Treatment Endpoint

TLC was powered to detect a 3-point difference in IQ between treatment arms. The Design and Recruitment paper (1998) describes the sample size calculation:

“The study was designed to have 80% power to detect a 3-point difference in IQ scores between treatment groups, assuming a standard deviation of 15 points and a two-sided alpha of 0.05.”

This calculation assumed that a 10 µg/dL sustained difference in blood lead levels would produce a 3-point IQ difference. The 10 µg/dL assumption was based on observational studies available at the time, which estimated roughly 2–3 IQ points lost per 10 µg/dL increase in blood lead.

The assumed 10 µg/dL separation never occurred. Rogan et al. (2001) showed that blood lead levels in the two groups converged rapidly after the initial treatment effect:

A sustained 10 µg/dL separation appears to have occurred for only a few weeks at the beginning of treatment. By the time of the primary IQ assessment at 36 months, the groups were nearly identical.

The rapid convergence was driven by two factors: (1) rebound of blood lead levels in the succimer group after chelation stopped, and (2) natural decline in blood lead levels in the placebo group as children aged out of peak hand-to-mouth behavior.

Subsequent evidence has established that the relationship between blood lead and IQ is nonlinear, with the steepest slope at the lowest concentrations. Lanphear et al. (2005), in a pooled analysis of seven international studies, showed:

The dose-response curve is log-linear, not linear. This means that reducing blood lead from 26 to 16 µg/dL (the range TLC was operating in) would produce far less cognitive benefit than reducing from 10 to 5 µg/dL.

Using Lanphear's estimates, a sustained 10 µg/dL reduction from TLC's mean of 26 µg/dL to 16 µg/dL would predict only ~1.5 IQ points of benefit — well below the 3-point difference TLC was powered to detect.

The TLC protocol specified that children would receive additional courses of succimer if their blood lead level rebounded above 15 µg/dL after treatment. This threshold determined how aggressively the trial pursued sustained BLL reduction.

At the time of trial design (early 1990s), 15 µg/dL was considered the upper limit of “acceptable” blood lead levels by some clinicians, though the CDC had lowered the level of concern to 10 µg/dL in 1991.

A lower retreatment threshold — say, 10 µg/dL or even 5 µg/dL — would have:

• Triggered more retreatment courses
• Maintained a larger BLL separation between groups
• Produced a greater expected IQ difference (per the log-linear dose-response)

The choice of 15 µg/dL as the retreatment threshold effectively limited the trial's ability to test whether chelation to low levels improves cognition.

TLC was not designed to test whether chelation to low levels improves cognition. It tested whether chelation to 15 µg/dL improves cognition — a narrower question whose answer has limited relevance to current practice.

The current blood lead reference value is 3.5 µg/dL. A treatment endpoint of <10 µg/dL, or ideally <5 µg/dL, would have been more consistent with the dose-response relationship and more relevant to modern treatment decisions.

The trial's null finding cannot be interpreted as evidence that chelation does not work. It is evidence that chelation to 15 µg/dL, in the context of ongoing environmental exposure and rapid BLL convergence, did not produce a detectable IQ benefit at 36 months.

Whether chelation to low levels — in conjunction with environmental remediation — can preserve cognitive function remains an unanswered question.