Delayed brainstem auditory evoked potential latencies in 14-year-old children exposed to methylmercury

To determine possible exposure-associated delays in (BAEP) latencies as an measure of toxicity

toxicity can produce widespread adverse effects within the nervous system, especially when exposures occur during brain development. [2][3] Early adverse effects have been characterized by administering neurobehavioral tests to children exposed in utero from maternal seafood diets. [4][5][6] Thus, a National Research Council (NRC) committee 7 recently concluded that intrauterine MeHg exposure was the most critical and emphasized the findings from a prospective birth cohort study carried out in the Faroe Islands. 5 The damage to the developing nervous system is thought to be potentially irreversible. 7 The possibility also exists that exposure during postnatal development may induce brain lesions; clinical 2,8 and experimental 9 information suggests that such effects would tend to be more focal and particularly involve the sensory cortex and the granular layer of the cerebellum.  13,14 and in laboratory animals. 15 In contrast to neuropsychological test outcomes, this measure is thought to be independent of socioeconomic covariates. 16 As illustrated by environmental exposure to lead, EP abnormalities constituted important objective evidence on neurotoxic effects in children. 17 I n a n e auditory evoked potentials (BAEPs) at age 14 years. We previously showed that increased intrauterine MeHg exposures were associated with delayed peak III latencies at age 7 years. 5,18 We hypothesized that these delays would remain at age 14 and that BAEP latencies would also 5 be sensitive to MeHg from adolescent seafood diets.

Study Population and Follow-up
A cohort of 1,022 births was assembled in the Faroe Islands during a 21-month period of 1986-1987. 19,20 The primary indicator of intrauterine exposure to MeHg was the mercury concentration in cord blood, and concentrations in maternal hair at parturition were also determined. 19 MeHg exposures varied considerably: 15% of the mothers had hair mercury concentrations above 10 g/g, while 4% were below 1 g/g, a level that corresponds to the exposure limit recommended by the NRC committee. 7 Concomitant exposure to polychlorinated biphenyls (PCBs) was determined from the concentration in umbilical cords from 438 cohort members. 5 The first follow-up examination was carried out seven years later and included hair-mercury assessment, evoked potentials and pediatric examination. 5 Hair samples were again obtained, and the proximal 2-cm segment was analyzed by flow-injection cold-vapor atomic absorption spectrometry after digestion of the hair sample in a microwave oven. 5 The total analytical imprecision for this analysis was estimated to be 4.3% and 5.5% at mercury concentrations of 4.7 g/g and 11.1 g/g, respectively. Accuracy was 6 ensured by participation in the Canadian Hair Mercury Quality Control Program; all our results were within one SD of the adjusted mean. The high analytical quality is comparable to previous performance. 5,19 Results in g may be converted to nmol by multiplying by 5.0.
The study protocol was approved by the ethical review committee for the Faroe Islands and the institutional review board at the U.S. institution, and parental informed consent was obtained.

Neurological examination
A thorough pediatric examination included otoscopy and assessment of neurological optimality. We used a four-channel electromyograph (Medelec Sapphire-4ME) also employed previously. 5,21 Click signals at an intensity of 65 dB HL (0.1 ms impulses of alternating polarity) were presented to the right ear through shielded ear phones at 20 Hz and 40 Hz (sampling time, 0.01 ms); the other ear was masked with white noise at an intensity of 45 dB HL. A frequency of 50 Hz was also attempted, but peak I was poorly defined at this click rate. EPs were recorded using three standard EEG electrodes placed on the vertex, the right mastoid ipsilateral to stimulation and the left mastoid (ground). While 1,024 responses were used seven years before, 5,21 the number was increased to 2,048 to improve the definition of peak I. Amplification and filtration were unchanged, and one replication of each condition was again carried out for calculation of average peak latencies. Peaks I, III, and V are thought to reflect the volume-conducted electric activity from the acoustic nerve, pons (superior olivary nucleus), and midbrain (inferior colliculi), respectively. 16

A u d i o m e t r y w a s c a r r i e d o u t b y a t r a i n e d n u r s e u s i n g I n t e r a c o u s t i c s D i a g n o s t i c
Audiometer AD229 with a Peltor H7A headphone in a sound-insulated room. The patient-controlled Hughson-Westlake procedure was used in accordance with ISO 8253-1. A threshold was defined as two out of three correct responses in a procedure with 5 dB increases and 10 dB decreases. Pure-tone air-conduction hearing thresholds were measured at 125, 250, 500, 750, 1000,1500, 2000, 3000, 4000, 6000, and 8000 Hz. Two children did not complete their audiometry examination.

Data analysis
Pearson's correlation coefficients were used to assess bivariate relationships between exposure parameters. Regression analysis was used to determine the association of MeHg exposure with the outcome variables. Age and sex may be important predictors of BAEP latencies 16,21 and were therefore included as independent variables along with the exposure parameters. In addition, confounders previously included in the analysis of neuropsychological test results 5 were screened for possible associations with the outcomes in the present study, but no pattern was found. Further models included as an independent variable the latency result obtained 7 years previously along with the age at that examination. Additional analyses also incorporated PCB dose-response curves at low dose levels and for determining exposure limits. 7, 18 The BMD is the dose of a substance that increases the risk of an abnormal response by a benchmark response (BMR), i.e., from P 0 (usually 5%) for an unexposed child to P 0 + BMR for a child exposed at the BMD. 23 The NRC committee used a BMR of 5% so that an exposure at the corresponding BMD will double the risk of an abnormal response. 7 To take the statistical uncertainty into account, a lower 95% confidence limit (BMDL) for the BMD is also determined. Using linear dose-response models, BMDLs expressed as the maternal hair mercury concentration were about 10 g/g for the most sensitive neuropsychological and BAEP outcomes in the Faroese children at age 7 years. 7,18,24 For comparison with these dose-response associations, we used the same default settings when calculating BMDL results for BAEP outcomes at age 14 years.

Prolonged Peak III and Peak V Latencies at Higher Prenatal MeHg Exposures Were Due to
Increased I-III Intervals That Were Prolonged Already 7 Years Before exposure had increased since the previous examination (p < 0.001). Approximately half of the children now exceeded the hair-mercury limit of 1 g/g, but the average corresponded to only one-fourth of the concentrations in maternal hair at child birth. Nonetheless, the different sets of exposure biomarkers correlated well.
T h e B A E P l a t e n c i e s w e r e s i m i l a r t o t h e r e s u l t s o b t a i n e d a t a g e 7 , 5,18,21 and again differed as expected 16 between boys and girls. Age had no effect within the limited range studied.
I n t r a u t e r i n e M e H g e x p o s u r e b i o m a r k e r s showed several statistically significant associations with the BAEP latencies, especially peaks III and V at both frequencies (Table II).
The same tendency was seen for the interpeak I-III latency, despite being affected by the greater imprecision of peak I determinations. Because peak I and interpeak III-V latencies were clearly not associated with the intrauterine exposure level, MeHg appeared to affect mainly the I-III interval. Neither sex nor age was associated with MeHg exposure levels, and confounder adjustment therefore did not affect the mercury regression coefficients.  (Table II, Fig 2). This association was not affected by inclusion of prenatal exposure biomarkers, and neither did the lower mercury concentrations at age 7 seem to affect this outcome parameter. At the same time, this interpeak variable was significantly associated with all other peak latencies, except for the peak I latency.  (Table III). The association with the peak III latency (Table III) (Table II)  BMDLs averaged about 5 g/g for the child's hair-mercury concentration at age 14 years.

DISCUSSION
The developing brain is thought to constitute the most vulnerable organ in regard to MeHg exposure. 1,7 Emphasis in risk assessment has therefore been placed on neurological functions of children with intrauterine exposure to this neurotoxicant, and previous studies have applied neuropsychological function as a key measure of adverse effects. [4][5][6] In parallel, neurophysiological tests, such as BAEP assessment, have found use in population studies as highly standardized, rapid, painless, and inexpensive procedures. 16 a n t d e t e r m i n a n t s c a n n o t be controlled a priori. However, the present study involved a large birth cohort that has been followed prospectively for 14 years and characterized in substantial detail with regard to developmental MeHg exposure levels. The participation rate at age 14 years was very high, thereby reducing the concern that the results may have been affected by differential follow-up rates. An important strength of this study is that the examinations relied on the same methodology as 7 years before, and the same examiner, who was blinded in regard to exposure data and prior peak latency results. The validity of the results was supported by extensive quality assurance data. In addition, the outcome measures were confirmed to be independent of socioeconomic confounders. The known 16 BAEP peak latency difference between boys and girls was replicated, but sex was not associated with MeHg exposure and therefore did not cause confounding. At age 7 years, 21  to PCBs, which occur in whale blubber sometimes eaten in the Faroes, did not influence the BAEP outcomes. Developmental exposure to PCBs is now thought to affect primarily cochlear function and impact on BAEP amplitudes rather than latencies. 32 In addition, the lead exposures were comparatively low and not associated with mercury. 19 The generalizability of this study would therefore not seem to be limited by concomitant exposures to other neurotoxicants. indicate that recent MeHg exposure as assessed at age 14 years is associated with EP delays that differ from those incurred from exposure in utero. The possibility that peak latencies may distinguish between effects incurred prenatally and postnatally deserves attention in future studies.
The potential postnatal vulnerability of the brain would mean that children ought to be protected against MeHg exposure to the same extent as pregnant women.
We are grateful to the cohort families for their loyal support, to the highly competent clinical staff in Tórshavn, and to Dr David A Otto for advice regarding the quality assurance for the BAEP measurements.