Timeline

Dec 1987

CaNa2EDTA chelation doubles brain lead in rats — blood lead drops but brain lead rises

Cory-Slechta et al. show that standard CaNa2EDTA chelation mobilizes lead from bone and kidney but redistributes it to the brain. A single injection doubles brain lead concentration despite lowering blood lead. No net brain lead reduction after a full 5-day course. This finding drives the shift to DMSA for pediatric chelation. View source ↑

Oct 1988

DMSA at 30 mg/kg/day cuts blood lead 78% in 5 days — outperforms CaNa2EDTA

Graziano et al. test three DMSA doses in 21 children with BLL 31-49. The highest dose (1050 mg/m2/day) reduces blood lead by 78% vs. 47% for IV CaNa2EDTA, with no adverse mineral excretion. This dose becomes the basis for the TLC protocol. Blood lead rebounds in all groups after treatment stops. View source ↑

May 1989

Only 20% of oral DMSA reaches systemic circulation — 88% biotransformed to disulfide forms

Aposhian et al. give a single DMSA dose to 6 healthy men. Only 20.6% is recovered in urine; 88% of that is in altered disulfide form. DMSA selectively chelates lead, copper, and zinc without affecting iron, calcium, or 27 other elements. These data provide the pharmacokinetic rationale for choosing DMSA over CaNa2EDTA in children. View source ↑

Oct 1989

6 million U.S. children at risk from lead paint; 1.2 million above 15 ug/dL

Mushak and Crocetti summarize an ATSDR report to Congress quantifying the scale of childhood lead exposure. Approximately 6 million children under 7 live in high-risk housing. Blood lead distributions are log-normal, meaning even moderate population means mask large numbers of severely exposed children in the upper tail. View source ↑

Feb 1990

Meta-analysis: 11 of 12 studies show lead lowers IQ (P < 0.0001)

Needleman and Gatsonis aggregate 12 studies meeting strict inclusion criteria. Joint P-values below 0.0001 for blood lead studies and 0.0005 for tooth lead studies. The effect is robust to removal of any single study. Most individual studies were underpowered, but the combined signal is overwhelming. This meta-analysis becomes a cornerstone for the 1991 CDC guideline revision. View source ↑

Apr 1994

No safe threshold: IQ declines by 2.6 points per 10 ug/dL, steeper at lower levels

Schwartz pools 7 studies using inverse-variance weighting and finds a 2.57-point IQ loss from 10 to 20 ug/dL. Studies with lower mean BLL have steeper slopes, suggesting a logarithmic dose-response. Nonparametric smoothing of the Boston cohort shows no threshold down to 1 ug/dL. This paper provides the epidemiologic foundation for TLC's IQ endpoint. View source ↑

Jan 1995

Decision model projects chelation prevents 22.5% of reading disability — but 'no definitive data' exist

Glotzer et al. model chelation cost-effectiveness assuming a 70% reduction in reading disability risk. EDTA chelation costs $1,778/child but saves $900M/year when remedial education costs are included. The authors explicitly note 'no definitive data exist regarding the effect of chelation with respect to lead-induced neurotoxicity.' Succimer is excluded from the model as too new. The TLC trial is designed to fill this gap. View source ↑

Jul 1998 TLC

TLC design paper reveals population under extreme stress: parental IQ 80, 97% on public assistance

The TLC Group describes the trial's architecture: 780 children randomized across 4 cities, extended 26-day succimer courses, home cleaning, vitamin supplements. Baseline developmental scores already 1 SD below norms (MDI 82). Population is 77% Black, 72% single-parent, 97% on public assistance. Iron-deficient children were treated pre-enrollment, but ferritin data for the randomized cohort go unreported. View source ↑

Mar 1999 R&M

CLEARS trial: intensive biweekly cleaning cuts BLL 34% — single TLC cleaning insufficient

Rhoads et al. randomize 113 children in Jersey City to biweekly household cleaning vs. controls. Homes cleaned 20+ times show a 34% BLL reduction (-3.9 ug/dL). Education alone has no effect. The dose-response relationship (more cleanings = lower BLL) implies the TLC trial's single baseline cleaning was inadequate to control ongoing exposure. View source ↑

Jul 1999 TLC

TLC screening cohort: 'no iron deficiency' — but ferritin threshold set at <10, below all modern standards

Serwint et al. compare 787 TLC screening children to 222 NHANES III controls. Iron depletion rates appear similar (11% vs. 13%), but the threshold is <10 ug/L — below WHO (<12), clinical (<15), and neurodevelopmental (<20-25) standards. MCV is significantly lower and RDW significantly higher in TLC children, suggesting subclinical iron compromise the ferritin threshold misses. Only arithmetic means reported for a right-skewed distribution. View source ↑

Aug 1999 TLC

TLC's own vitamins were contaminated with lead: ~36 ug per tablet, 628 children exposed

Rogan et al. report that iron-wax particles in the trial's vitamin supplements contained 250-600 ppm lead, producing ~36 ug lead per tablet. 628 children were potentially exposed from April to July 1997. Investigators report a mean 1.2 ug/dL BLL increase with no dose-response, concluding the contamination was inconsequential. The irony: the trial designed to address lead poisoning inadvertently gave children an additional lead source through the iron supplement. View source ↑

Nov 2000 TLC

Safety paper: succimer cuts BLL by 4.5 ug/dL for 6 months but 83% need retreatment; unexplained trauma excess

The TLC Group reports succimer lowered BLL by 4.5 ug/dL over 6 months (peak difference ~11 ug/dL at 1 week, then immediate rebound). 83% of succimer children required a second course; 83% of those required a third. Physical evidence of trauma was found in 14.9% of succimer vs. 9.9% of placebo children; 5 succimer and 0 placebo children were hospitalized for trauma. The authors acknowledged 'TLC is too small a trial to establish safety. View source ↑

Apr 2001

Blood lead predicts only 4 of 19 cognitive tests; cumulative bone-derived dose predicts 14 of 19

Links et al. apply linear systems theory to 535 former lead workers, comparing four dose metrics. Cumulative tibial lead (AUC') predicts 14 of 19 cognitive tests cross-sectionally, versus only 4 for current blood lead. Blood lead has a 30-day half-life vs. 27 years for bone lead. The TLC trial relied entirely on blood lead as both the treatment target and outcome metric — a fundamentally inadequate surrogate for brain dose. View source ↑

May 10, 2001 TLC

NEJM: Succimer lowers blood lead but produces no cognitive benefit at 36 months

The landmark TLC result. 780 children randomized, 96% retention. Succimer lowered BLL by 4.5 ug/dL over 6 months but the adjusted IQ difference was -1.1 points (succimer lower, not significant). No improvement on any of 15 cognitive, behavioral, or neuropsychological measures. Succimer children grew 0.35 cm less. The authors conclude chelation is 'not indicated' below 45 ug/dL. This paper becomes the evidence base for removing chelation from guidelines. View source ↑

Jul 2002

Needleman: 'How little can we spend?' — the retreat from primary prevention that made TLC necessary

Writing in response to the Kennedy-Krieger case, Needleman argues that the abandonment of comprehensive lead abatement ($34B estimated cost vs. $62B in benefits) created the conditions requiring cheap-alternative studies like TLC. The question shifted from eliminating lead in housing to whether pharmacologically lowering blood lead could reverse damage already done. View source ↑

Oct 2002 TLC

Paradox: naturally falling BLL predicts IQ gains in placebo group — drug-induced BLL drops do not

Liu et al. reanalyze TLC data by BLL change rather than treatment assignment. At 36 months, placebo children gain 4.0 IQ points per 10 ug/dL of natural BLL decline. Succimer children show no such association at any time point. This implies either chelation removes lead from a different compartment than what matters for the brain, or that families who reduced environmental exposure also provided more cognitive stimulation. View source ↑

Dec 2002 TLC

After professional cleaning, 69% of TLC homes still failed EPA lead dust standards

Ettinger et al. report that 765 of 780 TLC homes were professionally cleaned. Floor dust lead dropped 32%, windowsills 66%. But 69% of homes above EPA clearance standards before cleaning remained above after cleaning. Effectiveness varied dramatically by site: Baltimore 54% floor reduction vs. Philadelphia 9%. Lead in paint was never measured — it was simply assumed present. TLC children continued facing significant ongoing exposure throughout the trial. View source ↑

Apr 17, 2003

NEJM: IQ drops 7.4 points in the first 10 ug/dL — proportionally greater harm at lower levels

Canfield et al. follow 172 Rochester children with serial blood lead and IQ measures. A semiparametric model shows 7.4-point IQ decline from 1 to 10 ug/dL, but only ~2.5 additional points from 10 to 30. The nonlinear dose-response means TLC children (mean BLL 26) were on the flat part of the curve, making treatment effects harder to detect. Published in the same NEJM issue as Rogan and Ware's editorial on TLC implications. View source ↑

Apr 17, 2003 TLC

Rogan & Ware editorial: 'Prevention is the only plausible strategy' — chelation effects are irreversible

In an NEJM Perspective accompanying Canfield's paper, the TLC PI and biostatistician explicitly state that TLC chelation 'had no beneficial effects on tests of cognition, behavior, or neuropsychological function.' They note that lead damage appears long-lasting and irreversible, and that there is no safety margin at existing exposure levels. This is the trial leadership's own public concession. View source ↑

Feb 2004 TLC

Succimer-treated children grew 0.43 cm less — and this is the only TLC paper to report baseline ferritin

Peterson et al. find succimer-treated children grew significantly less in height: -0.27 cm at 9 months, -0.43 cm at 34 months. The deficit emerged during treatment and persisted. Buried in Table 1 as a covariate are the only published baseline ferritin values for the randomized cohort: 28.5 and 28.1 ug/L, with SDs nearly equal to the means, confirming a heavily right-skewed distribution. No analysis of ferritin as an effect modifier was ever performed. View source ↑

Jul 2004 TLC

Age 7: still no cognitive benefit; attention/executive function significantly worse in succimer group

Dietrich et al. assess 647 TLC children at school entry on an expanded battery (WISC-III, NEPSY, Woodcock, BASC). No benefit on any measure. One statistically significant finding favors placebo: NEPSY Attention/Executive Functions (P = .045). Succimer children are now 1.17 cm shorter. A quarter of all children still have BLL above 10 at age 7. Concurrent BLL is significantly associated with lower IQ (beta = -0.42 per ug/dL, P < .001). View source ↑

May 2005 TLC

Concurrent BLL at age 7 drives IQ (-5.4 points per 10 ug/dL) — peak BLL at age 2 does not

Chen et al. reanalyze TLC data and find concurrent BLL at age 7 is the strongest predictor of IQ (-5.4 points per 10 ug/dL), while baseline BLL at age 2 is not significant (-1.1 points, P = NS). When both are in the model, only concurrent BLL matters. This implies TLC's transient chelation (effective ~10 months) could not improve IQ at ages 5-7 when blood lead had re-equilibrated and ongoing exposure continued. View source ↑

Apr 2006 TLC

Succimer group has higher systolic blood pressure 1-5 years after treatment — possible calcium depletion

Chen et al. find succimer-treated children have systolic BP 1.09 mmHg higher than placebo from 12-60 months follow-up (P < 0.05). Despite succimer lowering BLL for 9-10 months, no blood pressure benefit occurred. The authors speculate the BP increase may result from calcium excretion during chelation. Another potential adverse effect of a treatment that produced no cognitive benefit. View source ↑

Feb 2007

Animal evidence: succimer reverses some lead-induced cognitive deficits — but causes lasting harm without lead exposure

Stangle et al. test succimer in a 3x2 factorial design with 120 rats. Lead-exposed rats given succimer show significant improvement in learning and arousal regulation — the first evidence any chelator can reverse lead's cognitive effects. But succimer given to non-lead-exposed rats produces lasting cognitive impairment comparable to high lead exposure. The authors note TLC achieved only a 4.5 ug/dL BLL difference — far less than the animal protocol that showed benefit. View source ↑

Mar 2007 TLC

Lead directly damages behavior independent of IQ — concurrent BLL at age 7 drives externalizing problems

Chen et al. use path analysis on TLC data to separate direct vs. IQ-mediated effects of lead on behavior. At age 7, concurrent BLL has significant direct effects on teacher-rated externalizing problems, school problems, and behavioral symptoms that are not mediated through IQ. Blood lead at age 2 has no significant effect on age-7 behavior. 39% of TLC children have teacher-rated school problems in the at-risk or clinical range. View source ↑

May 2007 TLC

Only positive TLC finding: succimer improves postural balance and gait in 161 Cincinnati children

Bhattacharya et al. assess 161 Cincinnati TLC children on force platform and gait walkway at age 5+. Succimer group shows significantly less postural sway during dynamic bending (6.6%, P = 0.04), 19% less medio-lateral sway during obstacle crossing (P = 0.001), and lower gait forces during normal walking. This is the only TLC publication showing significant benefit from succimer on any neuromotor measure. Authors caution the findings do not support changing clinical practice. View source ↑

Aug 2009

Age 6 BLL predicts IQ 7 points better than age 2 — TLC may have targeted the wrong developmental window

Hornung, Lanphear, and Dietrich combine Cincinnati and Rochester cohorts (n=397). Using a novel ratio approach to avoid collinearity, they show BLL at ages 5-6 is far more strongly associated with IQ than early childhood BLL. Children whose 6-year BLL was 50% higher than their 2-year BLL lost 7.0 IQ points. Criminal arrest rates were 3.35x higher. TLC's chelation at ages 1-3 may have targeted the wrong window entirely. View source ↑

Sep 2014 TLC

TLC cohort mean IQ 86.7 at age 7; low mercury levels rule out confounding neurotoxicant

Wang et al. measure methylmercury in 613 TLC children at age 7 and find very low levels (geometric mean 0.56 ug/L). Higher mercury is paradoxically associated with higher IQ (+2.1 points per ug/L), likely reflecting fish consumption benefits. Mercury is excluded as a confounding neurotoxicant. The persistent mean IQ of 86.7 — 13 points below the population norm — confirms lasting cognitive harm from lead exposure despite chelation. View source ↑