Abstract
Background
Neonatal withdrawal syndrome, which is associated most frequently with opioid use in pregnancy, is an emerging public health concern, with recent studies documenting an increase in the rate of U.S. infants diagnosed.
Purpose
This study examined neonatal withdrawal syndrome diagnosis among Michigan infants from 2000 to 2009 and hospital length of stay (LOS) between infants with and without the syndrome for a subset of years (2006–2009).
Methods
Michigan live birth records from 2000 to 2009 were linked with hospital discharge data to identify infants with neonatal withdrawal syndrome. Linked data were restricted to infants born between 2006 and 2009 to examine the difference in hospital LOS between infants with and without the syndrome. Multivariable regression models were constructed to examine the adjusted impact of syndrome diagnosis on infant LOS and fit using negative binomial distribution. Data were analyzed from July 2011 to February 2012.
Results
From 2000 to 2009, the overall birth rate of infants with neonatal withdrawal syndrome increased from 41.2 to 289.0 per 100,000 live births (p<0.0001). Among infants born from 2006 to 2009, the average hospital LOS for those with the syndrome was between 1.36 (95% CI=1.24, 1.49) and 5.75 (95% CI=5.41, 6.10) times longer than for infants without it.
Conclusions
Diagnosis of neonatal withdrawal syndrome increased significantly in Michigan with infants who had the syndrome requiring a significantly longer LOS compared to those without it.
Introduction
With prescription opioid use among women of reproductive age increasing since the late 1990s,1,2 concern regarding the effect on the fetus and newborn has grown. Neonatal withdrawal syndrome is an emerging public health concern,3–5 with recent studies documenting almost a threefold increase in the rate of U.S. infants diagnosed in the past decade.6,7 Neonatal withdrawal syndrome risk is greatest with maternal usage of opioids analgesics and heroin4,8; however, prenatal usage of non-narcotic drugs can cause similar neonatal withdrawal symptoms.3
Symptoms usually relate to central nervous system irritability and gastrointestinal dysfunction such as tremors, irritability, increased muscle tone, and difficulty sleeping.9,10 Neonatal withdrawal syndrome infants are more likely to be born preterm6; be admitted to a neonatal intensive care unit (NICU)11; and need longer hospitalization (measured as length of stay [LOS]).4–7 Little information is available that quantifies the additional LOS of neonatal withdrawal syndrome infants compared to other infants after controlling for maternal and infant factors.6 Therefore, the purpose of the study was to examine withdrawal syndrome diagnosis among Michigan infants from 2000 to 2009 and to compare LOS between infants with and without withdrawal syndrome for a subset of years (2006–2009). Infants with withdrawal syndrome were hypothesized to require a longer LOS than infants without withdrawal syndrome, after adjusting for factors affecting LOS.
Methods
Study Population and Data Sources
A retrospective cohort population-based study was conducted using live birth certificate and hospital discharge data for infants born to Michigan (resident) mothers. Hospital discharge data contained discharge diagnosis records for all Michigan hospitals. Live birth certificate data were linked with hospital discharge records by probabilistic, followed by deterministic, linkages performed separately for maternal and infant files. Overall, 88.5% of hospital discharge newborn records and 99.2% of maternal records are matched with birth certificate records (T Le, Michigan Department of Community Health, unpublished observations, 2005). Maternal hospital discharge records up to 1 year before and after delivery were included.
To examine maternal and infant variables related to neonatal withdrawal syndrome, most recent years of data were used (2006–2009). For withdrawal syndrome infants, the hospitalization with first withdrawal syndrome diagnosis was selected, whereas for infants without the syndrome, the first hospitalization was chosen. Hospitalizations for infants with transfers only included the first hospitalization. This study was reviewed and approved by the Michigan Department of Community Health IRB.
Measures
Live birth certificate records provided the following data: maternal age; race; education; number of named parents; tobacco use during pregnancy; Kotelchuck index (inadequate defined as <50% of expected prenatal visits12); Caesarean-section delivery; infant year of birth; gender; plurality; birthweight (low defined as <2500 g); and estimated gestational weeks. Maternal race was used as a proxy for infant’s race; only the mother acknowledged on the birth certificate was used as a proxy for social support at home.
Hospital discharge data provided infant LOS (measured in days); NICU usage; and main payment source (private insurance, Medicaid, self-pay, other). Infants with withdrawal syndrome were identified by any diagnosis of ICD-9-CM13 Code 779.5. Maternal morbidity was defined by discharge diagnoses related to mental illness (ICD-9-CM=290, 293–302, 306–319, 648.4) and preexisting chronic conditions (ICD-9-CM=240–246, 250, 393–398, 435, 440–448, 451–459, 648.0–648.1, 648.5–648.6, 745–747; VIG, unpublished observations, 2005) from maternal hospitalizations during pregnancy.
Data Analysis
Analyses were conducted from July 2011 to February 2012. Neonatal withdrawal syndrome birth rates per 100,000 live births in Michigan from 2000 to 2009 were calculated, with a chi-square test performed for linear trend analysis (Epi Info, version 7). Data from the most current years, 2006–2009, were used to measure demographic and health outcomes differences between infants with and without neonatal withdrawal syndrome using percentages (crude rate per 100) and 95% CIs.
Multivariable regression models examined the adjusted impact of withdrawal syndrome diagnosis on infant LOS and were fit using negative binomial distribution (SAS, version 9.2). Candidate covariates were included based on possible association with both withdrawal syndrome diagnosis and infant LOS and included: mother’s race; maternal hospitalization related to chronic disease and mental illness; Caesarean-section delivery; low birthweight (LBW); having only the mother acknowledged on the birth certificate; insurance type; inadequate Kotelchuck index; and smoking.14–17
Nonsignificant (p-value ≥0.002) terms (main effects and two-way interaction effects) were dropped sequentially from the model in reverse order of significance, commencing with the two-way interaction terms and removing collinear variables, until the most parsimonious model was obtained. If a two-way interaction term was significant, the main effects included in the interaction were retained. The final multivariable regression model included infant neonatal withdrawal syndrome and the following covariates: mother’s race (black, white, other); maternal hospitalization related to chronic disease and mental illness; Caesarean-section delivery; LBW; only the mother being acknowledged on the birth certificate; and inadequate Kotelchuck index. Neonatal withdrawal syndrome interaction terms included maternal hospitalization related to mental illness and LBW. Adjusted incidence rate ratios (IRRs) and 95% CIs were calculated.
Results
From 2000 to 2009, a total of 1509 infants were hospitalized with neonatal withdrawal syndrome. The syndrome birth rate per 100,000 live births increased during the study period, from 41.2 (2000) to 289.0 (2009; p<0.0001). When stratified by race, the increase was significant among white but not black infants (data not shown).
A comparison of infant and maternal characteristics and health outcomes for infants with and without withdrawal syndrome are shown in Table 1. Neonatal withdrawal syndrome infants were more likely to be white and have Medicaid as their hospital payment source. Adjusted IRRs of infant LOS for interactions between neonatal withdrawal syndrome status and LBW and maternal hospitalization related to mental illness status are shown in Table 2. Among infants with normal birthweight and whose mothers were not hospitalized because of mental illness, those with the syndrome had 5.75 times longer LOS than those without it (95% CI=5.41, 6.10). Among LBW infants, the LOS for those with the syndrome was 1.36 times on average longer than for those without it (95% CI=1.24, 1.49), whereas for infants with a maternal mental illness hospitalization, the average LOS for syndrome infants was 4.66 times longer than it was for infants without it (95% CI=4.39, 4.96).
Table 1.
Comparison of infant and maternal characteristics by neonatal withdrawal syndrome diagnosis, Michigan infants, 2006–2009
| Infants with neonatal withdrawal syndrome (n=935)a | Infants without neonatal withdrawal syndrome (n=470,329)a | |||
|---|---|---|---|---|
| % (95% CI) | Count | % (95% CI) | Count | |
| Infant demographics | ||||
| Male | 55.4 (50.7, 60.4) | 518 | 51.2 (51.0, 51.4) | 240,858 |
| White | 83.9 (78.1, 90.0) | 780 | 76.1 (75.9, 76.4) | 356,737 |
| Black | 13.7 (11.4, 16.2) | 127 | 18.5 (18.4, 18.6) | 86,724 |
| Other race | 2.5 (1.6, 3.7) | 23 | 5.4 (5.29, 5.42) | 25,093 |
| Medicaid hospital payment source | 78.5 (72.9, 84.4) | 732 | 42.3 (42.2, 42.5) | 198,341 |
| Delivery/infant outcomes | ||||
| Mode of delivery Caesarean-section | 37.0 (33.1, 41.2) | 331 | 31.4 (31.3, 31.6) | 140,669 |
| Low birthweight (<2500 g) | 23.2 (20.2, 26.5) | 215 | 8.0 (7.96, 8.12) | 37,648 |
| Premature (<37 weeks) | 24.5 (21.5, 27.9) | 228 | 10.3 (10.2, 10.4) | 48,121 |
| Small for gestational age (<10th percentile) | 15.9 (13.4, 18.7) | 146 | 4.8 (4.7, 4.9) | 22,096 |
| NICU admission | 29.6 (25.7, 33.9) | 204 | 5.9 (5.8, 6.0) | 19,969 |
| Maternal demographics | ||||
| <High school graduate | 29.1 (25.7, 32.9) | 256 | 16.5 (16.3, 16.6) | 72,895 |
| Age at delivery <21 years | 7.9 (6.2, 9.9) | 73 | 14.5 (14.4, 14.6) | 68,277 |
| Only mother acknowledged on birth certificate | 34.2 (30.6, 38.2) | 320 | 15.4 (15.3, 15.6) | 72,583 |
| Maternal health variables | ||||
| Inadequate Kotelchuck index | 35.8 (32.0, 40.0) | 315 | 11.4 (11.35, 11.55) | 51,551 |
| Smoking during pregnancy | 69.6 (64.4, 75.3) | 640 | 17.5 (17.3, 17.6) | 81,771 |
| Chronic disease–related hospitalization | 6.6 (5.1, 8.5) | 62 | 4.8 (4.7, 4.9) | 22,582 |
| Mental illness–related hospitalization | 50.7 (46.2, 55.5) | 473 | 6.4 (6.3, 6.5) | 30,097 |
In all calculations for percentages, missing values were excluded from the calculation in both the numerator and denominator. NICU, neonatal intensive care unit.
Table 2.
IRRs (95% CIs) of infant hospital length of stay for interactions terms, for infants with neonatal withdrawal syndrome, Michigan, 2006–2009
| Unadjusted | Adjusteda | |
|---|---|---|
| Low birthweight | ||
| Yes | 1.13 (1.04, 1.23) | 1.36 (1.24, 1.49) |
| No | 5.27 (5.04, 5.52) | 5.75 (5.41, 6.10) |
| Maternal hospitalization related to mental illness | ||
| Yes | 2.73 (2.52, 2.95) | 4.66 (4.39, 4.96) |
| No | 4.21 (3.90, 4.55) | 5.75 (5.41, 6.10) |
Note: N=426,030. The referent is infants without neonatal withdrawal syndrome.
Adjusted for mother’s race (black, white, other); maternal hospitalization related to chronic disease; maternal hospitalization related to mental illness; Caesarean-section delivery; low birthweight; inadequate Kotelchuck index; and having only the mother acknowledged on birth certificate.
IRR, incidence rate ratio.
Discussion
Neonatal withdrawal syndrome infants had a significantly longer LOS than infants without withdrawal syndrome after adjusting for factors that could affect LOS, consistent with previous literature documenting a higher LOS among infants with withdrawal syndrome,4,5,7 after adjusting for LBW or prematurity.6 The current study is unique by quantifying the additional LOS of infants with withdrawal syndrome after adjusting for maternal factors in addition to infant factors. Given that infants with withdrawal syndrome are more likely to be covered by Medicaid and to require a significantly longer LOS, these findings suggest that the majority of the withdrawal syndrome financial burden falls on state and federal governmental funding streams.
Withdrawal syndrome birth rates increased significantly among Michigan infants from 2000 to 2009, consistent with other recent population-based studies.3–7 Among infants with a maternal mental illness hospitalization, the average LOS was almost five times longer for withdrawal syndrome infants than for infants without withdrawal syndrome. Withdrawal syndrome infants with a maternal mental health illness hospitalization may have a longer LOS because of the need for a safe discharge plan to be in place before leaving the hospital.18
Limitations
Study limitations relate to validity and completeness of vital statistics and hospital discharge data. Identifying infants that have withdrawal syndrome by ICD-9-CM coding alone underestimates the true birth rate.19 When infants are transferred from one hospital to another, hospitalizations are counted as separate visits, thus lowering LOS into two shorter stays, suggesting that the study LOS underestimates the true LOS for some infants with the syndrome.
Misclassification of iatrogenic withdrawal infants as those with withdrawal syndrome is possible, but it appears to be uncommon. A recent study by Patrick et al. found that iatrogenic cases comprised less than 5% of all withdrawal syndrome births in their nationally representative sample (2000–2009).7 Although the present study cannot identify maternal drugs used to treat opioid dependence, differentiation in pharmacologic treatment may influence infant’s postneonatal LOS. Although methadone is considered standard treatment for opioid dependence during pregnancy,20,21 a recent study found infants of mothers treated with buprenorphine had a significantly shorter hospitalization.22
Conclusion
This study examined the increased LOS of neonatal withdrawal syndrome infants compared to other infants and adds to the literature by providing an estimate that adjusts for available maternal factors. Additionally, the increasing withdrawal syndrome trend with infants who are more likely to be covered by Medicaid suggests a large public financial burden. The estimated hospitalization charges in Michigan for infants with the syndrome born between 2000 and 2009 totaled more than $35 million.
Raising provider awareness is necessary to promote appropriate evaluation and treatment. The use of universal screening questions, brief intervention, and referral to treatment23 may help target and decrease drug abuse among women of childbearing age. Further research linking prescription drug monitoring data with maternal and infant hospital discharge data could identify scheduled drugs prescribed during pregnancy.
Acknowledgments
This study was led by an appointee to the Applied Epidemiology Fellowship Program administered by the Council of State and Territorial Epidemiologists (CSTE) and funded by the CDC Cooperative Agreement 5U38HM000414. Live birth certificate records were provided by the Division for Vital Records and Health Statistics (DVRHS) of the Michigan Department of Community Health. The hospital discharge records were made available through an agreement with the Michigan Health and Hospital Association. Thu Le, from DVRHS, assisted with data linkages and data file preparation.
The findings and conclusions of this study are those of the authors and do not necessarily represent the official position of the CDC.
Footnotes
No financial disclosures were reported by the authors of this paper.
References
- 1.Campbell CI, Weisner C, Leresche L, et al. Age and gender trends in long-term opioid analgesic use for noncancer pain. Am J Public Health. 2010;100(12):2541–7. doi: 10.2105/AJPH.2009.180646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kellogg A, Rose CH, Harms RH, Watson WJ. Current trends in narcotic use in pregnancy and neonatal outcomes. Am J Obstet Gynecol. 2011;204(3):259.e1–e4. doi: 10.1016/j.ajog.2010.12.050. [DOI] [PubMed] [Google Scholar]
- 3.Hudak ML, Tan RC. Neonatal drug withdrawal. Pediatrics. 2012;129(2):e540–e560. doi: 10.1542/peds.2011-3212. [DOI] [PubMed] [Google Scholar]
- 4.O’Donnell M, Nassar N, Leonard H, et al. Increasing prevalence of neonatal withdrawal syndrome: population study of maternal factors and child protection involvement. Pediatrics. 2009;123(4):e614–e621. doi: 10.1542/peds.2008-2888. [DOI] [PubMed] [Google Scholar]
- 5.Pan IJ, Yi HY. Prevalence of hospitalized live births affected by alcohol and drugs and parturient women diagnosed with substance abuse at liveborn delivery: U.S 1999–2008. Matern Child Health J. 2013;17(4):667–76. doi: 10.1007/s10995-012-1046-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Creanga AA, Sabel JC, Ko JY, et al. Maternal drug use and its effect on neonates: a population-based study in Washington state. Obstet Gynecol. 2012;119(5):924–33. doi: 10.1097/AOG.0b013e31824ea276. [DOI] [PubMed] [Google Scholar]
- 7.Patrick SW, Schumacher RE, Benneyworth BD, Krans EE, McAllister JM, Davis MM. Neonatal abstinence syndrome and associated health care expenditures: U.S., 2000–2009. JAMA. 2012;307(18):1934–40. doi: 10.1001/jama.2012.3951. [DOI] [PubMed] [Google Scholar]
- 8.Wagner CL, Katikaneni LD, Cox TH, Ryan RM. The impact of prenatal drug exposure on the neonate. Obstet Gynecol Clin North Am. 1998;25(1):169–94. doi: 10.1016/s0889-8545(05)70364-8. [DOI] [PubMed] [Google Scholar]
- 9.Neonatal drug withdrawal. American Academy of Pediatrics Committee on Drugs. Pediatrics. 1998;101(6):1079–88. [PubMed] [Google Scholar]
- 10.Lainwala S, Brown ER, Weinschenk NP, Blackwell MT, Hagadorn JI. A retrospective study of length of hospital stay in infants treated for neonatal abstinence syndrome with methadone versus oral morphine preparations. Adv Neonatal Care. 2005;5(5):265–72. doi: 10.1016/j.adnc.2005.06.003. [DOI] [PubMed] [Google Scholar]
- 11.Baxter FR, Nerhood R, Chaffin D. Characterization of babies discharged from Cabell Huntington Hospital during the calendar year 2005 with the diagnoses of neonatal abstinence syndrome. W V Med JA. 2009;105(2):16–21. [PubMed] [Google Scholar]
- 12.Kotelchuck M. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health. 1994;84(9):1414–20. doi: 10.2105/ajph.84.9.1414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.ICD-9-CM: International Classification of Diseases, 9th revision; clinical modification. Los Angeles: Practice Management Information Corporation (PMIC); 2006. [Google Scholar]
- 14.Hospital stay for healthy term newborns. Pediatrics. 2004;113(5):1434–6. doi: 10.1542/peds.113.5.1434. [DOI] [PubMed] [Google Scholar]
- 15.Adams EK, Miller VP, Ernst C, Nishimura BK, Melvin C, Merritt R. Neonatal health care costs related to smoking during pregnancy. Health Econ. 2002;11(3):193–206. doi: 10.1002/hec.660. [DOI] [PubMed] [Google Scholar]
- 16.Bernstein HH, Spino C, Finch S, et al. Decision-making for postpartum discharge of 4300 mothers and their healthy infants: the Life Around Newborn Discharge study. Pediatrics. 2007;120(2):e391–e400. doi: 10.1542/peds.2006-3389. [DOI] [PubMed] [Google Scholar]
- 17.Weiss M, Ryan P, Lokken L, Nelson M. Length of stay after vaginal birth: sociodemographic and readiness-for-discharge factors. Birth. 2004;31(2):93–101. doi: 10.1111/j.0730-7659.2004.00286.x. [DOI] [PubMed] [Google Scholar]
- 18.Hospital stay for healthy term infants. Pediatrics. 2010;125(2):405–9. doi: 10.1542/peds.2009-3119. [DOI] [PubMed] [Google Scholar]
- 19.Burns L, Mattick RP. Using population data to examine the prevalence and correlates of neonatal abstinence syndrome. Drug Alcohol Rev. 2007;26(5):487–92. doi: 10.1080/09595230701494416. [DOI] [PubMed] [Google Scholar]
- 20.Ebner N, Rohrmeister K, Winklbaur B, et al. Management of neonatal abstinence syndrome in neonates born to opioid maintained women. Drug Alcohol Depend. 2007;87(2–3):131–8. doi: 10.1016/j.drugalcdep.2006.08.024. [DOI] [PubMed] [Google Scholar]
- 21.Kaltenbach K, Berghella V, Finnegan L. Opioid dependence during pregnancy. Effects and management. Obstet Gynecol Clin North Am. 1998;25(1):139–51. doi: 10.1016/s0889-8545(05)70362-4. [DOI] [PubMed] [Google Scholar]
- 22.Jones HE, Kaltenbach K, Heil SH, et al. Neonatal abstinence syndrome after methadone or buprenorphine exposure. N Engl J Med. 2010;363(24):2320–31. doi: 10.1056/NEJMoa1005359. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.ACOG Committee Opinion No. 422. at-risk drinking and illicit drug use: ethical issues in obstetric and gynecologic practice. Obstet Gynecol. 2008;112(6):1449–60. doi: 10.1097/AOG.0b013e318192499b. [DOI] [PubMed] [Google Scholar]