Blinding & Placebo Comparability
Nine sources. Five different blinding mechanisms. Three silences. The TLC trial’s documentary record contains five mutually incompatible descriptions of how the trial blinded parents and clinic staff to treatment assignment, plus three published papers that assert “double-blind” without describing how blinding was achieved.
The only published papers that describe the mechanism at all (TLC Trial Group 2000, Dietrich et al. 2004) match just one of four protocol-internal descriptions — the one from Protocol v10 §4.6, a revision dated November 4, 1997, more than nine months after the last child was randomized. The amount of succimer in the blinding canister (200 mg) that appears in the published literature does not match the operative pre-randomization protocol (Protocol v9 §4.6, which specifies 100 mg).
The primary outcome paper (Rogan et al. 2001, NEJM) omits the blinding mechanism entirely. So does the 1998 design paper. The only acknowledgment that the blinding was partial appears in the 2000 Pediatric Research safety paper, in a sentence that also attributes 14% of succimer-arm attrition to the drug’s odor — a published admission that the blinding had observable behavioral consequences.
This page reads the sources in order, verbatim, with scholarly commentary. It ends with a cross-source contradiction map, open questions for records review, and a section on why this finding belongs near the center of the re-examination’s argument.
Sources, in order
- NIEHS RFP Amendment 01 to NIH-ES-92-31 — October 22, 1992, pre-contract
- Protocol v9 §4.6 Maintenance of Double-blind — August 23, 1994
- Protocol v9 §9.1.1 Repackaging of Trial Medications — August 23, 1994
- Protocol v10 §4.6 Maintenance of Double-blind — November 4, 1997
- Protocol v10 §9.1.1 Repackaging of Trial Medications — November 4, 1997
- TLC Trial Group 1998 design and recruitment paper — Paediatr Perinat Epidemiol
- TLC Trial Group 2000 safety and efficacy paper — Pediatric Research
- Rogan et al. 2001 primary outcome paper — NEJM
- Dietrich et al. 2004 7-year follow-up paper — Pediatrics
1. NIEHS RFP Amendment 01 to NIH-ES-92-31
Before any contract was awarded — and more than a year and a half before Protocol v9 was dated — NIEHS issued a written amendment to the TLC clinical-center RFP addressing questions raised by prospective offerors at the preproposal conference. The “BLINDING” subsection of that amendment is the earliest surviving document in which anyone responsible for the TLC trial publicly addressed how blinding should work.
How this document became public: The RFP Amendment was not voluntarily released by NIEHS. It surfaced through litigation discovery in Featherstone v. Kennedy Krieger Institute, Case 1:07-cv-01120-WMN (D. Md.), where it was filed as Exhibit D, Document 37-4 on September 20, 2007. The full original RFP has not been located; only this amendment has been made available, and only as a litigation exhibit fifteen years after it was issued.
Verbatim — BLINDING subsection
BLINDING
What justifies loss of blinding? What level of blinding is required?
Offerors must propose what they believe to be the best study design. Designs consistent with the basic trial include fully blinded designs with an escape, in which the study physician, the parent and child, and the psychometrician are all blind, to single blind designs, in which treatment assignment is random and the psychometrician is blinded but the child’s blood lead is managed openly. In general, the more open the design, the more likely it is that questions of bias will be raised about the results. Proposals that do not offer truly random treatment assignment and blind psychometric assessment are proposing in essence a different study than what was approved at NIEHS. The degree of blinding proposed by the offeror will be considered, among other things, in the evaluation of proposals, but the final decision will be made by the Steering Committee.
Can parents be made blind to treatment, if children given active drug smell like mercaptan?
Possibly not. Offerors could consider proposing to query parents about whether they knew if the child was getting active drug, and see if they get it right. NIEHS knows of no agent that smells like a mercaptan and could serve as a placebo.
2. Protocol v9 — §4.6 Maintenance of Double-blind
Protocol v9 is dated 8/23/94 in the running footer on every content page. This is the version in force during the entire randomization window (August 1994 – January 1997). A scanned copy with a September 21, 1994 title page also exists, but its content pages bear the same 8/23/94 footer — the September cover sheet is a court-discovery paperwork artifact, not a separate revision.
Verbatim — §4.6 Maintenance of Double-blind
Treatment will be blinded to the fullest extent possible. Both parents and clinic personnel will be blinded to the child’s PbB levels during treatment until six months after randomization.
Succimer emits a strong odor of sulfur, while the placebo for succimer emits a smell of alcohol. Therefore, it will not be possible to provide a fully comparable placebo. However, to provide a more sulfur-like smell to the placebo, a vented cylindrical plastic canister, 0.5 inches in diameter and 0.6 inches in length, will be filled with 100 mg. of succimer and added to all bottles of study drug (not just those containing succimer). The addition of the canister will change the odor of the placebo to one which is qualitatively similar to, but not as intense as, that of the active drug. Further, every effort will be made to avoid the need for any clinic personnel to open any subject’s medication bottle or otherwise deal directly with the study drug. The subjects taking succimer may themselves give off a strong odor; therefore, it may not be possible to blind clinic personnel entirely. For example, parents or caregivers may comment on the smell. Clinic personnel responsible for psychometric assessment, however, will not have contact with subjects or their caregivers during the treatment period.
As discussed above, local laboratory results will be reviewed by a physician who does not have direct subject or guardian contact during the treatment period. If a value is abnormal, the physician will order repeat testing. If a second abnormal value is obtained, the physician may recommend discontinuation of study drug. Blinding of treatment assignment will be maintained.
Mechanism described here: vented cylindrical plastic canister (0.5″ × 0.6″) filled with 100 mg of succimer, added symmetrically to every bottle — both active-drug and placebo — explicitly “not just those containing succimer.”
Admissions in the protocol’s own text: the protocol contains four separate acknowledgments that blinding could not be fully achieved: (i) “it will not be possible to provide a fully comparable placebo”; (ii) the canister odor is “qualitatively similar to, but not as intense as” the active drug; (iii) children taking succimer may themselves give off a strong odor; (iv) “parents or caregivers may comment on the smell.” None of these admissions appears in the 1998 design paper or the 2001 NEJM primary outcome paper.
3. Protocol v9 — §9.1.1 Repackaging of Trial Medications
The same protocol that describes a succimer-filled canister in §4.6 describes a completely different blinding mechanism in §9.1.1.
Verbatim — §9.1.1 (excerpt)
The Drug Distribution Center will repackage the medications in amber color glass unit-of-use containers, with child-resistant safety caps and a tamper-evident seal. The Drug Distribution Center will prepare two-thirds of the projected drug requirements before enrollment begins. When half of the projected total has been dispensed, the Drug Distribution Center will repackage the remaining one-third of the drug in proportions to be specified by the Data Coordinating Center based on actual trial experience.
In order to provide placebo with an odor comparable to that of succimer, the Drug Distribution Center will place a two cm² piece of filter paper which has been soaked with Mucomyst 20% solution into each bottle of placebo drug. An unsoaked piece of filter paper will be placed inside each bottle of succimer so that all bottles will appear the same.
Each bottle of drug will be assigned a unique number in random sequence at the Drug Distribution Center. Equal numbers of active and placebo drug will be placed in sequential order in shipping cartons. Each shipping carton will be assigned a unique identifying number. The Drug Distribution Center will provide to the Data Coordinating Center a database containing the bottle number, the carton number, and a code indicating whether active or placebo. This database will be used with a randomization algorithm different from that which was used at the Drug Distribution Center to further randomize drug assignments for trial participants.
Mechanism described here: a 2 cm² piece of filter paper soaked with Mucomyst 20% solution (Mucomyst is the brand name for N-acetylcysteine, a separate sulfur-containing compound with a sulfurous odor profile), placed in each bottle of placebo drug only. Bottles containing succimer receive an unsoaked (plain) piece of filter paper for visual parity.
This is a completely different mechanism from the one described in §4.6 of the same protocol. Different physical object (filter paper vs. plastic canister). Different chemical substance (Mucomyst vs. succimer). Different distribution pattern (placebo-only vs. symmetric across both arms). Protocol v9 is internally inconsistent on the blinding mechanism — the two sections cannot both be accurate descriptions of the same trial.
The Mucomyst mechanism appears in no other source in the TLC documentary record. The word “Mucomyst” appears zero times in Protocol v10, zero times in the 1998 design paper, zero times in the 2001 NEJM primary outcome paper, and zero times in any subsequent TLC publication. If this mechanism was ever implemented, the published literature has no trace of it. If it was abandoned in favor of the canister mechanism described in §4.6, there is no surviving amendment on the public record that documents the change.
4. Protocol v10 — §4.6 Maintenance of Double-blind
Protocol v10 is dated November 4, 1997. Randomization ended January 22, 1997. v10 is therefore a post-treatment revision of the protocol. It is the version NIEHS placed on its public website (dir.niehs.nih.gov/direb/tlc1/admin/prot_v10.html) and is the version the scientific community has treated as authoritative for understanding the trial’s design. The archived NIEHS HTML (captured by the Wayback Machine on October 30, 2004) matches the scanned PDF of Protocol v10 word-for-word at 99.73% similarity.
Verbatim — §4.6 Maintenance of Double-blind
Treatment will be blinded to the fullest extent possible. Both parents and clinic personnel will be blinded to the child’s PbB levels during treatment until six months after randomization.
Succimer emits a strong odor of sulfur, while the placebo for succimer emits a smell of alcohol. Therefore, it will not be possible to provide a fully comparable placebo. However, to provide a more sulfur-like smell to the placebo, a vented cylindrical plastic canister, 0.5 inches in diameter and 0.6 inches in length, will be filled with 200 mg of succimer and added to all bottles of study drug (not just those containing succimer). The addition of the canister will change the odor of the placebo to one which is qualitatively similar to, but not as intense as, that of the active drug. Further, every effort will be made to avoid the need for any clinic personnel to open any subject’s medication bottle or otherwise deal directly with the study drug. The subjects taking succimer may themselves give off a strong odor; therefore, it may not be possible to blind clinic personnel entirely. For example, parents or caregivers may comment on the smell. Clinic personnel responsible for psychometric assessment, however, will not have contact with subjects or their caregivers during the treatment period.
As discussed above, local laboratory results will be reviewed by a physician who does not have direct subject or guardian contact during the treatment period. If a value is abnormal, the physician will order repeat testing. If a second abnormal value is obtained, the physician may recommend discontinuation of study drug. Blinding of treatment assignment will be maintained.
What changed from v9 §4.6: exactly one number. 100 mg → 200 mg. Every other sentence is word-for-word identical to the v9 version, with only trivial editorial differences (a period after “100 mg.” in v9, dropped in v10; the word “the” before “psychometric assessment” in v9, dropped in v10). The canister dimensions (0.5″ × 0.6″), the symmetric distribution across both arms, the “(not just those containing succimer)” parenthetical, and all four admissions of partial unblinding are preserved verbatim from v9.
The dose doubled. Everything else stayed the same. This is either (a) a retrospective correction of v9’s 100 mg figure to reflect what was actually done during randomization, (b) a mid-trial dose change that should be documented in an IRB amendment, or (c) a post-hoc edit of the canister specification that does not reflect what was actually done. Nothing in the public record distinguishes these three possibilities.
5. Protocol v10 — §9.1.1 Repackaging of Trial Medications
If v10 §4.6 is a near-verbatim retention of v9 §4.6 with only the dose updated, v10 §9.1.1 is a complete rewrite. The Mucomyst filter paper mechanism that v9 §9.1.1 described is gone; in its place is a canister mechanism — but not the same canister mechanism §4.6 describes.
Verbatim — §9.1.1 (excerpt)
The Drug Distribution Center will repackage the medications in amber color glass unit-of-use containers, with child-resistant safety caps and a tamper-evident seal. The Drug Distribution Center will prepare two-thirds of the projected drug requirements before enrollment begins. When half of the projected total has been dispensed, the Drug Distribution Center will repackage the remaining one-third of the drug in proportions to be specified by the Data Coordinating Center based on actual trial experience.
In order to provide placebo with an odor comparable to that of succimer, the Drug Distribution Center will place a small canister containing 200 mg of active drug into each bottle of placebo drug. A canister containing 200 mg of placebo will be placed inside each bottle of succimer so that all bottles will appear the same.
Each bottle of drug will be assigned a unique number in random sequence at the Drug Distribution Center. Equal numbers of active and placebo drug will be placed in sequential order in shipping cartons.
Mechanism described here: swapped canisters. Placebo bottles receive a canister containing 200 mg of active drug (succimer). Succimer bottles receive a canister containing 200 mg of placebo. The canisters are asymmetric — different contents in each arm — so that all bottles appear the same from outside but the pharmacological contents are not duplicated in the succimer arm.
This contradicts §4.6 of the same protocol. v10 §4.6 says the canister is “filled with 200 mg of succimer and added to all bottles of study drug (not just those containing succimer)” — symmetric, every bottle contains a succimer canister. v10 §9.1.1 says succimer bottles contain a placebo canister, not a succimer canister — asymmetric, swapped contents. Both cannot be accurate descriptions of the same trial.
The simplest reading: v10 updated the operational Drug Distribution language in §9.1.1 but left the §4.6 overview text essentially untouched from v9 (only bumping the dose 100→200), creating an internal contradiction on what the canister in succimer bottles actually contained. Whichever description is correct, the other is an error of record on the face of the NIEHS-archived authoritative protocol.
Also: the word “Mucomyst” appears zero times in Protocol v10. “Filter paper” appears zero times. The mechanism described in v9 §9.1.1 has been silently removed from the authoritative public version of the protocol.
6. TLC Trial Group 1998 design and recruitment paper
This is the peer-reviewed paper whose explicit purpose is to document the TLC trial’s design and methods. It is the paper any subsequent researcher would cite to establish the trial’s methodological rigor. It is co-authored by the full TLC Trial Group.
Keyword frequency — blinding mechanism vocabulary
| mucomyst | 0 |
| canister | 0 |
| filter paper | 0 |
| smell | 0 |
| odor | 0 |
| sulfur | 0 |
| mercaptan | 0 |
| alcohol | 0 |
| mask / masking | 0 |
| blind (all occurrences) | 6 |
Every occurrence of “blind” in the paper is generic.
Verbatim — all six occurrences of “blind”
…a randomised, multicentre, placebo-controlled, double-blind clinical trial of the effects of treating lead-exposed children with succimer… (abstract)
The Treatment of Lead-exposed Children (TLC) Trial is a placebo-controlled, double-blind, randomised clinical trial… (methods, introduction)
…the data-coordinating centre assigned a study number corresponding to a blinded bottle of drug or placebo stored at the clinical site… (methods, randomization)
…the Bayley Scales of Infant Development-II (BSID2)…were performed on the same day as the first treatment visit by examiners blinded to treatment assignment. (methods, psychometric)
…An approximately equal number of children in the placebo group were selected at random for retreatment to retain blinding. Seventy-five per cent of children required a second course of drug… (methods, retreatment)
…Although clinical centres were blinded to blood lead levels during treatment, post-treatment levels were available to them. (methods, blood lead)
No sentence in the 1998 design paper describes the blinding mechanism. The trial is called “double-blind” in the abstract and methods. Bottles are called “blinded.” Examiners are called “blinded to treatment assignment.” Placebo patients are selected for retreatment “to retain blinding.” But there is no canister, no Mucomyst, no acknowledgment of the sulfur/alcohol smell disparity, no mention of partial unblinding, no unblinding-validation procedure. The protocol’s own four separate admissions that “it will not be possible to provide a fully comparable placebo” do not propagate into the public design paper.
CONSORT 1996 implication. The CONSORT 1996 guidelines — the applicable standard at the time of this paper — required reporting of “how blinding of those measuring the outcomes of interest and those giving the interventions was accomplished and its success evaluated.” This paper asserts blinding without documenting either the mechanism or any assessment of its success.
7. TLC Trial Group 2000 safety and efficacy paper
The 2000 Pediatric Research paper is the only publication in the core TLC literature that describes the blinding mechanism. It was published a year before the NEJM primary outcome paper, reports the safety profile and blood lead trajectories, and is the source the 2004 7-year follow-up paper cut-pasted its own blinding description from.
Verbatim — blinding mechanism passage
McNeil Consumer Products (Fort Washington, PA, U.S.A.) provided unmarked Chemet and placebo capsules of identical appearance. Because succimer has a strong, sulfurous, mercaptan odor, we packed 200 mg of succimer in a vented plastic cylinder in each bottle of placebo and succimer. Although this did not provide the placebo bottles with the room-filling odor of succimer, it did give the placebo an obvious aroma. The labels had a scratch-off area that identified the contents as succimer or placebo in case of medical emergency.
Verbatim — attrition-due-to-odor admission
…the families taking succimer had difficulty, with the additional 14% likely because of the odor of the drug.
This is the only published account of the blinding mechanism anywhere in the TLC core literature. It describes a canister (“vented plastic cylinder”) containing 200 mg of succimer placed in each bottle of placebo and succimer — symmetric distribution, both arms, same contents. This matches v10 §4.6 exclusively. It does not match v9 §4.6 (100 mg, not 200 mg), it does not match v9 §9.1.1 (Mucomyst filter paper, not canister), and it does not match v10 §9.1.1 (which specifies swapped canisters, with placebo in succimer bottles).
Two damaging admissions appear in this paper and nowhere else in the primary TLC literature. First: the placebo bottles “did not provide the room-filling odor of succimer” — the olfactory match was explicitly partial. Second: 14% of the succimer-arm adherence difficulty was attributed by the investigators themselves to the drug’s odor. That is a direct published acknowledgment that the compromised blinding had observable behavioral consequences, in the direction of differential adherence by arm.
Neither admission appears in the 2001 NEJM primary outcome paper. A reader encountering the TLC trial through the NEJM paper alone has no way to know the blinding was acknowledged to be partial or that odor-driven unblinding affected adherence.
8. Rogan et al. 2001 primary outcome paper
This is the primary outcome publication — the paper most often cited, relied on most heavily by subsequent clinical guidelines, and used to establish in the collective clinical understanding that chelation therapy did not improve cognition in children with blood lead levels of 20–44 µg/dL. It is the paper the AAP and CDC ACCLPP cite when they treat the TLC trial as authoritative.
Keyword frequency — blinding mechanism vocabulary
| mucomyst | 0 |
| canister | 0 |
| filter paper | 0 |
| smell | 0 |
| odor | 0 |
| sulfur | 0 |
| mercaptan | 0 |
| alcohol | 0 |
| mask / masking | 0 |
| blind (all occurrences) | 1 |
Verbatim — the single “blind” occurrence
…in a randomized, placebo-controlled, double-blind trial of up to three 26-day courses of treatment with succimer, a lead chelator that is administered orally…
The 2001 NEJM primary outcome paper contains no description whatsoever of how blinding was achieved. The trial is called “double-blind” once, in the opening methods summary. There is no canister, no Mucomyst, no acknowledgment that the placebo lacked the “room-filling odor of succimer,” no mention of the 14% attrition attributable to drug odor, and no description of any unblinding-validation procedure.
This silence is not an oversight driven by NEJM’s word limit alone. The 2000 Pediatric Research paper by the same TLC Trial Group had already published the blinding-mechanism description, including the “obvious aroma” admission and the 14% attrition estimate. Those facts were known to the authors at the time of the NEJM submission. The decision to omit them is a choice, not an accident.
This is the foundation of the clinical community’s understanding of the TLC trial as a double-blind RCT. Any defense of the trial’s methodological rigor that rests on citations to the NEJM paper must reconcile it with the protocol-internal contradictions above and the 2000 paper’s published admissions of partial unblinding.
9. Dietrich et al. 2004 7-year follow-up paper
The 7-year follow-up paper. Extends the primary outcome finding by showing the null effect persisted into elementary school. Cited alongside the 2001 NEJM paper as the long-term evidence against chelation for moderate BLLs.
Verbatim — blinding passage
…the Treatment of Lead-Exposed Children (TLC) study is a randomized, double-blind, placebo-controlled trial that was conducted between September 1994 and June 2003 in Philadelphia, Cincinnati, Newark, and Baltimore… [methods] …Because succimer has a strong, sulfurous, mercaptan odor, we packed 200 mg of succimer in a vented plastic cylinder in each bottle of placebo and succimer. Courses of therapy were 26 days…
Verbatim — discussion note on odor as an independent variable
…succimer has a prominent and unpleasant odor, and it is possible that administration of such a substance to small children 3 times daily over many weeks in early life could [affect cognition-behavior associations by a mechanism other than lead toxicity reduction]…
The 2004 paper repeats the 2000 paper’s blinding-mechanism description in substantially identical language. The phrase “Because succimer has a strong, sulfurous, mercaptan odor, we packed 200 mg of succimer in a vented plastic cylinder in each bottle of placebo and succimer” is lifted nearly verbatim from the 2000 Pediatric Research paper. This is not independent documentation — it is the same text reused. Across four years and two major publications, the TLC Trial Group has given one consistent account of the blinding mechanism, and that account is consistent with one (1) of the four protocol-internal descriptions.
The 2004 paper adds a deeper acknowledgment than the 2000 paper. In the discussion section, the authors speculate that administering “a substance” with a “prominent and unpleasant odor” to small children “3 times daily over many weeks” might itself affect the children’s later neuropsychological or behavioral trajectory. Read generously, this is the investigators openly entertaining the possibility that the drug’s smell was an exposure variable in its own right — a consideration that would complicate the interpretation of both the cognitive and behavioral outcomes, but that does not appear in the 1998 design paper or the 2001 NEJM paper.
Cross-source contradiction map
Which sources describe which component of the blinding mechanism. ✓ = source affirms the component; ✗ = source contradicts it; — = source is silent.
| v9 §4.6 | v9 §9.1.1 | v10 §4.6 | v10 §9.1.1 | 1998 design | 2001 NEJM | 2000 Ped Res | 2004 Dietrich | |
|---|---|---|---|---|---|---|---|---|
| Canister exists | ✓ | ✗ | ✓ | ✓ | — | — | ✓ | ✓ |
| Amount = 100 mg | ✓ | — | ✗ | ✗ | — | — | ✗ | ✗ |
| Amount = 200 mg | ✗ | — | ✓ | ✓ | — | — | ✓ | ✓ |
| Mucomyst used | ✗ | ✓ | ✗ | ✗ | — | — | ✗ | ✗ |
| Symmetric distribution (same contents in every bottle) | ✓ | ✗ | ✓ | ✗ | — | — | ✓ | ✓ |
| Swapped distribution (succimer in placebo bottles, placebo in succimer bottles) | ✗ | ✗ | ✗ | ✓ | — | — | ✗ | ✗ |
| Partial unblinding acknowledged in text | ✓ | — | ✓ | — | ✗ | ✗ | ✓ | ✓ |
| Parent-guess validation mentioned | ✗ | — | ✗ | — | ✗ | ✗ | ✗ | ✗ |
No row is unanimous. Every component of the blinding mechanism has at least one source that disagrees with the others.
CONSORT and ICMJE reporting implications
CONSORT 1996, the applicable reporting standard at the time of the 1998 design paper, required that randomized trials report “how blinding of those measuring the outcomes of interest and those giving the interventions was accomplished and its success evaluated.” The 1998 TLC design paper, the 2001 NEJM primary outcome paper, and the 2004 7-year follow-up paper each assert the trial was double-blind but none of them reports how blinding was accomplished in the primary methods text, and none reports whether its success was evaluated. The 2000 Pediatric Research paper alone describes the mechanism and acknowledges partial unblinding — but the 2000 paper is a safety/efficacy paper published between the design paper and the primary outcome paper, not the main paper of record, and its unblinding acknowledgment never propagates into the NEJM outcome paper or the subsequent clinical guidelines.
CONSORT 2010, the revision applicable at the time of the AAP 2016 and CDC ACCLPP 2012 guidance that cited TLC against chelation for moderate BLLs, requires explicit reporting of blinding methods (item 11a: “if done, who was blinded after assignment to interventions and how”) and similarity of interventions (item 11b: “if relevant, description of the similarity of interventions”). Re-evaluation of the TLC trial’s reporting quality against the CONSORT 2010 checklist would score the trial poorly on both items.
ICMJE requirements for describing methods in sufficient detail to allow replication are similarly not met by the primary outcome paper’s single generic “double-blind” reference.
Open questions for records review
Fact-questions the documentary record does not answer. IRB submissions, DSMB reports, Drug Distribution Center repackaging records, NIEHS contract files, or court discovery would be needed to settle them.
- Was the blinding mechanism actually Mucomyst-soaked filter paper, or was it a succimer-filled canister? What physical objects went into which bottles during the August 1994 – January 1997 randomization period? Drug Distribution Center records and the McNeil supply chain would resolve this definitively.
- If canisters, what was the amount — 100 mg or 200 mg? Was there a mid-trial change from 100 to 200 mg? If so, when, and under what amendment authority? IRB amendment logs at the four clinical centers should reflect any such change.
- Was the canister distribution symmetric or swapped? The two v10 descriptions cannot both be correct, and the operational answer should exist in Drug Distribution Center procedure documents.
- Did the investigators ever run the NIEHS-suggested parent-guess validation — querying parents on whether they knew what arm their child was in? If yes, where are the data, and why were they not published? If no, why not, given that NIEHS explicitly suggested the test in the pre-contract RFP Amendment?
- Why does the 2000 Pediatric Research paper describe the blinding mechanism and acknowledge partial unblinding, while the 2001 NEJM primary outcome paper omits both? Author-level documentation from the NEJM submission (drafts, reviewer comments, revisions) would indicate whether the omission was driven by word limits, editorial decision, or other factors.
- Does the 14% attrition-due-to-odor estimate in the 2000 paper have a primary source in the TLC internal data? If so, what was the specific question asked, who collected the data, and at what points during the trial?
- Why is the Mucomyst filter paper mechanism described in v9 §9.1.1 not mentioned in any published paper or any DSMB minutes known to be on the public record? If it was planned and then abandoned, the abandonment decision and its timing should be documentable. If it was actually implemented, the published literature is misreporting.
Relevance to the re-examination
This topic is a candidate for being the single most-cited argument in the TLC re-examination, for five reasons:
- The finding is evidentiary, not inferential. It does not require the reader to accept a contested statistical or epidemiological claim. It requires only that the reader look at five different source documents and notice that they say different things about the same physical procedure. Every claim is verifiable against the source text.
- The contradictions are about a core CONSORT-mandated methodological element — the blinding of a double-blind trial — and any defense of the 2001 NEJM paper’s conclusions must reconcile them.
- The 2000 Pediatric Research paper’s admissions are a direct published acknowledgment of partial unblinding, in the investigators’ own words. The re-examination does not need to allege unblinding; it only needs to quote the 2000 paper. The 14% attrition-due-to-odor estimate is, by itself, a published admission that the blinding had observable behavioral consequences in the treatment arm.
- The RFP Amendment establishes pre-contract knowledge of the problem. This makes the silence of the 2001 NEJM primary outcome paper harder to defend as an accident of word limits. The funder had explicitly told offerors that parental blinding was “possibly not” achievable and had suggested a specific validation (query parents on whether they could guess treatment assignment). Neither the validation nor the warning propagated into the primary publication.
- The v10 date (November 4, 1997, post-treatment) reframes the authoritative public protocol itself as a post-hoc revision. The NIEHS-hosted version of the protocol is dated ten months after the last child was randomized. Any defense of the trial’s design integrity that rests on “what the protocol said” must be explicit about which protocol, and when.
Source documents
- NIEHS RFP Amendment 01 to NIH-ES-92-31 (PDF) — dated October 22, 1992; surfaced as Exhibit D, Doc. 37-4 in Featherstone v. KKI, filed September 20, 2007
- Protocol v9 scan (PDF) — dated August 23, 1994
- Protocol v10 scan (PDF) — dated November 4, 1997
- Protocol v10 archived NIEHS HTML — Wayback snapshot October 30, 2004, matches PDF at 99.73% word similarity
- TLC Trial Group 1998 design paper (PDF) — Paediatr Perinat Epidemiol 1998;12(3):313–333
- TLC Trial Group 2000 safety/efficacy paper (PDF) — Pediatric Research 2000;48(5):593–599
- Rogan et al. 2001 NEJM primary outcome paper (PDF) — N Engl J Med 2001;344(19):1421–1426
- Dietrich et al. 2004 Pediatrics 7-year follow-up (PDF) — Pediatrics 2004;114(1):19–26
What this establishes. NIEHS knew, before the contract was awarded, that parental blinding might be impossible because of succimer’s mercaptan smell. The funder told prospective offerors that no existing pharmaceutical agent could credibly serve as a sensory placebo for succimer, and explicitly suggested a validation procedure: query parents on whether they knew which arm their child was in, and see if they guessed correctly. The published TLC literature never reports having performed this check. Whether it was ever run and the data suppressed, or never run at all, is a question court discovery or internal records would have to settle.
NIEHS also made clear that the minimum acceptable design was “random treatment assignment and blind psychometric assessment” — a standard that permits single-blind designs (in which clinicians manage blood lead openly) as long as the outcome assessors are masked. The trial’s advertised “double-blind” status was therefore not a protocol requirement imposed by the funder — it was a proposal choice by the investigators, evaluated as a scoring criterion.