Lead Less Toxic to the Well-Read?
Hardly a day goes by without some sensatinoal claims about what reading does for the mind, or brain. Some turn out to be true, but many skeptical. Here’s a recent news piece from HealthDay that circulated among national newspapers such as the Washington Post.
Lead was found to be 2.5 times more likely to have negative effects on the brains of adults with limited reading ability than on the brains of good readers, the researchers report in the July 31 issue ofNeurology.
How, you may ask, would reading prevent the toxicity of lead? It wouldn’t be because the moveable types were — until last century — made of lead, would it? No the answer given in the news piece is even sillier than that.
How might reading protect the brain? According to the researchers, an increased number of cortical synapses in larger brains might provide more brain capacity, the option to use alternative brain circuits if some are damaged, and the ability to process tasks more efficiently.
When something came out like this, one would immediately suspect the intelligence of the news reporter in interpreting academic publications. Sure enough, when we get to the source of the paper (subscription may be required), the authors mentioned the "brain size" as a potential theory in the 2nd paragraph:
There are multiple theories on how CR is protective against insults to the brain. These include an increased concentration of cortical synapses in larger brains that provide more brain reserve capacity,1 a greater ease of using alternative brain circuits,2 and the ability to process tasks more efficiently in presently used brain circuits.2
Their conclusion?
This finding would not support the passive brain reserve hypothesis that proposes a larger brain is able to tolerate more insult before clinical expression of impairment.
Couldn’t be clearer. Apparently the reporter didn’t bother to read beyond the 2nd paragraph.
Ok, then, what is the mechanism that the authors really think? They refer to something called "the cognitive reserve".
The theory of cognitive reserve (CR) is premised on the maintenance of cognitive performance in spite of ongoing underlying brain pathology.1,2 Factors which may contribute to CR include genetics,3 childhood cognitive ability,4 head size,1,5 education,4–6 linguistic ability,7 reading achievement,8,9 lifestyle,10,11 and occupational attainment.5,6,12 While the concept of CR evolved out of the Alzheimer disease (AD) literature,6,12–15 it is relevant in a variety of other dementias8,16 and in brain dysfunction associated with cerebrovascular disease,17,18 HIV-1 infection,19 sleep apnea,20 closed head injury,21 Parkinson disease (PD),22 multiple sclerosis (MS),23 temporal lobe epilepsy,24 and electroconvulsive therapy.25 Even those at risk for AD have increased use of CR during memory tasks as suggested by different activation patterns when compared to controls on functional MRI.26 The neural basis of CR has been demonstrated in a normal aging population27 using functional MRI28 and PET.29
Here’s when reading comes in, as a surrogate of "CR"
Surrogates of CR most commonly seen in the literature include years of education and measures of occupational attainment.6 Measures of reading achievement are considered better indicators of CR because they are less likely to be influenced by opportunity.8,9,30 Compared with years of education, reading achievement was shown to be a more accurate measure of educational experience and predictor of performance on neuropsychological testing in the elderly.9 A reading test identifies those self-educated individuals that left school early for a variety of sociocultural reasons and those individuals who graduated high school but are functionally illiterate.31
…
Cognitive reserve. The Wide Range Achievement Test-Revised for reading (WRAT)34 score provides a measure of reading achievement and premorbid intelligence,35,36 and is unlikely to be influenced by occupational neurotoxic insult. The WRAT, a measure of CR,30 is a word recognition test for 75 words. The words are of increasing complexity but only correct pronunciation not comprehension is required. When 10 consecutive words are mispronounced the test is terminated. Raw scores are then converted to an equivalent reading grade level. For the purpose of this study, a participant was classified as loCR if his performance on the WRAT was at a reading grade level of 11th grade or lower; hiCR participants had an equivalent reading grade level of 12th grade or higher.
The study is correlational by nature — you can’t randomly assign individuals to be good or bad readers, nor is it ethical to give one more dose of lead than others — but apparently it’s ethical to do so in an industry setting. So the Hi-CR and Lo-CR groups have many preexisting differences, in the psychological and motor measures (see Table 2). It would be foolish to simply compare psychological and motor performances between the two groups.
The authors did something much more sophisticated. Instead, they try to look for dose-response relations withing each group. In other words, they try to see if more (or less) lead exposure will decrease psychological and motor functions in individuals in the hi- or low-CR groups.
This makes sense, but things are nevery that simple.
As it turns out, the hi-CR group has a lot lower lead exposure (TWA) on average than the lo-CR group.
Visual inspection of the TWA distributions of the two CR groups showed a relative absence of loCR individuals with low TWAs and of hiCR individuals with high TWAs. This difference in the TWA distributions was reflected in the significant difference between loCR TWA (mean ± SD), 40.2 µg/dL ± 12.07 and hiCR TWA, 34.3 µg/dL ± 12.12 (p < 0.00).
I don’t know the clinical difference of 6 unit of TWA, but this is enough of a concern to the authors that they did something that is highly undesirable in a study like this — they attempted to match individuals from the two groups so that they’d have the same TWA.
Consequently, matched pairs were created using TWA with no more than 2 µ/dL difference between the pairs. For each member of the hiCR group the matching procedure consisted of the identification of all potential loCR matches. From this group of potential loCR matches one was randomly selected.
In other words, they dropped the majority of the 200 people in the lo-CR group and used only 56 that somehow matched the TWA of the hi-CR group. I would not put too much faith in the word "randomly selected" as they do not have much freedom in many cases, I suspect.
I will not critique problems of post-hoc matching in this case, as I don’t know how exactly things are done and the appropriateness of the subsequent analyses. Just one small surprise — the authors tried hard to claim that lead exposure affect only motor but not cognitive in the high-CR group. They apparently ignored the fact that the p-value of the dose-response analysis for the motor variables is only 0.09, NOT reaching statistical significant level as they stated (0.05, one tailed).
Anyhow, does reading protect you from lead exposure? You bet, but perhaps not in the way the author intended to convince you. By far the most non-contraversial finding here — one that makes most sense to anybody as long as you read — is that reading helps to land you a job that has less lead exposure.
So read, and become a manager.