Environmental and Perinatal Risk Factors of Cerebral Palsy: A California Birth Cohort Study
Date of Award
Spring 2024
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Public Health
First Advisor
Liew, Zeyan
Abstract
Cerebral palsy (CP) is the most prevalent neuromotor disorder that has clinical onset from childhood and persists for a lifelong period with no cure available. While most CP cases congenitally emerged before birth, their etiology is complex, multifactorial, and remained largely unknown. The documented adverse neonatal birth outcomes (e.g., birth asphyxia, preterm delivery) are strongly associated with CP but only occur in a small number of cases. Further, adverse birth outcomes are themselves the results of pathological processes during pregnancy and may be induced by preceding events like environmental exposures. There is a growing literature regarding the environmental factors on CP risk; nevertheless, evidence is scarce, especially from large population studies. Importantly, residual confounding bias has been extensively discussed but not rigorously examined in previous literature regarding perinatal risk factors of CP. In this context, my dissertation aimed to investigate the hypothesized association between CP risk and environmental exposures through a general exposure indicator, the season of conception, and a specific class of exposure, ambient air toxic contaminant. Furthermore, I developed the first population-based sibling comparison study for CP in the U.S. and I used such design to test the influence of residual confounding in the estimated effects of several established or suspected perinatal risk factors of CP. In the first aim, I worked with mentors to create a statewide birth cohort of over 4 million live births in California from 2007 to 2015 with 4,697 identified CP cases. The CP cases were ascertained using diagnostic records (up to 2021) from the California Department of Developmental Services. Using the season of conception as an indicator for seasonally varying environmental factors, we observed that children conceived in winter (January to March) or spring (April to June) were associated with an increased risk of CP (winter: RR=1.09, 95% CI: 1.01-1.19; spring: RR=1.10, 95% CI:1.02-1.20) compared with summer (July to September) conceptions. The associations were slightly stronger for mothers who lived in neighborhoods with a high social vulnerability and among spastic CP subtype, but no child sex differences were observed. The overall association was only minimally mediated through preterm delivery. Together, these findings suggested that seasonally varying environmental factors should be considered in the etiological research of CP. In the second aim, we investigated the environmental risk factors of CP by studying specific exposure to ambient toxic air contaminants (TACs) during pregnancy. In California, over 77 types of TACs have been routinely measured by air monitoring stations for every 12 days since 1990. We conducted a case-cohort study that includes all CP cases (N=906) and a 20% random sample of singleton births (N=184,954) who lived within a 5-mile radius around monitoring stations in California from 2005 to 2015. We focused on 14 volatile organic compounds (VOCs) and six metals with suspected neurotoxicity or developmental toxicity. We estimated the individual-level prenatal exposure to TACs by assigning monitored data based on the maternal residential address at delivery. For per interquartile-range increase of individual TAC exposures over the entire pregnancy, we observed a 3-18% increased risk of CP with six VOCs (benzene, toluene, 1,3-butadiene, acetone, acetonitrile, methylene chloride) and four metals (antimony, lead, nickel, vanadium) in the outdoor air. Stronger effects were observed for the mixture effects of all VOCs or all metals (VOCs, RR=1.24, 95% CI: 1.08-1.43; metals, RR=1.38, 95% CI: 1.20-1.58). To better evaluate uncontrolled confounding bias, we performed a negative control exposure analysis through the estimated effects of TAC exposures on CP risk in 36-48 months after delivery. We observed no associations in this negative control period, suggesting limited impact from uncontrolled time-invariant confounding. In the third aim, we developed the first sibling comparison design on CP and we used this design to test the influence of residual confounding on the association between CP and a range of a priori selected perinatal risk factors, including maternal prenatal lifestyle factors (pre-pregnancy BMI, cigarette smoking during pregnancy, gestational weight gain), pregnancy and obstetric complications (gestational preeclampsia, diabetes, infections), and neonatal adverse birth outcomes (preterm birth, low birth weight, low Apgar score). We focused on the entire live birth cohort defined in Aim 1, and further identified outcome-discordant siblings (1,213 CP cases and 1,544 non-CP sibling controls) born to the same biological mothers. We compared the effect estimates of the investigated perinatal risk factors and CP between the full cohort and discordant siblings. We observed robust associations between CP and most of the major neonatal birth outcomes and maternal labor complications we studied using the full cohort or the sibling comparison design. However, we found that residual confounding needs to be considered for certain maternal pregnancy complications, and the carryover effect might affect the findings for maternal lifestyle factors in the sibling design. Other assumptions, such as the generalizability of the sibling analysis to the full cohort, did not impact our findings. Overall, we demonstrated that sibling design can help enhance causal inference of the etiological research findings of CP if appropriately applied. In conclusion, my dissertation research established a statewide cohort of CP in California that is unique and valuable for CP studies. To my knowledge, no such efforts have been done in the US previously. Through my first two research aims, I found that seasonally varied environmental factors, such as maternal pregnancy exposure to toxic air contaminants in the outdoor air, may affect CP risk. My third aim found previously suspected perinatal risk factors of CP, notably neonatal birth outcomes, were robustly associated with CP using the sibling comparison design. The generated sibling sets and the statewide cohort by this dissertation are asset for CP research in the future. The collective findings from my studies offer insights into the CP etiology from environmental risk factors and strengthen the potential causal effects of perinatal risk factors.
Recommended Citation
Zhuo, Haoran, "Environmental and Perinatal Risk Factors of Cerebral Palsy: A California Birth Cohort Study" (2024). Yale Graduate School of Arts and Sciences Dissertations. 1463.
https://elischolar.library.yale.edu/gsas_dissertations/1463
