Association between lead levels and adverse health effects: findings from the birth to twenty cohort

Abstract

Introduction Environmental lead exposure continues to be a health hazard, particularly in children. Globally, populations in low or middle-income countries are generally the most exposed to environmental lead, thus increasing their risk of lead-related detrimental health effects (Abadin et al., 2007a). However, such exposure in these populations remains underresearched. To try and address this deficit in evidence from developing countries like South Africa, the aim of this thesis was to examine possible environmental lead toxicological contributions to specific types of aggressive behaviour, violent behaviour and altered pubertal progression in adolescents. Methods The Birth to Twenty Plus (BT20+) cohort is the "largest and longest-running" longitudinal birth cohort in the whole of Africa (Richter et al., 2004). In the beginning, the study was called the Birth to Ten (BTIO) cohort ( 1990-2000). The main aim of BTIO was to study children's health and development; including "growth, well-being and education" in the first 10 years of their lives. After IO years the study was extended for another IO years and the cohort was subsequently renamed to Birth to Twenty (BT20) cohort (2001-20 I 0). The BT20 cohort continued with the examination of children's health and development but also included another component. It sought to answer targeted questions related to risks associated with "lifestyle including sexual and reproductive disorders, cardiovascular diseases and diabetes" (Richter et al., 2004, Richter et al., 2007). Post 2011, the cohort is now referred to as Birth to Twenty Plus (BT20+). The cohort is comprised of all singleton births to women residing in Soweto, Johannesburg area, South Africa over a specified 7-week period at the public health facilities. Children born from the 23rd of April to the 8th of June 1989 were recruited into the study. Of the 5449 births during this period, 3273 fulfilled the inclusion criteria and were included in the study. Inclusion criteria included that mother and child were to reside in Soweto, Johannesburg area for at least six months after birth. This was required in order to exclude mothers from outside Johannesburg who had only come to give birth in the city because of perceived better access to healthcare than where they actually lived or other family reasons. Prior to the commencement of the BTIO study, pilot studies had shown great resistance from the private healthcare sector to participate in the study. As such, only mothers who gave birth in public health facilities were included in the study. This resulted in great underrepresentation of White study participants in the cohort because, under Apartheid, private healthcare was mainly reserved for Whites. In order to compensate for this discrepancy, after IO years 120 White study participants born during the same period of study enrolment but not in the same area were recruited into the study (Richter et al., 2004, Richter et al., 2007). For additional information regarding the cohort please see Richter et al., 2004 & 2007. For the current study, blood lead analyses were performed using blood samples collected at birth and at age 13 years. Aggressive and violent behaviour was assessed using data collected using YSR questionnaires at ages 14/ 15 years and 15/ 16 years, respectively. Data for sex; socio-demographic factors at birth such as maternal education, maternal marital status, hospital of birth, place of birth, maternal age; household items; and anthropometric measurements at birth and 8 years old were collected using structured questionnaires. Principal Component Analysis (PCA) was used for I) data reduction 2) to determine the specific type(s) of aggressive behaviour associated with lead exposure during mid-adolescence and 3) to determine the specific type(s) of violent behaviour associated with lead exposure in late adolescence. Aggressive behaviour and violent behaviour are multi-faceted variables, as such, in order to understand the structure of aggressive behaviour and violent behaviour items PCA was selected to reduce these data sets. PCA is a data-reducing statistical procedure used to assess if a number of measures really describe a single variable. That is, to examine if different variables actually reflect a single underlying variable (Field, 2009). As such, to measure the dimensionality of aggressive behaviour and violent behaviour we used principal components derived from PCA. For this study PCA was chosen over Exploratory Factor Analysis (FA) because unlike FA which is more suitable for latent variables which cannot be directly measured; the individual aggressive behaviour items and violent behaviour items examined in this thesis can be directly measured. PCA is a linear combination of variables which reduces data while retaining as much of the variance of the observed variances as possible (Conway and Huffcutt, 2003, Field, 2009). Therefore, PCA was a more appropriate method to use to address the aims and objectives of this thesis. The PCA-derived components were used to distinguish the dimensionality of direct and indirect aggressive behaviour. Additionally, pubertal progression was assessed using pubertal data collected at age 9 through to 16 years and measured with the Tanner Sexual Maturation Scale. Data for breast and pubic hair development in girls and genital and pubic hair development were collected. Using Mplus, Latent Class Growth Analysis (LCGA) was carried out to group study participants into specific classes based on a "common developmental trajectory for the Tanner Sexual Maturation Scale indicators of pubertal stage'' (Lundeen et al., 2016). Results First, in the study to examine the link between lead exposure and aggressive behaviour the BT20+ participants with aggressive behaviour data at mid-adolescence (n = 1086) had mean blood lead levels of 5.6 ± 2.3 μg/dL and geometric mean blood lead concentration of 5.1 μg/dL at age 13 years. Fifty-nine per cent of study participants had blood lead levels _>5 μg/dl. Blood lead levels 2".5 μg/dL were positively associated with direct aggression (p=0.02). A higher proportion of males compared to females displayed a propensity towards direct aggression while females were shown to be more likely to perpetrate indirect aggression during mid-adolescence. Consequently, there was a positive association between the male sex and direct aggression but a negative association with indirect aggression (p<0.05). Second, in the study to examine the association between lead exposure and violent behaviour during late adolescence, study participants with data for violent behaviour during late adolescence (n = 1332) had mean blood lead levels of 5.8 ± 2.6 μg/dl. There was a statistically significant association between blood lead category ~ 5 μg/dL and physical violence and fighting. Blood lead levels at age 13 years were only associated with physical violence during late adolescence (p<0.0001). Furthermore, being male was associated with violence using a weapon, physical violence, fighting and robbing others (p < 0.05).

Third, in the BT20+ sub-sample with data for pubertal growth trajectory classes for public hair development and breast development in girls (n=732) and pubic hair development and genital development in boys (n=684), 49.5% of females and 75.0% of males had blood lead levels _>5 μg/dL at age 13 years. The mean blood lead levels were 6.55 ±2.6 μg/dL and 4.97 ± 1.9 μg/dL for boys and girls, respectively. In girls blood lead levels _>5 μg/dL relative to blood lead levels <5 μg/dL were associated with the slower pubertal transition for pubic hair development (p<0.0001) and breast development (p<0.05) from age 9 through to 16 years. When data were analyzed for the association between cord blood lead levels and pubertal progression, in boys, elevated cord blood lead levels (~ 5 μg/dl) were associated with a 72.0% decreased risk of being in trajectory class 3 compared to trajectory class 1 for pubic hair development (p< 0.05). Conclusion The statistically significant association between blood lead levels ~5 μg/dL and direct aggressive behaviour in mid-adolescence, and physical violence and fighting in late adolescence should be of great concern considering the reported high rates of violent behaviour among adolescents in South African schools in particular. Lead exposure was also associated with altered pubertal progression in both girls and boys. Healthy pubertal development is pivotal for young people. These study findings speak to how environmental lead exposure may significantly change the trajectory and future of young people; and the need for nationally driven public health policies to monitor blood lead levels in children from high-risk communities in particular.