Pregnancy complications associated with hepatitis C: data from a 2003-2005 Washington state birth cohort
American Journal of Obstetrics & Gynecology July 2008
"HCV-positive pregnant women appear to be at risk for adverse neonatal and maternal outcomes"
Some of the data in this paper were presented at the 45th Annual Meeting of the Infectious Diseases Society of America, San Diego, CA, Oct. 4-7, 2007.
Steven A. Pergam, MD1Corresponding Author Informationemail address, Chia C. Wang, MD, MS1, Carolyn M. Gardella, MD, MPH2, Taylor G. Sandison, MD, MPH1, Warren T. Phipps, MD1, Stephen E. Hawes, PhD3
Received 15 September 2007; received in revised form 20 December 2007; accepted 21 March 2008. published online 20 May 2008.
Hepatitis C virus (HCV)-positive pregnant women appear to be at risk for adverse neonatal and maternal outcomes, according to the results of a population-based birth cohort study reported in the July issue of the American Journal of Obstetrics & Gynecology......Using Washington state birth records from 2003 to 2005, the investigators compared a cohort of HCV-positive pregnant women, based on birth certificate data (n = 506), with randomly selected HCV-negative mothers (n = 2022) and drug-using HCV-negative mothers (n = 1439).....Compared with infants of HCV-negative mothers, those of HCV-positive mothers were more likely to have low birth weight (odds ratio [OR], 2.17; 95% confidence interval [CI], 1.24 - 3.80), be small for gestational age (OR, 1.46; 95% CI, 1.00 - 2.13), require assisted ventilation (OR, 2.37; 95% CI, 1.46 - 3.85), and be admitted to the neonatal intensive care unit (NICU; OR, 2.91; 95% CI, 1.86 - 4.55).
......The risk for gestational diabetes was greater in HCV-positive mothers with excess weight gain (OR, 2.51; 95% CI, 1.04 - 6.03). Even in comparison with the drug-using cohort, infants of HCV-positive mothers were still at greater risk for NICU admission and assisted ventilation......Limitations of this study include possible ascertainment bias; HCV status identified through self-report on birth certificate data, likely underestimating the true prevalence; misclassification bias; lack of data concerning HCV viremia; use of International Classification of Diseases, Ninth Revision (ICD-9), codes to identify drug use, resulting in possible misclassification; possible residual confounding; and lack of data on maternal HIV status.......HCV infection may have a much greater effect on pregnancy and neonatal outcomes then [sic] previously reported, indicating that routine HCV screening in pregnant women may need to be reconsidered," the study authors conclude. "Future prospective studies evaluating placental involvement with HCV, role of viral genotype and HCV viremia, and further emphasis on drug use and socioeconomic status are warranted.... In addition, the ongoing development of new potent antiviral agents to treat hepatitis C may soon modify our approach to hepatitis C in pregnancy.....HCV-positive pregnant women appear to be at risk for adverse neonatal and maternal outcomes..."
The objective of the study was to determine the effect of hepatitis C virus (HCV) on selected maternal and infant birth outcomes.
This population-based cohort study using Washington state birth records from 2003 to 2005 compared a cohort of pregnant women identified as HCV positive from birth certificate data (n = 506) to randomly selected HCV-negative mothers (n = 2022) and drug-using HCV-negative mothers (n = 1439).
Infants of HCV-positive mothers were more likely to be low birthweight (odds ratio [OR], 2.17; 95% confidence interval [CI] 1.24, 3.80), to be small for gestational age (OR, 1.46; 95% CI, 1.00, 2.13), to need assisted ventilation (OR, 2.37; 95% CI, 1.46, 3.85), and to require neonatal intensive car unit (NICU) admission (OR, 2.91; 95% CI, 1.86, 4.55). HCV-positive mothers with excess weight gain also had a greater risk of gestational diabetes (OR, 2.51; 95% CI, 1.04, 6.03). Compared with the drug-using cohort, NICU admission and the need for assisted ventilation remained associated with HCV.
HCV-positive pregnant women appear to be at risk for adverse neonatal and maternal outcomes.
Approximately 4 million people (from Jules: its probably 5 million, as presented in oral by Brian Edlin at AASLD, I think 2006) in the United States are infected with hepatitis C virus (HCV).1, 2 In pregnant women, the prevalence of HCV is generally estimated to range between 1% and 2% but may be as high as 4% in some inner-city populations.3, 4, 5, 6, 7, 8, 9 With 4.1 million annual births in the United States, approximately 40,000-80,000 pregnant women per year would be expected to have a history of HCV infection.
Although HCV affects a significant number of pregnant women, there is little research regarding the impact of HCV on pregnancy outcomes. Prior studies of HCV and pregnancy have focused on vertical transmission of HCV infection, without examining the potential effects of chronic HCV infection on maternal health, complications during delivery, and neonatal health. Because identification of adverse outcomes could effect current screening recommendations, such information could have widespread public health implications.
In 2003, Washington state added maternal HCV status to its birth certificate data collection form, providing an opportunity to assess the potential effects of HCV on pregnancy. The objective of this study was to evaluate associations between maternal HCV infection, maternal pregnancy complications, and neonate health in Washington state from 2003 to 2005.
This study demonstrates that HCV infection in pregnant women is associated with several adverse pregnancy and neonatal outcomes. We found that women with HCV may be at higher risk for PROM, and those with excess weight gain during pregnancy have an increased risk of GDM. Likewise, infants born to HCV-infected women were more likely to be LBW, SGA, require NICU admission, and need assisted ventilation. NICU admission and need for assisted ventilation remained strongly associated with maternal HCV, regardless of drug use.
This large, population-based cohort demonstrates a previously unreported association between HCV and gestational diabetes. Studies have demonstrated an association between type 2 diabetes mellitus and HCV.19, 20, 21, 22, 23, 24, 25, 26 HCV has also been shown to be directly involved in the development of insulin resistance, particularly in individuals with genotype 3 HCV infection.27, 28, 29 Although the mechanism is unknown, HCV has been detected in the pancreas and may cause _-cell dysfunction.20, 30 Risk for GDM is also thought to be due to increases in insulin resistance and BMI,31 and it is possible that by working through similar pathways, HCV could be associated with an increased risk of GDM.
Few studies of HCV and pregnancy have ascertained pregnancy outcomes. Increased risks for obstetric complications associated with HCV infection have not been noted in previous studies, but these were limited by small sample sizes.32, 33, 34 In 1 study, viremia in HCV-seropositive mothers was associated with PROM.35 In this study, we also report evidence for a possible association between maternal HCV infection and PROM.
In neonates, the paucity of data, coupled with the relative infrequency of many adverse outcomes, has made it difficult to quantify risk associated with maternal HCV. In previous literature, Apgar scores for children born to HCV-infected and non-infected women appear similar.33, 34, 36 Studies evaluating prematurity, however, have been contradictory; 2 studies found no difference in HCV-positive women,33, 36 whereas another study found high rates of prematurity and spontaneous abortion in acute disease.37 Absent data regarding maternal drug use, socioeconomic status, and other risk factors for HCV acquisition and poor neonatal outcome limit these cross-sectional analyses.
Our findings suggest greater risk of neonatal morbidity associated with HCV than have been previously reported, even taking into account socioeconomic correlates and maternal drug use. We found that HCV was associated with having LBW, prematurity, and SGA. Additionally, NICU admission and assisted ventilation were associated with HCV. Drug use has a strong association with poor neonatal outcomes11, 12, 38 and has the potential to be a major driving force behind these results. To evaluate the role of drug use, drug-using members of the HCV cohort were compared with a HCV-negative, drug-using cohort. Although being LBW, prematurity, and SGA were not associated with HCV infection in maternal drug users, the need for assisted ventilation and NICU admission remained strongly associated with HCV.
HCV appears to be associated with multiple adverse outcomes, yet mechanisms for this increased risk are unknown. Epidemiologically, HCV infection may be a surrogate marker for other high-risk behaviors or factors that could increase the risk of poor outcomes. Physiologically, vascular compromise of the placenta can lead to poor neonatal outcomes.39, 40 Because HCV can cause vasculitis, involvement of the placental vasculature could explain the growth retardation and higher risk of complications after delivery. Pathologic examination of placental changes in women with HCV infection and further prospective evaluation of other cofactors may help elucidate reasons for increasing pregnancy risk associated with HCV.
This analysis does have limitations imposed by the data. Because universal HCV testing is not mandatory during pregnancy, patients noted to be HCV positive may have significant risk factors that initiated provider screening, introducing an ascertainment bias. Although this is unavoidable because of the limitations of birth certificate data collection, if our outcomes of interest were associated with an increased rate of screening mothers for HCV, we may be overestimating the risk associated with exposure. Missing data also limit our analysis when analyzing our drug-using, HCV-negative cohort.
The prevalence (0.2%) of HCV in this data set is substantially lower than that reported by seroprevalence studies conducted in Europe and North America.3, 4, 5, 6, 7, 8, 9 This is a reflection of the fact that in our data, HCV is identified through self-report on birth certificate data and therefore likely underestimates the true prevalence of HCV in the study population. It is likely not all exposed mothers were included in the HCV-positive cohort and that some HCV-positive mothers may have been included in our unexposed groups. This issue could result in a misclassification bias that may have led to an underestimation of the effect of HCV on maternal and neonatal outcomes. Furthermore, because our data set did not contain information about HCV viremia, but only HCV serostatus, the effect of replicating virus on the outcomes of interest could not be determined.
Regardless, the low prevalence of HCV infection in the general population means that these issues should not have altered our results substantially. Mothers documented with HCV in the birth certificates were also most likely a combination of those with and without viremia, and whether replicating virus had any effect on these outcomes could not be determined.
In addition, because of our use of ICD-9 codes to identify drug use, we cannot rule out misclassification by drug-use status. The lack of an association in the drug-using cohorts for certain outcomes associated with drug use, such as being LBW, premature, and SGA, also suggests that drug use may still be a residual confounder in some of these results. Similarly, underreporting may have occurred with respect to maternal smoking and alcohol use during pregnancy, resulting in misclassification and the potential for residual confounding. However, because more accurate assessment tools of self-reported drug use have been shown to be inadequate, prospective data may not help to limit further confounding.41, 42, 43
A final limitation of our analysis is that data on maternal HIV status are not included in the data set. Because of concerns about patient privacy, HIV status is protected information in Washington state and is not included in the data set. However, approximately 6% of HCV infected individuals in the United States are believed to be coinfected with HIV.2 If this prevalence is similar in the present study cohort, then confounding because of HIV infection should not have had a great impact on our results.
Testing at-risk mothers for HCV is advocated by both the Centers for Disease Control and Prevention44 and the American College of Obstetricians and Gynecologists,45 but routine HCV screening of all pregnant patients is not considered cost effective.46 However, practitioners need to be aware that risk-factor screening is not without limitations. These guidelines may miss numerous at-risk women because up to 40% of pregnant women with HCV have no identifiable risk factor.5 At-risk screening may detect only half of the women exposed to HCV, indicating that this method of screening may be inadequate.7 Furthermore, such risk factors are underascertained by health care providers during pregnancy.47, 48
Pregnancy also is an opportunity to identify early HCV infection because many who are found to be positive during pregnancy are unaware of their serologic status.34 Regardless, further options for therapy during pregnancy, methods of decreasing perinatal transmission, and future studies on HCV's effect on pregnancy are needed.
HCV infection may have a much greater effect on pregnancy and neonatal outcomes then previously reported, indicating that routine HCV screening in pregnant women may need to be reconsidered. In this study, HCV infection in women with excess weight gain was strongly associated with gestational diabetes. Further studies are needed to clarify this association and evaluate the additional potential role of HCV on development of gestational diabetes.
HCV was also associated with poor neonatal outcomes, including being low birth weight, being small for gestational age, the need for assisted ventilation, and NICU admission. Future prospective studies evaluating placental involvement with HCV, role of viral genotype and HCV viremia, and further emphasis on drug use and socioeconomic status are warranted. Data from such studies may impel clinicians to reexamine current guidelines for HCV screening in pregnant women. In addition, the ongoing development of new potent antiviral agents to treat hepatitis C may soon modify our approach to hepatitis C in pregnancy.
Materials and Methods
We conducted a population-based cohort study, using Washington state singleton birth records from 2003 to 2005. Birth certificate data for mothers and infants were linked to the Comprehensive Hospital Abstract Reporting System (CHARS), created by the Washington State Department of Health, which contains hospital discharge records of inpatients in nonfederal facilities in the state. Birth certificates use a check-box format to collect information on demographic characteristics, complications, procedures, and newborn conditions. Birth certificates are completed by trained medical records staff with information abstracted from patient records.10
The exposed cohort consisted of women noted to be HCV positive on a birth certificate check box (n = 506). The unexposed cohort consisted of 4 HCV-negative mothers per exposed and were randomly selected from the same data set and frequency matched by birth year (n = 2024). Two women in this comparison group were excluded after CHARS data indicated they were HCV positive (n = 2022). To create a second comparison group consisting of HCV-negative mothers with a history of drug use, the entire CHARS data set was queried a second time to select women without HCV infection but with a history of drug use identified in CHARS using the methods described below (n = 1439). Finally, we identified women in the HCV-positive group with a history of drug use using similar methods.
Defining drug use and prenatal care
Maternal drug use was ascertained by reviewing CHARS data and identifying all International Classification of Diseases, ninth revision (ICD-9) codes associated with a history of opioid, cocaine, and methamphetamine use as has been demonstrated in other studies.11, 12 Drug use was considered positive if the following ICD-9 codes were assigned to the mother: 304.0 (opioid dependence), 304.2 (cocaine dependence), 304.4 (amphetamine dependence), 304.7 (combinations of opioid type drug and any other), 304.9 (unspecified drug dependence), 305.5 (opioid abuse), 305.6 (cocaine abuse), 305.7 (amphetamine or related acting sympathomimetic abuse), and 292 (drug withdrawal syndromes) or the neonate: 760.72 (noxious influences on fetus or newborn via placenta or breast milk, narcotics, excluding anesthetic and analgesic drugs administered during labor and delivery), 760.75 (noxious influences on fetus or newborn via placenta or breast milk, cocaine), and 779.5 (drug withdrawal syndrome in newborn). Prenatal care was assessed using the Adequacy of Prenatal Care Utilization (APCU) index, which uses trimester at first prenatal visit and number of visits to determine a patient's level of care (inadequate, intermediate, adequate, adequate-plus).13
Defining maternal and neonatal outcomes
Maternal pregnancy complications of gestational diabetes (GDM) and premature ruptured membranes (PROM) were determined from birth certificate data. Because GDM has been previously validated using supplemental CHARS data, we added additional cases using the ICD-9 code for gestational diabetes (648.8).14
Mothers were classified by body mass index (BMI), using their prepregnancy weight, as underweight (less than 18.5 kg/m2), normal (18.5 to 24.9 kg/m2), overweight (25.0 to 29.9 kg/m2), and obese (greater than 30 kg/m2).15 Pregnancy-related weight gain was described using the Institute of Medicine's guidelines, which are based on kilograms of weight gained and mother's baseline BMI. These guidelines recommend pregnancy weight gain of 28-40 pounds for women with a BMI of less than 19.8 kg/m2, 25-35 pounds for women with a BMI of 19.8 to 26 kg/m2, and 15-25 pounds for women with a BMI of 26.1 to 29.0 kg/m2.16 Studies suggest that a weight gain of 15-25 pounds is also appropriate for women with a BMI greater than 29.0 kg/m2.17 Patients under recommended levels were considered as insufficient, those who gained above recommended guidelines as excess, and those within guidelines as appropriate.
Neonatal outcomes determined from birth certificate data included having low birthweight (LBW) (less than 2500 g), prematurity (less than 37 weeks), neonatal jaundice, low Apgar score (less than 7 at 5 minutes), neonatal intensive care unit (NICU) admission, and any need for assisted ventilation. Being small for gestational age (SGA) was determined using previously described methods,18 applied to Washington data from 1989 to 2001 (a total of 1,037,558 births). Births were classified as SGA if they were in the lowest 10% of birthweights for a given gestational age. Neonatal jaundice data were collected using the ICD-9 code 774 (neonatal jaundice) from CHARS data.
We also reviewed ICD-9 diagnostic codes to evaluate reasons for NICU admission. Using CHARS data, reasons were classified as being due to respiratory, infectious, cardiac, metabolic/gastrointestinal, hematologic, congenital abnormalities, or drug use/withdrawal causes. The ICD-9 codes utilized for infectious etiologies included 771 (infections specific to the perinatal period), 770 (congenital pneumonia), and V02.51 (carrier or suspected carrier of infectious disease-group B streptococcus), among others. All reasons were not mutually exclusive.
Multivariate logistic regression methods were used to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for outcomes associated with maternal HCV infection. A priori confounders included maternal age, race (white, black, Hispanic, Asian/Pacific Islander, Native American), tobacco use, alcohol use, drug use, and prenatal care usage. Because of missing prenatal care data in the HCV-negative, drug-using cohort, this variable was removed from the model when comparing drug-using cohorts.
Other factors, including mother's insurance status, occupation, educational level, pregnancy weight gain, parity, and maternal infections (hepatitis B virus, Chlamydia trachomatis, Neisseria gonorrhea, herpes simplex-2 virus, group B streptococcus, syphilis, and chorioamnionitis), were also considered potential confounders. We adjusted for these confounders if they altered the OR for the outcomes of interest by 10% of more. To evaluate the role of drug use in these outcomes, we also performed a subanalysis by stratifying by drug use and then using similar multivariate methods. In this subanalysis when comparing drug users, the mother's occupation and maternal pregnancy weight gain were not included in the model because of significant missing data.
To explore factors associated with NICU admission, Mantel-Haensel stratified analysis was used to determine condition-specific ORs and 95% CIs between children born to HCV-positive mothers and those in the other cohorts stratified by drug use. Because of small numbers, we adjusted only for maternal smoking and prenatal care utilization when comparing the non-drug-using cohorts. In the drug-using cohorts, we adjusted for maternal smoking only because of the large percentage of missing data in prenatal care usage in the HCV-negative drug-using cohort. The University of Washington Institutional Review Board approved of this study. All statistical analysis was performed using STATA 9.0 (StataCorp, College Station, TX).
During 2003-2005, 506 mothers (0.2%) of 240,131 singleton deliveries were reported to have HCV, and 2022 women were randomly selected to be HCV-negative controls. In addition, 1439 (0.6%) HCV-negative women and 124 (25%) of the HCV-positive mothers were identified as drug users.
As shown in Table 1, HCV-positive women were similar to HCV-negative drug users in many respects and less similar to the randomly selected comparison group. Drug-using women, regardless of HCV status, were more likely to be unmarried, have less education, have inadequate prenatal care, have received Medicaid, and smoke than their randomly selected HCV-negative counterparts. The lowest rates of prenatal care and highest rates of Medicaid use were found in the HCV-negative, drug-using group.
HCV-positive women were somewhat more likely to have GDM than HCV-negative women (OR, 1.53; 95% CI, 0.85, 2.27) (Table 2). However, among women with excess weight gain during pregnancy, HCV infection was strongly associated with GDM (OR, 2.51; 95% CI, 1.04, 6.03). This association was not observed in women with insufficient or adequate weight gain (OR, 0.89; 95% CI, 0.28, 2.80 and OR, 1.14; 95% CI, 0.38, 3.39, respectively). There was also a trend toward an association between HCV and PROM (OR, 1.66; 95% CI, 0.93, 2.96).
To determine the impact of drug use, a stratified analysis was performed based on drug use history (Table 3). Among non-drug-using mothers, HCV positivity (n = 382) remained associated with GDM in women with excess weight gain (OR, 3.09; 95% CI, 1.29, 7.42). In addition, a similar trend toward an association between HCV infection and PROM (OR, 1.74; 95% CI, 0.98, 3.25) was observed.
Similarly, when the analysis was limited to HCV-positive (n = 124) and HCV-negative (n = 1439) mothers with a history of drug use, HCV remained somewhat associated with increased risk of PROM (OR, 1.33; 95% CI, 0.60, 2.95) (Table 3). Because of the small number of HCV-positive drug users with GDM (n = 2), no comment could be made in regard to this outcome.
HCV infection was associated with an increased risk of the infant being LBW (OR, 2.17; 95% CI, 1.24, 3.80), SGA (OR, 1.46; 95% CI, 1.00, 2.13), requiring NICU admission (OR, 2.91; 95% CI, 1.86, 4.55), and needing assisted ventilation (OR, 2.37; 95% CI, 1.46, 3.85) (Table 2). There were nonsignificant trends for low Apgar score (OR, 1.53; 95% CI, 0.93, 2.54), prematurity (OR, 1.54; 95% CI, 0.97, 2.43), and neonatal jaundice (OR, 1.25; 95% CI, 0.82, 1.90) being associated with HCV.
Analyses stratified by drug use again yielded similar findings for non-drug-using mothers (Table 3). HCV was associated with increased risk of being LBW (OR, 2.15; 95% CI, 1.19, 3.90), SGA (OR, 1.62; 95% CI, 1.07, 2.43), requiring NICU admission (OR, 3.00; 95% CI, 1.85, 4.86), needing assisted ventilation (OR, 2.25; 95% CI, 1.40, 3.63), and prematurity (OR, 1.69; 95% CI, 1.04, 2.73). Neonates born to HCV-positive mothers were somewhat more likely to have a low Apgar score (OR, 1.42; 95% CI, 0.85, 2.37), but this association did not reach statistical significance. In non-drug-using mothers, HCV was not associated with neonatal jaundice (OR, 1.06; 95% CI, 0.67, 1.69).
Finally, when analysis was restricted to mothers with a history of drug use, maternal HCV remained strongly associated with NICU admission (OR, 2.80; 95% CI, 1.83, 4.29) and need for assisted ventilation (OR, 1.82; 95% CI, 1.03, 3.22) (Table 3). Neonates born to HCV-positive, drug-using mothers did not have an increased risk of being LBW, being premature, or having a low Apgar score.
Reasons for NICU admission
Children born to HCV-positive mothers were admitted to the NICU (n = 93) for a variety of reasons, most commonly with respiratory problems (38.7%), because of maternal drug use/withdrawal (35.5%), early gestational age/prematurity (37.6%), and infections (26.9%). When stratified by drug use, HCV status in non-drug-using mothers was associated with admission to the NICU for respiratory (OR, 3.25; 95% CI, 1.62, 6.51) and congenital reasons (OR, 4.17; 95% CI, 1.18, 14.8) (Table 4). In these same 2 cohorts, there were also nonsignificant trends toward an association between maternal HCV status and gestational age/prematurity (OR, 1.71; 95% CI, 0.83, 3.56), infectious (OR, 2.08; 95% CI, 0.92, 4.72), cardiac (OR, 1.41; 95% CI, 0.49, 4.01), metabolic /gastrointestinal (OR, 1.24; 95% CI, 0.49, 3.18), and hematologic reasons (OR, 3.41; 95% CI, 0.87, 13.3) for NICU admission.
In drug-using women, similar findings were seen (Table 4). Maternal HCV status was associated with admission to the NICU for respiratory (OR, 1.91; 95% CI, 1.05, 3.46), infectious (OR, 2.41; 95% CI, 1.31, 4.44), and metabolic/gastrointestinal reasons (OR, 2.80; 95% CI, 1.49, 5.28). For the remaining outcomes in the drug-using cohorts, there were again nonsignificant trends in all other reasons for admission. Because of small numbers, no conclusions could be drawn in NICU admissions because of congenital abnormalities.