered those infants, including 174,221 women with repeat deliveries. Over 99% of birth certifıcates have been linked to their corresponding maternal or infant hospital discharge record. Records are stripped of identifying information prior to analysis. Additional information on the PELL linkage methodology has been published elsewhere.14
Study Population This retrospective cohort study included Massachusetts resident women of African descent with in-state deliveries (live April 2010
S543
birth or fetal death) during 1998 –2006. Massachusetts birth certifıcate and fetal death data contain questions on the self-identifıcation of race and ancestry; African descent was defıned by self-identifıcation on either record as being black, Haitian, Jamaican, Barbadian, other West Indian/Caribbean, African-American, Nigerian, Cape Verdean, or of other African ancestry. Puerto Rican and Dominican women were also included in the study based on historical African presence in these countries (Table 1). All live births and fetal deaths at ⱖ20 weeks and ⱖ350 g were included in the study. Fetal deaths at ⬍20 weeks and ⬍350 g (n⫽21, 0.02%) were excluded because these deaths are not required to be reported to the state. Using information available on birth certifıcates, fetal death, and delivery and nondelivery hospital utilization data up to 1 year before and after a delivery date, women were classifıed into three categories: SCD, no documented hemoglobinopathy, or other non-SCD hemoglobinopathy. Women with SCD were those identifıed on the birth certificate or fetal death report with a positive response on the maternal condition check box as having sickle cell anemia or who had any related hospital discharge codes (ICD-9) for sickle cell (282.60 –282.69); acute chest (517.3); or splenic sequestration (289.52). Women with sickle--thalassemia (282.41, 282.42) were included with the SCD group for this analysis because they have similar clinical manifestations during pregnancy.15 Women with other non-SCD hemoglobinopathies (thalassemia [282.49] and other nonspecifıc hemoglobinopathies [282.7]) were excluded from the current SCD analysis. Further details are shown in Table 2.
Covariates and Dependent Variables Covariates included several sociodemographic characteristics available from the birth certifıcate: maternal age at delivery (categorized into age groups: ⬍20, 20 –24, 25–29, 30 –34, 35–39, and ⬎40 years); completed years of education (⬍12, 12, 13–15, and ⱖ16 years); parity (1st pregnancy, 2nd pregnancy, 3rd pregnancy, and 4th or higher pregnancy); plurality (singleton versus multiple gestation); adequacy of prenatal care (Adequacy of Prenatal Care Utilization Index [APNCU]: adequate plus, adequate, intermediate, inadequate, or missing)17; smoking during pregnancy (yes, no); and health insurance (public, private, other/unknown) derived from the method of payment on the infant’s hospital discharge record following birth. Chronic and pregnancy-related maternal medical risk factors were characterized using information available on birth certifıcate and fetal death record check boxes in addition to antenatal, delivery, and postpartum hospital utilization data using ICD-9 codes (Table 3). Diabetes prior to pregnancy and gestational diabetes were classifıed together to reduce potential misclassifıcation.18 A similar convention was followed for the coding of chronic and pregnancyinduced hypertension. Eclampsia was considered separately from other hypertensive disorders, including preeclampsia,
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
S544
Table 1. Maternal delivery characteristics by hemoglobinopathy status: PELL, 1998 –2006 (n [%] unless otherwise indicated) No hemoglobinopathies (nⴝ115,160)
Sickle cell disease (nⴝ663)
p-value*
⬍0.0001
Maternal ancestry American
67,096 (58.3)
322 (48.6)
African
16,317 (14.2)
96 (14.5)
Caribbean
14,752 (12.8)
160 (24.1)
South or Central American
15,881 (13.8)
80 (12.1)
Asian, Indian, other
a
970 (0.8)
4 (0.6)
Maternal age (years) ⬍20
18,629 (16.2)
97 (14.6)
20–24
32,065 (27.8)
175 (26.4)
25–29
28,348 (24.6)
178 (26.9)
30–34
21,744 (18.9)
125 (18.9)
35–39
14,405 (9.9)
ⱖ40
0.49
74 (11.2)
2,967 (2.6)
14 (2.1)
⬍12
33,057 (28.7)
173 (26.1)
12
40,977 (35.6)
222 (34.5)
13–15
27,003 (23.5)
194 (29.3)
ⱖ16
14,123 (12.3)
74 (11.2)
Education (years) 0.006
Data Analysis
Parity 1
46,601 (40.5)
276 (41.8)
2
35,061 (30.5)
196 (29.7)
3
19,501 (17.0)
107 (16.2)
ⱖ4
13,821 (12.0)
82 (12.4)
113,186 (98.3)
649 (97.9)
0.87
Plurality Singleton Twin or higher
1974 (1.7)
14 (2.1)
Private
49,896 (43.3)
246 (37.1)
Public
61,531 (53.4)
388 (58.5)
3,733 (3.2)
29 (4.4)
0.43
Insurance during pregnancy
Other/unknown
0.003
Prenatal care utilization index 4,983 (4.3)
33 (5.0)
Inadequate
Missing
16,610 (14.4)
100 (15.1)
Intermediate
10,069 (8.7)
Adequate
45,286 (39.3)
211 (31.8)
Adequate plus
38,212 (33.2)
276 (41.6)
105,234 (91.5)
589 (89.4)
⬍0.0001
43 (6.5)
Smoking during pregnancy None Any a
9800 (8.5)
because of the severity and specifıc nature of this condition. Lung disease was characterized as disorders such as chronic airway obstruction, acute and chronic respiratory failure, and nonspecifıc lung disease; asthma, bronchitis, emphysema, and allergic reactions were not included in this risk group. Ischemic heart disease, atherosclerotic conditions, myocardial degeneration, and heart failure were included under heart disease. Chronic and end-stage renal disease and unspecifıed renal diseases in pregnancy were grouped together under kidney disease. Outcome measures included maternal, fetal, and infant death. Infant death included both neonatal (⬍28 days) and postneonatal (28 –364 days) events. Preterm delivery (⬍37 weeks gestation); low birth weight (⬍2500 g); small for gestational age (SGA, birth weight ⬍10th percentile for gestational age and gender); and intrauterine growth restriction (IUGR, birth weight ⬍3rd percentile for gestational age and gender) were also measured.19 Vaginal or cesarean delivery (ICD-9 codes: 740, 742, 744, 749.9) and induction of labor (ICD-9 codes: 730.1, 731, 734, 964.9) were identifıed from both vital records and delivery hospitalization data. Infant gender, gestational age (completed weeks), and birth weight (grams) at delivery considered in these analyses were obtained from birth certifıcate or fetal death data.
0.05
70 (10.6)
Includes Chinese, Vietnamese, Cambodian, Korean, Filipino, Japanese, Laotian, Thai, other Asian or Pacific Islander, Asian Indian, Pakistani, Lebanese, Iranian, Israeli, other Middle Eastern, European, other, and unknown *p-value⬍0.05 is considered significant PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System
Bivariate analyses were conducted to compare selected demographic and clinical characteristics between deliveries to women with SCD and no documented hemoglobinopathy. Overlapping 95% CIs were also used to examine differences in the occurrence of maternal medical complications. For analyses examining fetal/infant outcomes, deliveries resulting in fetal deaths were analyzed separately from deliveries resulting in live births. Logistic regression analysis for correlated data was used to examine differences in fetal death between deliveries to women with SCD and no documented hemoglobinopathy.20 Among deliveries resulting in live births, bivariate and multivariable logistic regression analyses for correlated data were used to examine differences in the method of delivery (vaginal or cesarean delivery) and in the occurrence of adverse infant outcomes to include infant death (neonatal and postneonatal) preterm delivery, low birth weight, SGA, and IUGR. Multivariable logistic regression analyses adjusted for maternal age, education, parity, plurality, adequacy of prenatal care, insurance status, smoking during pregnancy, and infant gender. These variables were chosen as covariates because of their documented associations with maternal, perinatal, and delivery outcomes in previous studies. Maternal clinical factors were not included as covariates, as their independent relationship from the SCD diagnosis itself cannot be established. Data were analyzed with SAS software. Signifıcance was tested at ␣⫽0.05 for all analyses. This study was approved by the MDPH Research and Data Access Review (RaDAR)/IRB Committee, the Division www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
S545
Table 2. Classification of sickle cell disease and other hemoglobinopathies, ICD-9
Ninety percent of all delivering women Sickle cell disease Non–sickle cell disease hemoglobinopathies with SCD during this a 282.41 (thalassemia-Hb-S without crisis) 282.49 (other thalassemias) time period were of African descent. 282.42 (thalassemia-Hb-S with crisis)a 282.7 (other hemoglobinopathies) Among SCD women b 282.6 (sickle cell anemia, NOS) of African descent, 282.61 (Hb-S without crisis)b 49% were African282.62 (Hb-S with crisis)b Americans; 24% were b Caribbean; the others 282.63 (sickle cell/Hb-C disease) were predominately b 282.64 (sickle cell/Hb-C disease with crisis) from Africa or South/ 282.68 (other sickle cell without crisis)b Central America. Ma282.69 (sickle cell anemia, NOS) ternal characteristics, such as maternal age a Added to the ICD-9 in October 1, 200316 b and most education Modified October 1, 2003 categories were simiNOS, not otherwise specified lar among women with and without of Healthcare Finance and Policy Data Protection Committee, SCD (Table 1). Parity and plurality did not differ signifıthe BU School of Public Health, and the CDC Human Subjects cantly among the groups. Over 58% of SCD deliveries Research Committees in June 2009, and analysis was conwere covered by public insurance, yet nearly 26% of deducted June–August 2009. liveries to women with SCD had less-than-adequate prenatal care utilization, despite higher percentages of adeResults quate-plus utilization compared to women with no reported hemoglobinopathies (41.6% vs 33.2%). The Among 708,467 Massachusetts resident deliveries during prevalence of smoking during pregnancy among women 1998 –2006, 116,097 were to women of African descent with SCD was higher than among those with non-SCD (Figure 1). There were 116,076 live births and reportable hemoglobinopathies but barely signifıcant (10.6% vs fetal deaths delivered by 84,561 unique women of African 8.5%, p⫽0.05). descent during this period. Among the study population, Based on non-overlapping 95% CIs, preeclampsia, 488 women had SCD, and 83,877 were identifıed with no lung disease, and heart disease were also reported more reported hemoglobinopathy.
Table 3. Prevalence of maternal medical conditions among deliveries: PELL, 1998 –2006
Maternal medical risk conditions
ICD-9 codes
No hemoglobinopathies n (% [95% CI])
Sickle cell disease n (% [95% CI])
Diabetes
250 (250.00–250.33, 250.40–250.93), 648.0, 648.8
5948 (5.2 [5.0, 5.3])
48 (7.2 [5.3, 9.2])
Hypertension
401.0, 401.1, 401.9, 402.00–405.99, 403.0, 403.1, 403.9, 437.2, 642.00–642.24, 642.30–642.34, 642.90–642.94, 997.91
9052 (7.9 [7.7, 8.0])
66 (10.0 [7.7, 12.2])
Eclampsia
642.60–642.64
2104 (1.8 [1.7, 1.9])
15 (2.3 [1.1, 3.34])
Preeclampsia
642.40–642.44, 642.50–642.54, 642.70–642.74
5137 (4.5 [4.3, 4.6])
50 (7.5 [5.5, 9.6])
Lung disease
496, 518.89, 518.8, 518.81–518.84
7864 (6.8 [6.7, 7.0])
73 (11.0 [8.6, 13.4])
Heart disease
429–429.9, 428–428.4, 428.9, 414–414.3, 414.8
1035 (0.9 [0.8, 1.0])
18 (2.7 [1.5, 4.0])
Kidney disease
646.2, 593.9, 585, 585.6, 585.9, 646.2–646.24
1193 (1.0 [0.9, 1.1])
8 (1.2 [0.4, 2.0])
PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System
April 2010
S546
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
Discussion
Figure 1. Study population selection and case classification: PELL, 1998 –2006 PELL, Pregnancy to Early Life Longitudinal Data System; SCD, sickle cell disease
often among deliveries in women with SCD compared to deliveries in women with no reported hemoglobinopathies (Table 3). Reporting of diabetes trended toward being higher among deliveries to women with SCD but was not statistically different. The two groups showed no differences for hypertension, eclampsia, and kidney disease. Crude and adjusted ORs of the association between maternal SCD and perinatal outcomes are shown in Table 4. There were fewer than seven maternal and infant deaths among deliveries to women with SCD (exact number not shown because of confıdentiality requirements for small cell size suppression); therefore, ORs were not calculated for these outcomes. Fetal demise was twice as likely among SCD deliveries relative to deliveries to women with no reported hemoglobinopathies. Less-advantageous infant outcomes were associated with SCD. The odds of preterm delivery were 50% higher, and the odds of low birth weight were 70% higher among SCD deliveries than among non-SCD deliveries in multivariate models. SGA and IUGR in infants were 30% higher among deliveries to women with SCD compared to women with no hemoglobinopathy after adjustment for covariates; but the OR for IUGR did not reach significance. Cesarean delivery was signifıcantly associated (AOR⫽1.3; 95% CI⫽1.1, 1.5) and induction of labor was marginally associated (AOR⫽1.2, 95% CI⫽1.0, 1.5) with SCD compared to no hemoglobinopathies, in adjusted models.
Among a population of Massachusetts women of African ancestry, SCD remains a major risk factor for pregnancyand delivery-related complications and is associated with adverse infant outcomes, despite adjustment for demographic, behavioral characteristics, access to insurance, and prenatal care utilization. Massachusetts is a state with a well-regionalized system of perinatal care and advanced medical technology, where women delivering with SCD have mostly similar demographic characteristics to other women of African descent. However, women with SCD still have higher risks of fetal death, preterm births, low– birth weight births, cesarean delivery, and induction of labor. Using a population-based linked data system with race/ancestry, sociodemographic and clinical data adds to the current body of knowledge regarding the association between SCD and adverse pregnancy and delivery outcomes. These data also enable the analysis of women affected with SCD, rather than solely the analysis of SCD hospitalizations as seen in most current cross-sectional studies.6,21 Overall, intensive prenatal care utilization was highest among women with SCD compared to women without SCD. Further analysis of the APNCU Index (data not shown) suggests that women with SCD do not enter care earlier than other women of African descent, but the services they receive once in care are considerably more intensive, which may indicate a heightened clinical sensitivity to their potential pregnancy-related complications. Nearly 22% of women in both groups had lessthan-adequate prenatal care, including almost 15% with inadequate or no care. Among women with SCD, in particular, prenatal care plays an important role in the management of anemia and amelioration of crisis episodes through iron and folate supplementation, aggressive treatment of infections, and monitoring of fetal growth and well-being.4,15,22 Additionally, enhanced preconception care could potentially create opportunities for early baseline health maintenance; preventive health practices (including smoking cessation); and counseling.23,24 Medical conditions during pregnancy involving the heart and lungs were more common among women with SCD. A recent study of hospital discharges in the U.S. reported that women with non-inflammatory vasculopathies (SCD, primary pulmonary hypertension, and systemic sclerosis) have increased odds of hypertensive disorders including preeclampsia,21 and this may explain the higher percentages of preeclampsia seen in the current study. This population-based study reaffırms the increased risk of preterm delivery for women with SCD reported by other, more-restricted studies.6,8,9 Yet the mechanism for www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
Table 4. Association between maternal sickle cell disease (SCD) and perinatal outcomes: PELL, 1998 –2006
S547
number of pregnancies among women with SCD. Prior to No hemoglobinopathy SCD Unadjusted a October 2003, ICD-9 Outcomes (nⴝ115,160; n) (nⴝ663; n) OR (95% CI) AOR (95% CI) codes lacked specifıcFetal mortality 916 14 2.7 (1.6, 4.5)* 2.2 (1.2,4.2)* ity to distinguish Infant outcomesb between sickle-⫹ thalassemia (282.41, Pretermc 14,405 127 1.7 (1.4,2.0)* 1.5 (1.2,1.8)* 282,42);  thalasse⬍34 weeks 4,711 44 1.7 (1.2,2.3)* 1.5 (1.1,2.0)* mia (a milder hemo34–36 weeks 9,694 83 1.6 (1.2,2.0)* 1.4 (1.1,1.8)* globinopathy); and Low birth weightd 2,126 23 1.9 (1.3,3.0)* 1.7 (1.1,2.6)* other thalassemias (282.49).16 The curSGAe 10,733 79 1.3 (1.0,1.7)* 1.3 (1.0,1.7)* rent study also does IUGRf 4,649 34 1.3 (0.9,1.8) 1.3 (0.9,1.8) not include spontaDelivery outcomesb neous fırst trimester Cesarean 29,748 207 1.3 (1.1,1.5)* 1.3 (1.1,1.5)* or elective abortions because MassachuInduction 24,160 163 1.2 (1.0,1.5)* 1.2 (1.0,1.5)* setts does not require a Adjusted for maternal age, maternal years of education, parity, plurality, insurance status at delivery, reporting of fetal adequacy of prenatal care utilization, smoking during pregnancy, and infant gender b deaths at ⬍20 weeks Live births only c Less than 37 weeks gestation gestation and ⬍350 g d Less than 2500 g birth weight. Addie Small-for-gestational age; ⬍10th birth weight percentile for gestational age and gender tionally, the data f Intrauterine growth restriction; ⬍3rd birth weight percentile for gestational age and gender used in this study *p-value ⬍0.05 IUGR, intrauterine growth restriction; PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System; were not validated by SCD, sickle cell disease; SGA, small-for-gestational-age chart review and are subject to the quality of administrative preterm delivery among women with SCD remains unrecord data. However, hospital discharge data have been known. Prematurity may be the major cause of low birth shown to provide greater sensitivity in reporting maternal weight among infants of mothers with SCD in this study; medical conditions16 than birth and fetal death certifıcate modest increases in SGA and IUGR among women with data alone.27–30 SCD were also seen despite defıning IUGR more strictly Although the current study is strengthened by includthan in previous studies (⬍3rd percentile) and adjusting ing a population of women of African descent, it is possifor gestational age and infant gender. Vascular stasis of ble that it inadvertently excluded SCD women who did the uteroplacental unit among gravid women with SCD not report any of the race and ancestry groups selected. has been proposed as a potential mechanism for fetal However, the self-reporting nature of the birth certifıcate growth restriction,25 and this may occur in the last tridata improves selection, and the Massachusetts-specifıc mester of pregnancy.26 Cesarean delivery during the last variable of ancestry increases the sensitivity of selecting trimester, for fetal (e.g., poor intrauterine growth) or women of African ancestry rather than using simply maternal (e.g., preeclampsia) medical indications/risks black race.31 Of the 10% of SCD deliveries excluded from could potentially have decreased the incidence of IUGR this study, most mothers were self-identifıed as South or among SCD women in the current study compared to Central American (n⫽33) or Asian (n⫽11); 16 were earlier studies.6,20 missing maternal ancestry or were classifıed as other. It was found that maternal mortality appeared elevated These women were not included, because the general among women with SCD compared to women with no population of women in Massachusetts who are of South/ documented hemoglobinopathy, similar to fındings in Central American or Asian ancestry are demographically other studies.6 However, because of confıdentiality redifferent than women of African ancestry and, as a genstrictions for small cell sizes, detailed data, including the eral population, do not tend to have SCD. cause of death, are not shown. Women with SCD in which the diagnosis was not Several limitations of the study are noted. Because of reported on hospital discharge records or birth and/or the lack of systematic data on maternal genotype (i.e., SS, fetal death data may have been missed in the study; this SC, S-thalassemia), the study may underestimate the April 2010
S548
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
would also make the statistical estimates closer to the null, because of “misclassifıcation” of SCD cases into nonhemoglobinopathies. However, because SCD is a serious, life-long condition with important implications for medical, surgical, and anesthetic management,3,32 it is unlikely to be missed in hospitalizations over time, based on the linked nature of the current data. Lastly, the Massachusetts population of women in the current study may be different than women of African descent in other geographic locations, and not representative of the U.S. as a whole, because women in Massachusetts are generally older and more highly educated.33 This study illustrates how longitudinally linked data are useful for assessing factors associated with adverse health outcomes among women with specifıc medical conditions, such as SCD. Cross-sectional studies may overestimate pregnancy and delivery risk, yet underestimate individual severity of illness. The redundancy inherent in linked data systems improves the identifıcation of chronic medical conditions and complications, which may not be reported in all clinical encounters. Future SCD studies will evaluate the impact of pre- and postdelivery hospitalizations on maternal, fetal, and infant health outcomes. Despite declines in maternal and perinatal mortality rates in the U.S., sickle cell disease remains an important factor associated with pregnancy- and delivery-associated complications among deliveries to women of African descent. Although further research is needed to understand the specifıc biologic mechanisms that contribute to SCD and adverse pregnancy outcomes, there are current opportunities to improve care to women with SCD through improved preconception counseling and care; education and prevention of high-risk behaviors (e.g., smoking); early and adequate prenatal care; and appropriate identifıcation and treatment of pregnancy/labor/delivery complications. In order to disentangle the numerous contributors to racial and ethnic disparities in U.S. perinatal outcomes, it is important to understand the impact of specifıc medical conditions that are more prevalent among women of African descent, such as SCD. We thank Terry Njoroge, MPH, for her assistance in the literature review of this manuscript, and Daksha Gopal, PhD, for her help in the initial preparation of the analytic PELL database for this study. The fındings and conclusions of this report are those of the authors and do not necessarily represent the offıcial position of the CDC. This study was funded by CDC contracts 200 –200722782 and 200-2006-15969 to Boston University School of Public Health.
No fınancial disclosures were reported by the authors of this paper.
References 1. Steiner CA, Miller JL. Sickle cell disease in U.S. hospitals. Healthcare Cost and Utilization Project, Statistical Brief # 21. Agency for Healthcare Research and Quality, USDHHS, 2006. 2. Panepinto JA, Brousseau DC, Hillery CA, Scott JP. Variation in hospitalizations and hospital length of stay in children with vaso-occlusive crises in sickle cell disease. Pediatr Blood Cancer 2005;44(2):182– 6. 3. Adam S, Jonassaint J, Kruger H, et al. Surgical and obstetric outcomes in adults with sickle cell disease. Am J Med 2008;121:916 –21. 4. American College of Obstetricians and Gynecologists. Hemoglobinopathies in pregnancy. ACOG Practice Bulletin No. 78. Obstet Gynecol 2007;109:229 –37. 5. Creary M, Williamson D, Kulkarni R. Sickle cell disease: current activities, public health implications, and future directions. J Womens Health (Larchmt) 2007;16(5):575– 82. 6. Villers MS, Jamison MG, De Castro LM, James AH. Morbidity associated with sickle cell disease in pregnancy. Am J Obstet Gynecol 2008:199:125.e1–125.e5. 7. James AH, Jamison MG, Biancazio LR, Meyers ER. Venous thromobembolism during pregnancy and the postpartum period: incidence, risk factors, and mortality. Am J Obstet Gynecol 2006;194:1311–5. 8. Nabukera SK, Wingate MS, Owen J, et al. Racial disparities in perinatal outcomes and pregnancy spacing among women delaying initiation of childbearing. Matern Child Health J 2009;13:81–9. 9. Lu MC, Halfon N. Racial and ethnic disparities in birth outcomes: a life-course perspective. Matern Child Health J 2003; 7:13–30. 10. Shen JJ, Tymkow ND, MacMullen N. Disparities in maternal outcomes among four ethnic populations. Ethn Dis 2005;492–7. 11. Serjeant, GR, Look Loy L, Crowther M, Hambleton IR, Thame M. Outcome of pregnancy in homozygous sickle cell disease. Obstet Gynecol 2004;103:1278 – 85. 12. Sun PM, Wilburn W, Raynor BD, Jamieson D. Sickle cell disease in pregnancy: Twenty years of experience at Grady Memorial Hospital, Atlanta, Georgia. Am J Obstet Gynecol 2001;184:1127–30. 13. Smith JA, Espeland M, Bellevue R, Bonds D, Brown AK, Koshy M. Pregnancy in sickle cell disease: experience of the Cooperative Study of Sickle Cell Disease. Obstet Gynecol 1996;87: 199 –204. 14. Clements KM, Barfıeld W, Kotelchuck M, Lee KG, Wilber N. Birth characteristics associated with early intervention referral, evaluation for eligibility, and program eligibility in the fırst year of life. Matern Child Health J 2006;10:433– 41. 15. Rappaport VJ, Velazquez M, Williams K. Hemoglobinopathies in pregnancy. Obstet Gynecol Clin N Am 2004;31: 287–317. 16. Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project Clinical Classifıcations Software (CCS). Rockville MD. www.hcup-us.ahrq.gov/toolssoftware/ ccs/ccs.jsp. www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549 17. Kotelchuck M. The Adequacy of Prenatal Care Utilization Index: its U.S. distribution and association with low birthweight. Am J Public Health 1994; 84:1486 –9. 18. Lyndon-Rochelle MT, Holt VL, Cardenas V, et al. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certifıcates and in hospital discharge data. Am J Obstet Gynecol 2005:193:125–34. 19. Oken E, Kleinman KP, Rich-Edwards J, Gillman MW. A nearly continuous measure of birth weight for gestational age using a U.S. national reference. BMC Pediatr 2003;3:6. 20. Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364 –75. 21. Chakravarty EF, Khanna D, Chung L. Pregnancy outcomes in systemic sclerosis, primary pulmonary hypertension, and sickle cell disease. Obstet Gynecol 2008;111:927–34. 22. Thurman AR, Steed LL, Hulsey T, Soper DE. Bacteriuria in pregnant women with sickle cell trait. Am J Obstet Gynecol 2006;194:1366 –70. 23. Lottenberg R, Hassell K. An evidence-based approach to the treatment of adults with sickle cell disease. Hematology Am Soc Hematol Educ Program 2005:58 – 65. 24. CDC. Recommendations to improve preconception health and health care—U.S. MMWR 2006;55(RR06):1. 25. Rathod KB, Jaiswal KN, Shrivastava AC, Shrikhande AV. Study of placenta in sickle cell disorders. Indian J Pathol Microbiol 2007;50:698 –701.
26. Thame M, Lewis J, Troutman H, Hambleton I, Serjeant G. The mechanism of low birth weight in infants of mothers with homozygous sickle cell disease. Pediatrics 2007;120:e686 – e693. 27. Lyndon-Rochelle MT, Holt VL, Cardenas V, et al. Accuracy of reporting maternal in-hospital diagnosis and intrapartum procedures in Washington State linked birth records. Paediatr Perinat Epidemiol 2005;19:450 –71. 28. DiGiuseppe DL, Aron DC, Ranborn L, Harper DL, Rosenthal GE. Reliability of birth certifıcate data: a multi-hospital comparison to medical records information. Matern Child Health J 2002;6:169 –79. 29. Roohan PJ, Josberger RE, Acar K, Dabir P, Feder HM, Gagliano PJ. Validation of birth certifıcate data in New York State. J Community Health 2003;28:335– 46. 30. Northam S, Knapp TR. The reliability and validity of birth certifıcates. J Obstet Gynecol Neonatal Nurs 2006;28:3–12. 31. Friedman DJ, Cohen BB, Dunn VH, et al. Topics in minority health ethnic variation and maternal risk characteristics among blacks—Massachusetts, 1987 and 1988. MMWR 1991; 40:409 –11. 32. Camous J, N’da A, Etienne-Julan M, Stéphan F. Anesthetic management of pregnant women with sickle cell disease— effect on postnatal sickling complications. Can J Anesth 2008; 55:276 – 83. 33. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2007. Natl Vital Stat Rep 2007;57:1–103.
Did you know? The AJPM online archive includes issues back to 1998. Visit www.ajpm-online.net today!
April 2010
S549
Purpose: This study compared perinatal outcomes among women of African ancestry with and without SCD in a large, population-based sample. Methods: Data from the Massachusetts Pregnancy to Early Life Longitudinal (PELL) Data System were analyzed during June–August 2009 to identify in-state deliveries to resident women of African descent. Logistic regression analyses compared perinatal outcomes for deliveries among women with and without SCD, adjusted for maternal age, education, parity, plurality, insurance status, adequacy of prenatal care, smoking during pregnancy, and infant gender.
Results: During 1998 –2006, there were 116,076 deliveries to 84,561 women; SCD prevalence was 0.6%. Adjusted odds of fetal death among deliveries to women with SCD were 2.2 times those among women without SCD (95% CI⫽1.2, 4.2). Compared to women without SCD, the odds of preterm delivery, low birth weight, and having babies small for gestational age (SGA) among women with SCD were 1.5 (95% CI⫽1.2, 1.8); 1.7 (95% CI⫽1.1, 2.6); and 1.3 (95% CI⫽1.0, 1.7), respectively. Sickle cell disease was positively associated with cesarean delivery and inductions. Conclusions: Population-based linked data systems are useful for assessing risks of adverse health outcomes among women with specifıc medical conditions, such as SCD. Women with SCD should seek preconception care to identify and modify risk behaviors and receive counseling regarding potential adverse sequelae associated with pregnancy-related morbidity and preterm delivery. (Am J Prev Med 2010;38(4S):S542–S549) Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine
Background
S
ickle cell disease (SCD), an inherited hematologic disorder, affects approximately 50,000 African Americans.1 The disease causes anemia and acute vaso-occlusion affecting the major organ systems, including the long bones, heart, lungs, kidneys, intestinal tract, and brain. Individuals affected with SCD may require
From the Division of Reproductive Health (Barfıeld, Barradas, Manning, Shapiro-Mendoza), National Center for Chronic Disease Prevention and Health Promotion, CDC, Atlanta, Georgia; Massachusetts Department of Health (Manning); and Boston University School of Public Health (Kotelchuck), Boston, Massachusetts Address correspondence and reprint requests to: Wanda D. Barfıeld, MD, MPH, Division of Reproductive Health, National Center for Chronic Disease Prevention and Health Promotion, CDC, 4770 Buford Highway NE, MS-K22, Atlanta GA 30341. E-mail: wbarfı[email protected]. 0749-3797/00/$17.00 doi: 10.1016/j.amepre.2009.12.020
S542 Am J Prev Med 2010;38(4S):S542–S549
prolonged and recurrent hospitalizations for disorders such as acute pain crises, infections, cardiac problems, renal failure, and acute chest syndrome.2,3 Pregnant women with SCD experience these medical risks, as well as vascular effects to the gravid uterus and placenta, which create additional risk for both women and their fetuses.4 Yet data on pregnancy outcomes for U.S. women with SCD and their offspring are limited.5 An important contribution to our understanding of pregnancy outcomes among women with SCD involves a nationwide sample of inpatient delivery hospitalizations during 2000 –2003.6 This study indicates that women with SCD are at signifıcantly increased risk of mortality and morbidity, including cerebral and deep venous thrombosis, postpartum infection, sepsis, and pneumonia. However, this cross-sectional study of individual birth hospitalizations does not account for correlations
Published by Elsevier Inc. on behalf of American Journal of Preventive Medicine
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
among individual women who may have more than one delivery over time. This and other hospital discharge studies compare deliveries of pregnant women with SCD to all other deliveries with limited information about maternal race or ancestry;7 however, women of African descent have an increased risk of adverse pregnancy outcomes compared to European women.8 –10 Most studies of pregnancy-associated complications of SCD among African-American, Caribbean, or African women are not population-based, but limited to individual hospitals or multiple referral centers.11–13 This study describes maternal characteristics and delivery outcomes among a population of women of African descent who delivered in Massachusetts with SCDrelated hospitalizations compared to women without SCD, identifıed by using longitudinally linked antenatal, delivery, and postpartum hospitalizations (emergency room, observational stay, delivery, and inpatient hospitalizations). The aim of examining the contribution of SCD to maternal and perinatal outcomes among women of African ancestry is to understand the unique reproductive health burden of SCD on maternal and infant health.
Methods Data Source Data for this analysis were derived from the Massachusetts Pregnancy to Early Life Longitudinal (PELL) Data System, a partnership between the Massachusetts Department of Public Health (MDPH); Boston University (BU) School of Public Health; and the Centers for Disease Control and Prevention (CDC). PELL provides a unique, population-based reproductive data system, with multiple linked data sets that can be used for cross-sectional and longitudinal analyses. The PELL core consists of the annual linkage of all Massachusetts’ birth certifıcates, fetal death reports, and deliveryrelated hospital discharge records for mothers and infants. This core is longitudinally linked to statewide pre- and postdelivery hospital utilization records (inpatient visits, observational stays, and emergency room records) for both mother and child. PELL contains data from 1998 to 2006 on 736,987 live births, 3,896 fetal deaths, and 505,703 unique women who delivered those infants, including 174,221 women with repeat deliveries. Over 99% of birth certifıcates have been linked to their corresponding maternal or infant hospital discharge record. Records are stripped of identifying information prior to analysis. Additional information on the PELL linkage methodology has been published elsewhere.14
Study Population This retrospective cohort study included Massachusetts resident women of African descent with in-state deliveries (live April 2010
S543
birth or fetal death) during 1998 –2006. Massachusetts birth certifıcate and fetal death data contain questions on the self-identifıcation of race and ancestry; African descent was defıned by self-identifıcation on either record as being black, Haitian, Jamaican, Barbadian, other West Indian/Caribbean, African-American, Nigerian, Cape Verdean, or of other African ancestry. Puerto Rican and Dominican women were also included in the study based on historical African presence in these countries (Table 1). All live births and fetal deaths at ⱖ20 weeks and ⱖ350 g were included in the study. Fetal deaths at ⬍20 weeks and ⬍350 g (n⫽21, 0.02%) were excluded because these deaths are not required to be reported to the state. Using information available on birth certifıcates, fetal death, and delivery and nondelivery hospital utilization data up to 1 year before and after a delivery date, women were classifıed into three categories: SCD, no documented hemoglobinopathy, or other non-SCD hemoglobinopathy. Women with SCD were those identifıed on the birth certificate or fetal death report with a positive response on the maternal condition check box as having sickle cell anemia or who had any related hospital discharge codes (ICD-9) for sickle cell (282.60 –282.69); acute chest (517.3); or splenic sequestration (289.52). Women with sickle--thalassemia (282.41, 282.42) were included with the SCD group for this analysis because they have similar clinical manifestations during pregnancy.15 Women with other non-SCD hemoglobinopathies (thalassemia [282.49] and other nonspecifıc hemoglobinopathies [282.7]) were excluded from the current SCD analysis. Further details are shown in Table 2.
Covariates and Dependent Variables Covariates included several sociodemographic characteristics available from the birth certifıcate: maternal age at delivery (categorized into age groups: ⬍20, 20 –24, 25–29, 30 –34, 35–39, and ⬎40 years); completed years of education (⬍12, 12, 13–15, and ⱖ16 years); parity (1st pregnancy, 2nd pregnancy, 3rd pregnancy, and 4th or higher pregnancy); plurality (singleton versus multiple gestation); adequacy of prenatal care (Adequacy of Prenatal Care Utilization Index [APNCU]: adequate plus, adequate, intermediate, inadequate, or missing)17; smoking during pregnancy (yes, no); and health insurance (public, private, other/unknown) derived from the method of payment on the infant’s hospital discharge record following birth. Chronic and pregnancy-related maternal medical risk factors were characterized using information available on birth certifıcate and fetal death record check boxes in addition to antenatal, delivery, and postpartum hospital utilization data using ICD-9 codes (Table 3). Diabetes prior to pregnancy and gestational diabetes were classifıed together to reduce potential misclassifıcation.18 A similar convention was followed for the coding of chronic and pregnancyinduced hypertension. Eclampsia was considered separately from other hypertensive disorders, including preeclampsia,
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
S544
Table 1. Maternal delivery characteristics by hemoglobinopathy status: PELL, 1998 –2006 (n [%] unless otherwise indicated) No hemoglobinopathies (nⴝ115,160)
Sickle cell disease (nⴝ663)
p-value*
⬍0.0001
Maternal ancestry American
67,096 (58.3)
322 (48.6)
African
16,317 (14.2)
96 (14.5)
Caribbean
14,752 (12.8)
160 (24.1)
South or Central American
15,881 (13.8)
80 (12.1)
Asian, Indian, other
a
970 (0.8)
4 (0.6)
Maternal age (years) ⬍20
18,629 (16.2)
97 (14.6)
20–24
32,065 (27.8)
175 (26.4)
25–29
28,348 (24.6)
178 (26.9)
30–34
21,744 (18.9)
125 (18.9)
35–39
14,405 (9.9)
ⱖ40
0.49
74 (11.2)
2,967 (2.6)
14 (2.1)
⬍12
33,057 (28.7)
173 (26.1)
12
40,977 (35.6)
222 (34.5)
13–15
27,003 (23.5)
194 (29.3)
ⱖ16
14,123 (12.3)
74 (11.2)
Education (years) 0.006
Data Analysis
Parity 1
46,601 (40.5)
276 (41.8)
2
35,061 (30.5)
196 (29.7)
3
19,501 (17.0)
107 (16.2)
ⱖ4
13,821 (12.0)
82 (12.4)
113,186 (98.3)
649 (97.9)
0.87
Plurality Singleton Twin or higher
1974 (1.7)
14 (2.1)
Private
49,896 (43.3)
246 (37.1)
Public
61,531 (53.4)
388 (58.5)
3,733 (3.2)
29 (4.4)
0.43
Insurance during pregnancy
Other/unknown
0.003
Prenatal care utilization index 4,983 (4.3)
33 (5.0)
Inadequate
Missing
16,610 (14.4)
100 (15.1)
Intermediate
10,069 (8.7)
Adequate
45,286 (39.3)
211 (31.8)
Adequate plus
38,212 (33.2)
276 (41.6)
105,234 (91.5)
589 (89.4)
⬍0.0001
43 (6.5)
Smoking during pregnancy None Any a
9800 (8.5)
because of the severity and specifıc nature of this condition. Lung disease was characterized as disorders such as chronic airway obstruction, acute and chronic respiratory failure, and nonspecifıc lung disease; asthma, bronchitis, emphysema, and allergic reactions were not included in this risk group. Ischemic heart disease, atherosclerotic conditions, myocardial degeneration, and heart failure were included under heart disease. Chronic and end-stage renal disease and unspecifıed renal diseases in pregnancy were grouped together under kidney disease. Outcome measures included maternal, fetal, and infant death. Infant death included both neonatal (⬍28 days) and postneonatal (28 –364 days) events. Preterm delivery (⬍37 weeks gestation); low birth weight (⬍2500 g); small for gestational age (SGA, birth weight ⬍10th percentile for gestational age and gender); and intrauterine growth restriction (IUGR, birth weight ⬍3rd percentile for gestational age and gender) were also measured.19 Vaginal or cesarean delivery (ICD-9 codes: 740, 742, 744, 749.9) and induction of labor (ICD-9 codes: 730.1, 731, 734, 964.9) were identifıed from both vital records and delivery hospitalization data. Infant gender, gestational age (completed weeks), and birth weight (grams) at delivery considered in these analyses were obtained from birth certifıcate or fetal death data.
0.05
70 (10.6)
Includes Chinese, Vietnamese, Cambodian, Korean, Filipino, Japanese, Laotian, Thai, other Asian or Pacific Islander, Asian Indian, Pakistani, Lebanese, Iranian, Israeli, other Middle Eastern, European, other, and unknown *p-value⬍0.05 is considered significant PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System
Bivariate analyses were conducted to compare selected demographic and clinical characteristics between deliveries to women with SCD and no documented hemoglobinopathy. Overlapping 95% CIs were also used to examine differences in the occurrence of maternal medical complications. For analyses examining fetal/infant outcomes, deliveries resulting in fetal deaths were analyzed separately from deliveries resulting in live births. Logistic regression analysis for correlated data was used to examine differences in fetal death between deliveries to women with SCD and no documented hemoglobinopathy.20 Among deliveries resulting in live births, bivariate and multivariable logistic regression analyses for correlated data were used to examine differences in the method of delivery (vaginal or cesarean delivery) and in the occurrence of adverse infant outcomes to include infant death (neonatal and postneonatal) preterm delivery, low birth weight, SGA, and IUGR. Multivariable logistic regression analyses adjusted for maternal age, education, parity, plurality, adequacy of prenatal care, insurance status, smoking during pregnancy, and infant gender. These variables were chosen as covariates because of their documented associations with maternal, perinatal, and delivery outcomes in previous studies. Maternal clinical factors were not included as covariates, as their independent relationship from the SCD diagnosis itself cannot be established. Data were analyzed with SAS software. Signifıcance was tested at ␣⫽0.05 for all analyses. This study was approved by the MDPH Research and Data Access Review (RaDAR)/IRB Committee, the Division www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
S545
Table 2. Classification of sickle cell disease and other hemoglobinopathies, ICD-9
Ninety percent of all delivering women Sickle cell disease Non–sickle cell disease hemoglobinopathies with SCD during this a 282.41 (thalassemia-Hb-S without crisis) 282.49 (other thalassemias) time period were of African descent. 282.42 (thalassemia-Hb-S with crisis)a 282.7 (other hemoglobinopathies) Among SCD women b 282.6 (sickle cell anemia, NOS) of African descent, 282.61 (Hb-S without crisis)b 49% were African282.62 (Hb-S with crisis)b Americans; 24% were b Caribbean; the others 282.63 (sickle cell/Hb-C disease) were predominately b 282.64 (sickle cell/Hb-C disease with crisis) from Africa or South/ 282.68 (other sickle cell without crisis)b Central America. Ma282.69 (sickle cell anemia, NOS) ternal characteristics, such as maternal age a Added to the ICD-9 in October 1, 200316 b and most education Modified October 1, 2003 categories were simiNOS, not otherwise specified lar among women with and without of Healthcare Finance and Policy Data Protection Committee, SCD (Table 1). Parity and plurality did not differ signifıthe BU School of Public Health, and the CDC Human Subjects cantly among the groups. Over 58% of SCD deliveries Research Committees in June 2009, and analysis was conwere covered by public insurance, yet nearly 26% of deducted June–August 2009. liveries to women with SCD had less-than-adequate prenatal care utilization, despite higher percentages of adeResults quate-plus utilization compared to women with no reported hemoglobinopathies (41.6% vs 33.2%). The Among 708,467 Massachusetts resident deliveries during prevalence of smoking during pregnancy among women 1998 –2006, 116,097 were to women of African descent with SCD was higher than among those with non-SCD (Figure 1). There were 116,076 live births and reportable hemoglobinopathies but barely signifıcant (10.6% vs fetal deaths delivered by 84,561 unique women of African 8.5%, p⫽0.05). descent during this period. Among the study population, Based on non-overlapping 95% CIs, preeclampsia, 488 women had SCD, and 83,877 were identifıed with no lung disease, and heart disease were also reported more reported hemoglobinopathy.
Table 3. Prevalence of maternal medical conditions among deliveries: PELL, 1998 –2006
Maternal medical risk conditions
ICD-9 codes
No hemoglobinopathies n (% [95% CI])
Sickle cell disease n (% [95% CI])
Diabetes
250 (250.00–250.33, 250.40–250.93), 648.0, 648.8
5948 (5.2 [5.0, 5.3])
48 (7.2 [5.3, 9.2])
Hypertension
401.0, 401.1, 401.9, 402.00–405.99, 403.0, 403.1, 403.9, 437.2, 642.00–642.24, 642.30–642.34, 642.90–642.94, 997.91
9052 (7.9 [7.7, 8.0])
66 (10.0 [7.7, 12.2])
Eclampsia
642.60–642.64
2104 (1.8 [1.7, 1.9])
15 (2.3 [1.1, 3.34])
Preeclampsia
642.40–642.44, 642.50–642.54, 642.70–642.74
5137 (4.5 [4.3, 4.6])
50 (7.5 [5.5, 9.6])
Lung disease
496, 518.89, 518.8, 518.81–518.84
7864 (6.8 [6.7, 7.0])
73 (11.0 [8.6, 13.4])
Heart disease
429–429.9, 428–428.4, 428.9, 414–414.3, 414.8
1035 (0.9 [0.8, 1.0])
18 (2.7 [1.5, 4.0])
Kidney disease
646.2, 593.9, 585, 585.6, 585.9, 646.2–646.24
1193 (1.0 [0.9, 1.1])
8 (1.2 [0.4, 2.0])
PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System
April 2010
S546
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
Discussion
Figure 1. Study population selection and case classification: PELL, 1998 –2006 PELL, Pregnancy to Early Life Longitudinal Data System; SCD, sickle cell disease
often among deliveries in women with SCD compared to deliveries in women with no reported hemoglobinopathies (Table 3). Reporting of diabetes trended toward being higher among deliveries to women with SCD but was not statistically different. The two groups showed no differences for hypertension, eclampsia, and kidney disease. Crude and adjusted ORs of the association between maternal SCD and perinatal outcomes are shown in Table 4. There were fewer than seven maternal and infant deaths among deliveries to women with SCD (exact number not shown because of confıdentiality requirements for small cell size suppression); therefore, ORs were not calculated for these outcomes. Fetal demise was twice as likely among SCD deliveries relative to deliveries to women with no reported hemoglobinopathies. Less-advantageous infant outcomes were associated with SCD. The odds of preterm delivery were 50% higher, and the odds of low birth weight were 70% higher among SCD deliveries than among non-SCD deliveries in multivariate models. SGA and IUGR in infants were 30% higher among deliveries to women with SCD compared to women with no hemoglobinopathy after adjustment for covariates; but the OR for IUGR did not reach significance. Cesarean delivery was signifıcantly associated (AOR⫽1.3; 95% CI⫽1.1, 1.5) and induction of labor was marginally associated (AOR⫽1.2, 95% CI⫽1.0, 1.5) with SCD compared to no hemoglobinopathies, in adjusted models.
Among a population of Massachusetts women of African ancestry, SCD remains a major risk factor for pregnancyand delivery-related complications and is associated with adverse infant outcomes, despite adjustment for demographic, behavioral characteristics, access to insurance, and prenatal care utilization. Massachusetts is a state with a well-regionalized system of perinatal care and advanced medical technology, where women delivering with SCD have mostly similar demographic characteristics to other women of African descent. However, women with SCD still have higher risks of fetal death, preterm births, low– birth weight births, cesarean delivery, and induction of labor. Using a population-based linked data system with race/ancestry, sociodemographic and clinical data adds to the current body of knowledge regarding the association between SCD and adverse pregnancy and delivery outcomes. These data also enable the analysis of women affected with SCD, rather than solely the analysis of SCD hospitalizations as seen in most current cross-sectional studies.6,21 Overall, intensive prenatal care utilization was highest among women with SCD compared to women without SCD. Further analysis of the APNCU Index (data not shown) suggests that women with SCD do not enter care earlier than other women of African descent, but the services they receive once in care are considerably more intensive, which may indicate a heightened clinical sensitivity to their potential pregnancy-related complications. Nearly 22% of women in both groups had lessthan-adequate prenatal care, including almost 15% with inadequate or no care. Among women with SCD, in particular, prenatal care plays an important role in the management of anemia and amelioration of crisis episodes through iron and folate supplementation, aggressive treatment of infections, and monitoring of fetal growth and well-being.4,15,22 Additionally, enhanced preconception care could potentially create opportunities for early baseline health maintenance; preventive health practices (including smoking cessation); and counseling.23,24 Medical conditions during pregnancy involving the heart and lungs were more common among women with SCD. A recent study of hospital discharges in the U.S. reported that women with non-inflammatory vasculopathies (SCD, primary pulmonary hypertension, and systemic sclerosis) have increased odds of hypertensive disorders including preeclampsia,21 and this may explain the higher percentages of preeclampsia seen in the current study. This population-based study reaffırms the increased risk of preterm delivery for women with SCD reported by other, more-restricted studies.6,8,9 Yet the mechanism for www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
Table 4. Association between maternal sickle cell disease (SCD) and perinatal outcomes: PELL, 1998 –2006
S547
number of pregnancies among women with SCD. Prior to No hemoglobinopathy SCD Unadjusted a October 2003, ICD-9 Outcomes (nⴝ115,160; n) (nⴝ663; n) OR (95% CI) AOR (95% CI) codes lacked specifıcFetal mortality 916 14 2.7 (1.6, 4.5)* 2.2 (1.2,4.2)* ity to distinguish Infant outcomesb between sickle-⫹ thalassemia (282.41, Pretermc 14,405 127 1.7 (1.4,2.0)* 1.5 (1.2,1.8)* 282,42);  thalasse⬍34 weeks 4,711 44 1.7 (1.2,2.3)* 1.5 (1.1,2.0)* mia (a milder hemo34–36 weeks 9,694 83 1.6 (1.2,2.0)* 1.4 (1.1,1.8)* globinopathy); and Low birth weightd 2,126 23 1.9 (1.3,3.0)* 1.7 (1.1,2.6)* other thalassemias (282.49).16 The curSGAe 10,733 79 1.3 (1.0,1.7)* 1.3 (1.0,1.7)* rent study also does IUGRf 4,649 34 1.3 (0.9,1.8) 1.3 (0.9,1.8) not include spontaDelivery outcomesb neous fırst trimester Cesarean 29,748 207 1.3 (1.1,1.5)* 1.3 (1.1,1.5)* or elective abortions because MassachuInduction 24,160 163 1.2 (1.0,1.5)* 1.2 (1.0,1.5)* setts does not require a Adjusted for maternal age, maternal years of education, parity, plurality, insurance status at delivery, reporting of fetal adequacy of prenatal care utilization, smoking during pregnancy, and infant gender b deaths at ⬍20 weeks Live births only c Less than 37 weeks gestation gestation and ⬍350 g d Less than 2500 g birth weight. Addie Small-for-gestational age; ⬍10th birth weight percentile for gestational age and gender tionally, the data f Intrauterine growth restriction; ⬍3rd birth weight percentile for gestational age and gender used in this study *p-value ⬍0.05 IUGR, intrauterine growth restriction; PELL, (Massachusetts) Pregnancy to Early Life Longitudinal Data System; were not validated by SCD, sickle cell disease; SGA, small-for-gestational-age chart review and are subject to the quality of administrative preterm delivery among women with SCD remains unrecord data. However, hospital discharge data have been known. Prematurity may be the major cause of low birth shown to provide greater sensitivity in reporting maternal weight among infants of mothers with SCD in this study; medical conditions16 than birth and fetal death certifıcate modest increases in SGA and IUGR among women with data alone.27–30 SCD were also seen despite defıning IUGR more strictly Although the current study is strengthened by includthan in previous studies (⬍3rd percentile) and adjusting ing a population of women of African descent, it is possifor gestational age and infant gender. Vascular stasis of ble that it inadvertently excluded SCD women who did the uteroplacental unit among gravid women with SCD not report any of the race and ancestry groups selected. has been proposed as a potential mechanism for fetal However, the self-reporting nature of the birth certifıcate growth restriction,25 and this may occur in the last tridata improves selection, and the Massachusetts-specifıc mester of pregnancy.26 Cesarean delivery during the last variable of ancestry increases the sensitivity of selecting trimester, for fetal (e.g., poor intrauterine growth) or women of African ancestry rather than using simply maternal (e.g., preeclampsia) medical indications/risks black race.31 Of the 10% of SCD deliveries excluded from could potentially have decreased the incidence of IUGR this study, most mothers were self-identifıed as South or among SCD women in the current study compared to Central American (n⫽33) or Asian (n⫽11); 16 were earlier studies.6,20 missing maternal ancestry or were classifıed as other. It was found that maternal mortality appeared elevated These women were not included, because the general among women with SCD compared to women with no population of women in Massachusetts who are of South/ documented hemoglobinopathy, similar to fındings in Central American or Asian ancestry are demographically other studies.6 However, because of confıdentiality redifferent than women of African ancestry and, as a genstrictions for small cell sizes, detailed data, including the eral population, do not tend to have SCD. cause of death, are not shown. Women with SCD in which the diagnosis was not Several limitations of the study are noted. Because of reported on hospital discharge records or birth and/or the lack of systematic data on maternal genotype (i.e., SS, fetal death data may have been missed in the study; this SC, S-thalassemia), the study may underestimate the April 2010
S548
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549
would also make the statistical estimates closer to the null, because of “misclassifıcation” of SCD cases into nonhemoglobinopathies. However, because SCD is a serious, life-long condition with important implications for medical, surgical, and anesthetic management,3,32 it is unlikely to be missed in hospitalizations over time, based on the linked nature of the current data. Lastly, the Massachusetts population of women in the current study may be different than women of African descent in other geographic locations, and not representative of the U.S. as a whole, because women in Massachusetts are generally older and more highly educated.33 This study illustrates how longitudinally linked data are useful for assessing factors associated with adverse health outcomes among women with specifıc medical conditions, such as SCD. Cross-sectional studies may overestimate pregnancy and delivery risk, yet underestimate individual severity of illness. The redundancy inherent in linked data systems improves the identifıcation of chronic medical conditions and complications, which may not be reported in all clinical encounters. Future SCD studies will evaluate the impact of pre- and postdelivery hospitalizations on maternal, fetal, and infant health outcomes. Despite declines in maternal and perinatal mortality rates in the U.S., sickle cell disease remains an important factor associated with pregnancy- and delivery-associated complications among deliveries to women of African descent. Although further research is needed to understand the specifıc biologic mechanisms that contribute to SCD and adverse pregnancy outcomes, there are current opportunities to improve care to women with SCD through improved preconception counseling and care; education and prevention of high-risk behaviors (e.g., smoking); early and adequate prenatal care; and appropriate identifıcation and treatment of pregnancy/labor/delivery complications. In order to disentangle the numerous contributors to racial and ethnic disparities in U.S. perinatal outcomes, it is important to understand the impact of specifıc medical conditions that are more prevalent among women of African descent, such as SCD. We thank Terry Njoroge, MPH, for her assistance in the literature review of this manuscript, and Daksha Gopal, PhD, for her help in the initial preparation of the analytic PELL database for this study. The fındings and conclusions of this report are those of the authors and do not necessarily represent the offıcial position of the CDC. This study was funded by CDC contracts 200 –200722782 and 200-2006-15969 to Boston University School of Public Health.
No fınancial disclosures were reported by the authors of this paper.
References 1. Steiner CA, Miller JL. Sickle cell disease in U.S. hospitals. Healthcare Cost and Utilization Project, Statistical Brief # 21. Agency for Healthcare Research and Quality, USDHHS, 2006. 2. Panepinto JA, Brousseau DC, Hillery CA, Scott JP. Variation in hospitalizations and hospital length of stay in children with vaso-occlusive crises in sickle cell disease. Pediatr Blood Cancer 2005;44(2):182– 6. 3. Adam S, Jonassaint J, Kruger H, et al. Surgical and obstetric outcomes in adults with sickle cell disease. Am J Med 2008;121:916 –21. 4. American College of Obstetricians and Gynecologists. Hemoglobinopathies in pregnancy. ACOG Practice Bulletin No. 78. Obstet Gynecol 2007;109:229 –37. 5. Creary M, Williamson D, Kulkarni R. Sickle cell disease: current activities, public health implications, and future directions. J Womens Health (Larchmt) 2007;16(5):575– 82. 6. Villers MS, Jamison MG, De Castro LM, James AH. Morbidity associated with sickle cell disease in pregnancy. Am J Obstet Gynecol 2008:199:125.e1–125.e5. 7. James AH, Jamison MG, Biancazio LR, Meyers ER. Venous thromobembolism during pregnancy and the postpartum period: incidence, risk factors, and mortality. Am J Obstet Gynecol 2006;194:1311–5. 8. Nabukera SK, Wingate MS, Owen J, et al. Racial disparities in perinatal outcomes and pregnancy spacing among women delaying initiation of childbearing. Matern Child Health J 2009;13:81–9. 9. Lu MC, Halfon N. Racial and ethnic disparities in birth outcomes: a life-course perspective. Matern Child Health J 2003; 7:13–30. 10. Shen JJ, Tymkow ND, MacMullen N. Disparities in maternal outcomes among four ethnic populations. Ethn Dis 2005;492–7. 11. Serjeant, GR, Look Loy L, Crowther M, Hambleton IR, Thame M. Outcome of pregnancy in homozygous sickle cell disease. Obstet Gynecol 2004;103:1278 – 85. 12. Sun PM, Wilburn W, Raynor BD, Jamieson D. Sickle cell disease in pregnancy: Twenty years of experience at Grady Memorial Hospital, Atlanta, Georgia. Am J Obstet Gynecol 2001;184:1127–30. 13. Smith JA, Espeland M, Bellevue R, Bonds D, Brown AK, Koshy M. Pregnancy in sickle cell disease: experience of the Cooperative Study of Sickle Cell Disease. Obstet Gynecol 1996;87: 199 –204. 14. Clements KM, Barfıeld W, Kotelchuck M, Lee KG, Wilber N. Birth characteristics associated with early intervention referral, evaluation for eligibility, and program eligibility in the fırst year of life. Matern Child Health J 2006;10:433– 41. 15. Rappaport VJ, Velazquez M, Williams K. Hemoglobinopathies in pregnancy. Obstet Gynecol Clin N Am 2004;31: 287–317. 16. Agency for Healthcare Research and Quality. Healthcare Cost and Utilization Project Clinical Classifıcations Software (CCS). Rockville MD. www.hcup-us.ahrq.gov/toolssoftware/ ccs/ccs.jsp. www.ajpm-online.net
Barfıeld et al / Am J Prev Med 2010;38(4S):S542–S549 17. Kotelchuck M. The Adequacy of Prenatal Care Utilization Index: its U.S. distribution and association with low birthweight. Am J Public Health 1994; 84:1486 –9. 18. Lyndon-Rochelle MT, Holt VL, Cardenas V, et al. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certifıcates and in hospital discharge data. Am J Obstet Gynecol 2005:193:125–34. 19. Oken E, Kleinman KP, Rich-Edwards J, Gillman MW. A nearly continuous measure of birth weight for gestational age using a U.S. national reference. BMC Pediatr 2003;3:6. 20. Hanley JA, Negassa A, Edwardes MD, Forrester JE. Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 2003;157:364 –75. 21. Chakravarty EF, Khanna D, Chung L. Pregnancy outcomes in systemic sclerosis, primary pulmonary hypertension, and sickle cell disease. Obstet Gynecol 2008;111:927–34. 22. Thurman AR, Steed LL, Hulsey T, Soper DE. Bacteriuria in pregnant women with sickle cell trait. Am J Obstet Gynecol 2006;194:1366 –70. 23. Lottenberg R, Hassell K. An evidence-based approach to the treatment of adults with sickle cell disease. Hematology Am Soc Hematol Educ Program 2005:58 – 65. 24. CDC. Recommendations to improve preconception health and health care—U.S. MMWR 2006;55(RR06):1. 25. Rathod KB, Jaiswal KN, Shrivastava AC, Shrikhande AV. Study of placenta in sickle cell disorders. Indian J Pathol Microbiol 2007;50:698 –701.
26. Thame M, Lewis J, Troutman H, Hambleton I, Serjeant G. The mechanism of low birth weight in infants of mothers with homozygous sickle cell disease. Pediatrics 2007;120:e686 – e693. 27. Lyndon-Rochelle MT, Holt VL, Cardenas V, et al. Accuracy of reporting maternal in-hospital diagnosis and intrapartum procedures in Washington State linked birth records. Paediatr Perinat Epidemiol 2005;19:450 –71. 28. DiGiuseppe DL, Aron DC, Ranborn L, Harper DL, Rosenthal GE. Reliability of birth certifıcate data: a multi-hospital comparison to medical records information. Matern Child Health J 2002;6:169 –79. 29. Roohan PJ, Josberger RE, Acar K, Dabir P, Feder HM, Gagliano PJ. Validation of birth certifıcate data in New York State. J Community Health 2003;28:335– 46. 30. Northam S, Knapp TR. The reliability and validity of birth certifıcates. J Obstet Gynecol Neonatal Nurs 2006;28:3–12. 31. Friedman DJ, Cohen BB, Dunn VH, et al. Topics in minority health ethnic variation and maternal risk characteristics among blacks—Massachusetts, 1987 and 1988. MMWR 1991; 40:409 –11. 32. Camous J, N’da A, Etienne-Julan M, Stéphan F. Anesthetic management of pregnant women with sickle cell disease— effect on postnatal sickling complications. Can J Anesth 2008; 55:276 – 83. 33. Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2007. Natl Vital Stat Rep 2007;57:1–103.
Did you know? The AJPM online archive includes issues back to 1998. Visit www.ajpm-online.net today!
April 2010
S549