Original Article
515
Neonatal Outcomes in Very Preterm Singleton Infants Conceived Using Assisted Reproductive Technologies Laura Chiarelli, MSc1 Lucia Mirea, PhD1,2 Junmin Yang, MSc1 Shoo K. Lee, MD, PhD1,3 Prakesh S. Shah, MD, MSc1,3; on behalf of the Canadian Neonatal Network 1 Maternal-Infant Care Research Centre, Mount Sinai Hospital, Toronto,
Ontario, Canada 2 Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada 3 Department of Paediatrics, University of Toronto, Toronto, Ontario, Canada
Address for correspondence Lucia Mirea, PhD, Maternal-Infant Care Research Centre, Mount Sinai Hospital, 700 University Avenue, Suite 8-500, Toronto, Ontario M5G 1X6, Canada (e-mail: [email protected]).
Abstract
Keywords
► ► ► ► ►
fertility mode of conception singleton gestation mortality morbidities of prematurity
Objective To compare neonatal mortality, severe morbidities and hospital length of stay in very preterm singleton infants conceived using assisted reproductive technologies (ARTs) or spontaneously (SP). Study Design Singleton infants born at 23 to 32 weeks gestation during 2010–2012 were retrospectively identified from the Canadian Neonatal Network database. A composite outcome indicating a mortality or severe morbidity (grade 3 intraventricular hemorrhage, periventricular leukomalacia, stage 3 retinopathy of prematurity, bronchopulmonary dysplasia, or stage 2 necrotizing enterocolitis) was compared between ART and SP infants using multivariable logistic regression. Length of stay was examined in multivariable time-to-event analyses adjusting for competing risk of mortality. Results Eligible subjects included 346 (4.4%) ART and 7,578 (95.6%) SP infants. ART mothers were older, with fewer single parents, higher rates of nulliparity, diabetes, hypertension, antenatal corticosteroids, and prenatal care, but less smoking and substance use than SP mothers. No significant differences were detected in the composite outcome (odds ratio: 0.79; 95% confidence interval: 0.54–1.17) or length of stay (hazard ratio: 0.84; 95% confidence interval: 0.63–1.12) between ART and SP infants after adjustment for potential confounders and risk factors. Conclusion Among singleton infants born very preterm, mode of conception is not associated with overall mortality/morbidity or length of stay.
Assisted reproductive technologies (ARTs) have become widely accepted therapies to facilitate conception and pregnancy, accounting for 1.7 to 4.0% of total births and a large proportion (18%) of multiple births in many industrialized countries including Canada.1,2 The increase in ART usage has been linked to rising rates of preterm birth and subsequent adverse perinatal outcomes.3,4 To reduce the burden of
perinatal morality and morbidities due to multiple births, public health initiatives are increasingly recommending single embryo transfers in ART treatments.5,6 However, even for singleton infants conceived using ARTs, there is evidence of higher preterm birth, low birth weight, small for gestational age births, and perinatal mortality compared with infants conceived spontaneously.3,4,7 Furthermore, conception using
received August 21, 2014 accepted after revision October 8, 2014 published online December 24, 2014
Copyright © 2015 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0034-1396699. ISSN 0735-1631.
Downloaded by: UPSTATE Medical University. Copyrighted material.
Am J Perinatol 2015;32:515–522.
Assisted Conception and Outcomes in Preterm Singleton Infants ARTs has been associated with higher admission rates to the neonatal intensive care unit (NICU)3,4 and longer duration of hospital stay.8,9 The majority of neonatal mortality and morbidities occur in infants who are born preterm and require care in a NICU; however, the evidence for association between ART conception and adverse neonatal outcomes among singleton infants born preterm or with low birth weight is limited with conflicting results.10–15 As single-embryo ART procedures are applied more widely, and with improved success, the number of singleton infants conceived using ART is expected to rise. Therefore, it is important to understand the burden of illness and the demands on health care resources due to ART conception, especially for infants who are born very preterm and require specialized care. The objectives of this study were to examine the association of mode of conception with (1) neonatal mortality and morbidities and (2) hospital length of stay, in singleton infants born at 23 to 32 weeks gestational age and admitted to a level-three NICU in Canada between 2010 and 2012.
Subjects and Methods This retrospective observational study examined a cohort of singleton infants born at 23 to 32 weeks gestational age and admitted to a NICU participating in the Canadian Neonatal Network (CNN) between January 2010 and December 2012. The CNN population coverage included all infants admitted to 30 level-three NICUs in Canada. At each CNN site, trained abstractors collected data for each infant throughout their NICU stay according to a standard protocol, with information from patient charts entered electronically into a data-entry program with built-in error checking.16 Approval was granted at each site by the local research ethics board or through an institutional quality improvement process. Study subjects excluded infants who were declared moribund on admission and for whom no aggressive treatment was provided due to imminent mortality. Furthermore, infants were excluded if they were born with severe congenital anomalies (life threatening or requiring immediate attention), or if data for mode of conception or gender was missing. Conception using ARTs included in vitro fertilization (IVF), intracytoplasmic sperm injection, embryo transfer, intrauterine insemination, and any other mechanical method in which both the oocytes and sperm are manipulated to facilitate conception and implantation. For each infant, a binary composite outcome was defined indicating mortality (all causes) during NICU stay, or any of the following severe neonatal morbidities: grade 3 intraventricular hemorrhage (IVH) or periventricular leukomalacia17,18; stage 3 retinopathy of prematurity (ROP)19; bronchopulmonary dysplasia (BPD)20; or stage 2 necrotizing enterocolitis (NEC).21,22 Length of stay was measured in days from the date of NICU admission to the date of hospital discharge or transfer. Possible destinations following NICU discharge included home, another area of the hospital, a separate community or tertiary hospital, transfer out of the country, or discharge due to mortality or for palliative care. American Journal of Perinatology
Vol. 32
No. 6/2015
Chiarelli et al.
Covariates available included (1) maternal factors: maternal age, nulliparity, smoking, substance use, and single parent (no one other than the mother is regularly involved in the social care of the child), (2) pregnancy/obstetrics factors: gestational diabetes, hypertension, use of antenatal corticosteroids, inborn/outborn (born in a hospital with a level-three NICU/transferred after birth to a level-three NICU), cesarean/vaginal delivery and any prenatal care (at least one visit before delivery); and (3) infant characteristics: gender, mild congenital anomalies (not the cause of admission to the NICU), gestational age (complete weeks), birth weight, small for gestational age (SGA) corresponding to birth weight below the 10th percentile for gestational age,23 the presence of a primary infection at birth, illness severity on admission to the NICU quantified by a score for neonatal acute physiology version II of 20,24 and sepsis (isolation of a pathogenic organism in symptomatic neonate from blood or cerebrospinal fluid after 3 days postnatal age). Gestational age was defined as the best estimate based on date of IVF, early ultrasound, last menstrual period, obstetric estimate, or pediatric estimate, in that hierarchical order.16 The distribution of each covariate was compared between ART and SP infants using the Pearson chi-square or Fisher exact test for categorical variables, and the Student ttest or Wilcoxon rank sum test for continuous measures. Similar univariate analyses compared the composite outcome, mortality and each morbidity component, and the destinations following NICU discharge, between the ART and SP conception groups. The association between the mode of conception and the composite outcome was examined using multivariable logistic regression. To account for potential correlation of the composite outcome among infants from the same NICU site, analyses employed generalized estimating equations assuming an independent working correlation structure. The association between the mode of conception and time-to-discharge home was assessed using competing risk analyses as described by Fine and Gray.25 The event of interest was discharge home, and infants discharged to another area of the hospital, a separate community or tertiary hospital, or transferred out of the county were considered right censored. Discharge due to mortality or for palliative care was considered a competing event. Model building for both logistic regression and competing risk analyses was performed using a self-administered forward selection process with a significance level cutoff of 0.05. A series of multivariable models were developed with adjustment for potential confounders and/or important risk factors, and considering possible intermediate factors between the mode of conception and neonatal mortality/morbidity. Birth weight was available but not included in multivariable models due to the high correlation with gestational age (Pearson correlation coefficient ¼ 0.79). Statistical analyses were performed using SAS V.9.3 (SAS Institute Inc., Cary, NC), and R V.2.15 software (R Core Team), including the “cmprsk” package.26 All statistical tests were two-tailed with significance evaluated at the 0.05 level.
Downloaded by: UPSTATE Medical University. Copyrighted material.
516
Assisted Conception and Outcomes in Preterm Singleton Infants
Among the total 12,691 infants born at 23 to 32 weeks gestational age between 2010 and 2012 identified from the CNN database, 7,924 (62.4%) singleton infants met the study criteria; details regarding infants excluded are provided in ►Fig. 1. The distribution of each covariate among 346 (4.4%) infants conceived using ART and 7,578 (95.6%) infants conceived spontaneously is presented in ►Table 1. Mothers of ART infants were of older age, fewer were single parents, with higher rates of nulliparity, diabetes, hypertension, use of antenatal corticosteroids, and any prenatal care, but lower rates of smoking, substance use, and outborn birth, than SP mothers (►Table 1). A borderline significant association was observed for SGA birth, with a higher rate in ART infants. The composite mortality/morbidity outcome was present in 102 (30%) ART and 2,449 (32%) SP infants. No statistically significant differences were detected in the distribution of the composite outcome, mortality or any morbidity between the ART and SP groups (►Table 2). Similarly, mode of conception was not associated with the composite mortality/morbidity outcome in multivariable logistic regression models adjusted for potential confounders and risk factors (►Table 3). The distribution of NICU discharge destinations varied significantly by mode of conception, with a larger proportion
517
of SP infants discharged home, whereas a larger proportion of ART infants were transferred to a community/tertiary hospital or out of the country (►Table 2). No association between mode of conception and time-to-discharge home was detected in competing risk multivariable analyses (►Table 3). The probability of discharge to home and the probability of discharge due to mortality/palliative care were similar among ART and SP infants (►Fig. 2A, B, respectively).
Discussion As ART treatments using single embryo transfer are expected to continue increasing, our study assessed the association between mode of conception and outcomes within the population of high-risk singleton infants who are born preterm and require acute care in a level-three NICU. In this contemporary Canadian cohort of singleton infants born between 23 and 32 weeks of gestational age and admitted to a level-three NICU, mode of conception was not associated with a composite outcome of neonatal mortality or morbidities, or length of stay. Our findings are consistent with previous results showing no association of IVF conception with mortality, NEC, IVH, cerebral morbidity, and BPD in singleton infants born with very low weight (1,500 g).10,15 In contrast to our null
Fig. 1 Infants excluded from the study sample.
American Journal of Perinatology
Vol. 32
No. 6/2015
Downloaded by: UPSTATE Medical University. Copyrighted material.
Results
Chiarelli et al.
Assisted Conception and Outcomes in Preterm Singleton Infants
Chiarelli et al.
Table 1 Distribution of maternal, pregnancy/obstetric, and infant characteristics among infants conceived using ART or spontaneously, born at 23 to 32 weeks gestational age from 2010 to 2012 Mode of conception, N/total (%)
Characteristic Infants, N (%)
ART
Spontaneous
346 (4.4)
7,578 (96.6)
p-Valuea
Maternal Maternal age, mean (SD)
34.3 (5.8)
29.9 (5.9)
515
Neonatal Outcomes in Very Preterm Singleton Infants Conceived Using Assisted Reproductive Technologies Laura Chiarelli, MSc1 Lucia Mirea, PhD1,2 Junmin Yang, MSc1 Shoo K. Lee, MD, PhD1,3 Prakesh S. Shah, MD, MSc1,3; on behalf of the Canadian Neonatal Network 1 Maternal-Infant Care Research Centre, Mount Sinai Hospital, Toronto,
Ontario, Canada 2 Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada 3 Department of Paediatrics, University of Toronto, Toronto, Ontario, Canada
Address for correspondence Lucia Mirea, PhD, Maternal-Infant Care Research Centre, Mount Sinai Hospital, 700 University Avenue, Suite 8-500, Toronto, Ontario M5G 1X6, Canada (e-mail: [email protected]).
Abstract
Keywords
► ► ► ► ►
fertility mode of conception singleton gestation mortality morbidities of prematurity
Objective To compare neonatal mortality, severe morbidities and hospital length of stay in very preterm singleton infants conceived using assisted reproductive technologies (ARTs) or spontaneously (SP). Study Design Singleton infants born at 23 to 32 weeks gestation during 2010–2012 were retrospectively identified from the Canadian Neonatal Network database. A composite outcome indicating a mortality or severe morbidity (grade 3 intraventricular hemorrhage, periventricular leukomalacia, stage 3 retinopathy of prematurity, bronchopulmonary dysplasia, or stage 2 necrotizing enterocolitis) was compared between ART and SP infants using multivariable logistic regression. Length of stay was examined in multivariable time-to-event analyses adjusting for competing risk of mortality. Results Eligible subjects included 346 (4.4%) ART and 7,578 (95.6%) SP infants. ART mothers were older, with fewer single parents, higher rates of nulliparity, diabetes, hypertension, antenatal corticosteroids, and prenatal care, but less smoking and substance use than SP mothers. No significant differences were detected in the composite outcome (odds ratio: 0.79; 95% confidence interval: 0.54–1.17) or length of stay (hazard ratio: 0.84; 95% confidence interval: 0.63–1.12) between ART and SP infants after adjustment for potential confounders and risk factors. Conclusion Among singleton infants born very preterm, mode of conception is not associated with overall mortality/morbidity or length of stay.
Assisted reproductive technologies (ARTs) have become widely accepted therapies to facilitate conception and pregnancy, accounting for 1.7 to 4.0% of total births and a large proportion (18%) of multiple births in many industrialized countries including Canada.1,2 The increase in ART usage has been linked to rising rates of preterm birth and subsequent adverse perinatal outcomes.3,4 To reduce the burden of
perinatal morality and morbidities due to multiple births, public health initiatives are increasingly recommending single embryo transfers in ART treatments.5,6 However, even for singleton infants conceived using ARTs, there is evidence of higher preterm birth, low birth weight, small for gestational age births, and perinatal mortality compared with infants conceived spontaneously.3,4,7 Furthermore, conception using
received August 21, 2014 accepted after revision October 8, 2014 published online December 24, 2014
Copyright © 2015 by Thieme Medical Publishers, Inc., 333 Seventh Avenue, New York, NY 10001, USA. Tel: +1(212) 584-4662.
DOI http://dx.doi.org/ 10.1055/s-0034-1396699. ISSN 0735-1631.
Downloaded by: UPSTATE Medical University. Copyrighted material.
Am J Perinatol 2015;32:515–522.
Assisted Conception and Outcomes in Preterm Singleton Infants ARTs has been associated with higher admission rates to the neonatal intensive care unit (NICU)3,4 and longer duration of hospital stay.8,9 The majority of neonatal mortality and morbidities occur in infants who are born preterm and require care in a NICU; however, the evidence for association between ART conception and adverse neonatal outcomes among singleton infants born preterm or with low birth weight is limited with conflicting results.10–15 As single-embryo ART procedures are applied more widely, and with improved success, the number of singleton infants conceived using ART is expected to rise. Therefore, it is important to understand the burden of illness and the demands on health care resources due to ART conception, especially for infants who are born very preterm and require specialized care. The objectives of this study were to examine the association of mode of conception with (1) neonatal mortality and morbidities and (2) hospital length of stay, in singleton infants born at 23 to 32 weeks gestational age and admitted to a level-three NICU in Canada between 2010 and 2012.
Subjects and Methods This retrospective observational study examined a cohort of singleton infants born at 23 to 32 weeks gestational age and admitted to a NICU participating in the Canadian Neonatal Network (CNN) between January 2010 and December 2012. The CNN population coverage included all infants admitted to 30 level-three NICUs in Canada. At each CNN site, trained abstractors collected data for each infant throughout their NICU stay according to a standard protocol, with information from patient charts entered electronically into a data-entry program with built-in error checking.16 Approval was granted at each site by the local research ethics board or through an institutional quality improvement process. Study subjects excluded infants who were declared moribund on admission and for whom no aggressive treatment was provided due to imminent mortality. Furthermore, infants were excluded if they were born with severe congenital anomalies (life threatening or requiring immediate attention), or if data for mode of conception or gender was missing. Conception using ARTs included in vitro fertilization (IVF), intracytoplasmic sperm injection, embryo transfer, intrauterine insemination, and any other mechanical method in which both the oocytes and sperm are manipulated to facilitate conception and implantation. For each infant, a binary composite outcome was defined indicating mortality (all causes) during NICU stay, or any of the following severe neonatal morbidities: grade 3 intraventricular hemorrhage (IVH) or periventricular leukomalacia17,18; stage 3 retinopathy of prematurity (ROP)19; bronchopulmonary dysplasia (BPD)20; or stage 2 necrotizing enterocolitis (NEC).21,22 Length of stay was measured in days from the date of NICU admission to the date of hospital discharge or transfer. Possible destinations following NICU discharge included home, another area of the hospital, a separate community or tertiary hospital, transfer out of the country, or discharge due to mortality or for palliative care. American Journal of Perinatology
Vol. 32
No. 6/2015
Chiarelli et al.
Covariates available included (1) maternal factors: maternal age, nulliparity, smoking, substance use, and single parent (no one other than the mother is regularly involved in the social care of the child), (2) pregnancy/obstetrics factors: gestational diabetes, hypertension, use of antenatal corticosteroids, inborn/outborn (born in a hospital with a level-three NICU/transferred after birth to a level-three NICU), cesarean/vaginal delivery and any prenatal care (at least one visit before delivery); and (3) infant characteristics: gender, mild congenital anomalies (not the cause of admission to the NICU), gestational age (complete weeks), birth weight, small for gestational age (SGA) corresponding to birth weight below the 10th percentile for gestational age,23 the presence of a primary infection at birth, illness severity on admission to the NICU quantified by a score for neonatal acute physiology version II of 20,24 and sepsis (isolation of a pathogenic organism in symptomatic neonate from blood or cerebrospinal fluid after 3 days postnatal age). Gestational age was defined as the best estimate based on date of IVF, early ultrasound, last menstrual period, obstetric estimate, or pediatric estimate, in that hierarchical order.16 The distribution of each covariate was compared between ART and SP infants using the Pearson chi-square or Fisher exact test for categorical variables, and the Student ttest or Wilcoxon rank sum test for continuous measures. Similar univariate analyses compared the composite outcome, mortality and each morbidity component, and the destinations following NICU discharge, between the ART and SP conception groups. The association between the mode of conception and the composite outcome was examined using multivariable logistic regression. To account for potential correlation of the composite outcome among infants from the same NICU site, analyses employed generalized estimating equations assuming an independent working correlation structure. The association between the mode of conception and time-to-discharge home was assessed using competing risk analyses as described by Fine and Gray.25 The event of interest was discharge home, and infants discharged to another area of the hospital, a separate community or tertiary hospital, or transferred out of the county were considered right censored. Discharge due to mortality or for palliative care was considered a competing event. Model building for both logistic regression and competing risk analyses was performed using a self-administered forward selection process with a significance level cutoff of 0.05. A series of multivariable models were developed with adjustment for potential confounders and/or important risk factors, and considering possible intermediate factors between the mode of conception and neonatal mortality/morbidity. Birth weight was available but not included in multivariable models due to the high correlation with gestational age (Pearson correlation coefficient ¼ 0.79). Statistical analyses were performed using SAS V.9.3 (SAS Institute Inc., Cary, NC), and R V.2.15 software (R Core Team), including the “cmprsk” package.26 All statistical tests were two-tailed with significance evaluated at the 0.05 level.
Downloaded by: UPSTATE Medical University. Copyrighted material.
516
Assisted Conception and Outcomes in Preterm Singleton Infants
Among the total 12,691 infants born at 23 to 32 weeks gestational age between 2010 and 2012 identified from the CNN database, 7,924 (62.4%) singleton infants met the study criteria; details regarding infants excluded are provided in ►Fig. 1. The distribution of each covariate among 346 (4.4%) infants conceived using ART and 7,578 (95.6%) infants conceived spontaneously is presented in ►Table 1. Mothers of ART infants were of older age, fewer were single parents, with higher rates of nulliparity, diabetes, hypertension, use of antenatal corticosteroids, and any prenatal care, but lower rates of smoking, substance use, and outborn birth, than SP mothers (►Table 1). A borderline significant association was observed for SGA birth, with a higher rate in ART infants. The composite mortality/morbidity outcome was present in 102 (30%) ART and 2,449 (32%) SP infants. No statistically significant differences were detected in the distribution of the composite outcome, mortality or any morbidity between the ART and SP groups (►Table 2). Similarly, mode of conception was not associated with the composite mortality/morbidity outcome in multivariable logistic regression models adjusted for potential confounders and risk factors (►Table 3). The distribution of NICU discharge destinations varied significantly by mode of conception, with a larger proportion
517
of SP infants discharged home, whereas a larger proportion of ART infants were transferred to a community/tertiary hospital or out of the country (►Table 2). No association between mode of conception and time-to-discharge home was detected in competing risk multivariable analyses (►Table 3). The probability of discharge to home and the probability of discharge due to mortality/palliative care were similar among ART and SP infants (►Fig. 2A, B, respectively).
Discussion As ART treatments using single embryo transfer are expected to continue increasing, our study assessed the association between mode of conception and outcomes within the population of high-risk singleton infants who are born preterm and require acute care in a level-three NICU. In this contemporary Canadian cohort of singleton infants born between 23 and 32 weeks of gestational age and admitted to a level-three NICU, mode of conception was not associated with a composite outcome of neonatal mortality or morbidities, or length of stay. Our findings are consistent with previous results showing no association of IVF conception with mortality, NEC, IVH, cerebral morbidity, and BPD in singleton infants born with very low weight (1,500 g).10,15 In contrast to our null
Fig. 1 Infants excluded from the study sample.
American Journal of Perinatology
Vol. 32
No. 6/2015
Downloaded by: UPSTATE Medical University. Copyrighted material.
Results
Chiarelli et al.
Assisted Conception and Outcomes in Preterm Singleton Infants
Chiarelli et al.
Table 1 Distribution of maternal, pregnancy/obstetric, and infant characteristics among infants conceived using ART or spontaneously, born at 23 to 32 weeks gestational age from 2010 to 2012 Mode of conception, N/total (%)
Characteristic Infants, N (%)
ART
Spontaneous
346 (4.4)
7,578 (96.6)
p-Valuea
Maternal Maternal age, mean (SD)
34.3 (5.8)
29.9 (5.9)
