Latent subgroups of cognitive performance in lead- and manganese-exposed Uruguayan children: Examining behavioral signatures

Abstract

Objectives: Lead (Pb) and manganese (Mn) are confirmed neurotoxins but it is unclear to what extent low-level exposure produces a unique behavioral signature. The objective of this study was to investigate latent cognitive profiles among children (6-8 years) from Montevideo, Uruguay co-exposed to these metals.

Method: Among 345 children, blood Pb and hair Mn were measured using atomic absorption spectroscopy and ICP-MS, respectively. Sixteen measures, reflecting multiple domains of cognitive functioning were gathered: (1) three tests from Cambridge Neuropsychological Test Automated Battery (CANTAB): Intra-Extra Dimensional Shift (IED), Spatial Span (SSP) and Stockings of Cambridge (SOC), (2) ten tasks from Woodcock-Muñoz Achievement Battery, Revised (WM): Visual-Motor Integration, Verbal Comprehension (Vocabulary, Synonyms, Antonyms, Analogies), Visual-Auditory Comprehension, Concept Formation, Visual Spatial Thinking, Number Inversion and Spatial Relations, (3) Bender Gestalt task, and (4) Weschler block design task. Scores were modeled using latent profile analysis (LPA). Association between blood Pb and hair Mn on performance profiles was assessed using ordinal regression, controlling for confounders. An interaction between Pb and Mn was tested.

Results: Mean ± SD of blood Pb was 4.1 ± 2.1 μg/dL and 35% of children had blood Pb ≥ 5 μg/dL. Median [5%, 95%] hair Mn level was 0.8 [0.3, 4.1] ppb. Three latent cognitive performance profiles were identified: high (n = 46, 13%), average (n = 209, 61%) and low (n = 90, 26%). Each one-unit increase in blood Pb was associated with a 28% greater likelihood of belonging to a poorer-performing profile. The association was non-linear, with the effect of Pb on profile membership strongest at lower levels of exposure. There was no meaningful interaction between Pb and Mn.

Conclusions: A behavioral signature for low-level Pb & Mn exposure was not identified, but the likelihood of membership in low-performing profile was higher at lowest levels of blood Pb. There was no effect measure modification between Pb and Mn. Future research should address how complex environments created by chemical exposures and the social context relate to cognitive performance in young children.

Keywords: Behavioral signature; Childhood; Cognitive domains; Global health; Lead; Manganese.

Figure 1.. Standardized Test Means by Cognitive Profile (n=345).

Latent profiles are plotted with cognitive testing variables on the X-axis, and grouped according to CHC cognitive domains. All variables were standardized in Mplus to ease with interpretation across tests. Mean differences demonstrate a three tiered level of cognitive performance, generally with low, average and high performers. Note: log – log-transformed variable; SSP – Spatial Span Task; IED – Intra-dimensional Extra-dimensional Shift Task; SOC – Stockings of Cambridge Task; W-M – Woodcock-Muñoz; BGT – Bender-Gestalt Test; Gsm – Short term memory; Gf – Fluid Intelligence; Gv – Visual spatial processing; Gc – Crystalized Intelligence; Glr – Long-Term Retrieval; Ga – Auditory Processing; Gs – Processing Speed

Figure 2.

Model Predicted Probabilities of Membership in Each Cognitive Profile among 6–8 Year-Old Children.

Figure 2.

Model Predicted Probabilities of Membership in Each Cognitive Profile among 6–8 Year-Old Children.

Figure 3.

Model Predicted Probabilities of Membership in Each Cognitive Profile by High and Low Levels of Blood Pb and Hair Mn

Figure 3.

Model Predicted Probabilities of Membership in Each Cognitive Profile by High and Low Levels of Blood Pb and Hair Mn