Aluminum's Complex Influence in Provocation Testing: A Multi-Method Validation Characterizing Iatrogenic Artifacts, Endogenous Interactions, and Mitigation Strategies

Provocation testing with chelating agents like Na₂Ca-EDTA serves as a powerful clinical tool that enhances the detection of urinary trace metals and offers valuable insights into the relationships between body burden and environmental exposure. However, the simultaneous mobilization and transport of multiple metals—competing for the same binding sites on provocation chelators and within physiological storage, transport, and excretion mechanisms—can create complex, sometimes counterintuitive, urinary excretion patterns that complicate interpretation. Additionally, trace-metal impurities in the chelating solution can significantly skew results [1], which is a particular concern in high-precision applications such as environmental epidemiology.

Our investigation was initially prompted by puzzling clinical observations: repeated concurrent urinary spikes of uranium and aluminum, both non-essential and potentially toxic elements (Fig. 1B). With urinary aluminum typically exceeding uranium by three orders of magnitude and no other apparent confounders, their association most likely indicates shared exposure pathways (either environmental or iatrogenic) or an aluminum-driven enhancement of uranium excretion, rather than the reverse. Besides its association with uranium, aluminum also shows strong positive associations with a range of essential metals (e.g., manganese, iron, chromium) and other non-essential or toxic metals, such as cadmium, lead, and uranium (Figs. 1A–3), raising concerns about the pervasive influence of iatrogenic confounding.

The spring 2017 uranium spike [2] presents an intriguing characteristic: it occurred without a concurrent general aluminum surge (Fig. 1B). Yet, Figure 1A reveals an even stronger association between enhanced uranium excretion and relatively low aluminum levels (<100 µg/g creatinine) within this spike cohort. These aluminum concentrations, while not minimal in absolute terms, are substantially lower than those typically resulting from iatrogenic sources (e.g., Na₂Ca-EDTA impurities) and are presumed to reflect an endogenous burden from prior environmental deposition. Critically, it is this environmentally derived, chronically deposited aluminum that correlated most strongly with uranium excretion. This apparent paradox points to aluminum's influence on uranium excretion being multifaceted, involving at least two distinct components—one linked to chronic environmental/endogenous aluminum and another to acute, high-concentration iatrogenic aluminum from chelators—a distinction corroborated by the bimodal distribution of aluminum itself, as revealed by Gaussian Mixture Models (Fig. 11A). Each component subsequently exhibits distinct dynamics in its interaction with uranium. Indeed, the data indicate that chronically deposited environmental aluminum more significantly amplifies provoked uranium excretion. This enhanced efficacy likely stems from its prolonged tissue residence, allowing for deeper systemic distribution and more extensive disruption of metal-protein bonds [3]. In contrast, acute iatrogenic aluminum—often introduced with its chemical reactivity already largely shielded by the chelator and having limited tissue interaction time—consequently shows a less potent effect. This is reflected in the Al-U dose-response, which is considerably steeper for the environmental aluminum subset, signifying a greater excretory impact per unit of chronic aluminum despite its lower overall concentration.

Motivated by these compelling clinical observations and the broader challenge of deciphering complex inter-metal correlations in provoked urine, we undertook a systematic, multi-phase investigation. This study aimed to:

1. Rigorously measure trace metals in all infusion solutions in a first QC Phase: DMPS and glutathione showed negligible impurities, but Na₂Ca-EDTA proved to be the dominant source of aluminum, iron, and manganese contamination (Fig. 8A; Excel 'Aluminum_CEMET_QC_Modelling'), corroborating earlier findings [1].
2. Dissect the dual impact of aluminum, both from environmental/endogenous origins and iatrogenic Na₂Ca-EDTA impurities, on post-provocation urinary metal excretion profiles using a multi-method approach. This involved:
   - Extensive correlation analyses (Table 1, Figs. 4–10, Excel 'Data_Sets_Imputation_Correlation'),
   - Pharmacokinetic (PK) modeling (Fig. 10),
   - Gaussian Mixture Models (GMMs; Figs. 11A–C), and
   - Comparative provocation with cleaner Zn-DTPA to differentiate exposure and biological interactions from iatrogenic impurity artifacts (Figs. 8E, 12), including a comparative Gaussian Mixture Model (GMM) analysis of aluminum distributions in Na₂Ca-EDTA- vs. Zn-DTPA provoked urines (cf. Figs. 11A and 11D).
3. Establish a framework, including potential mitigation strategies, for more robust analysis of provoked urine data in future environmental and clinical research.
   

Key Findings & Conclusions:

1. Significant and highly variable Na₂Ca-EDTA impurities: Quality control assays revealed highly variable aluminum (3.2–38.7 mg/L) and iron (1.9–29.0 mg/L) content in commercial Na₂Ca-EDTA solutions (Table 1A, Figs. 5, 8A). This iatrogenic metal load directly drives elevated urinary aluminum and possibly iron excretion in a dose-responsive manner (Table 1B, Figs. 6, 10).
   
   
2. PK modeling examines aluminum sources: A PK modeling framework, utilizing known Al content from individual Ca-EDTA vials and cohort-optimized parameters (model diagnostics in Fig. 10), was employed to estimate the contributions of endogenous and iatrogenic (solution-derived) aluminum to post-provocation urinary excretion. This modeling indicated that iatrogenic aluminum represents a substantial fraction of total urinary aluminum (e.g., cohort average ~44%) with individual model-derived estimates varying widely. However, the model's considerable residual variability (Fig. 10B), even after accounting for vial-specific iatrogenic Al, emphasizes that variations in individual endogenous aluminum burdens are also a major determinant of urinary Al levels following provocation. The precision of specific source apportionment is further affected by model limitations and parameter uncertainties.
   
   
3. Al–U interaction is primarily biological: Despite negligible direct uranium contamination in Na₂Ca-EDTA solutions, a consistent positive correlation is observed between urinary aluminum and uranium with both Na₂Ca-EDTA provocation (partial Spearman ρ≈0.45; Figs. 4, 8C-E) and even with the less-impurity-prone Zn-DTPA provocation (Spearman ρ≈0.36; Figs. 8E, 12). This persistence suggests an intrinsic Al–U interaction, likely where aluminum (endogenous and, to a lesser extent, iatrogenic) enhances uranium excretion via competitive displacement from transport proteins (e.g., transferrin, albumin), consistent with established aluminum physiology [3–7].
   
   
4. Gaussian Mixture Modeling (GMM) Highlights Chelator-Specific Metal Excretion Patterns: GMM reveals clear, chelator-specific patterns for each metal:
   
   Aluminum (Al)
   Na₂Ca-EDTA (Fig. 11a): Two components (N=2 optimal):
   Component 1 (Endogenous; Weight ≈ 0.29): Lower-concentration peak reflecting environmental burden.
   Component 2 (Iatrogenic; Weight ≈ 0.71): Higher-concentration peak from Na₂Ca-EDTA Al impurities
   Zn-DTPA (Fig. 11d): Single component (N=1 optimal by BIC), confirming negligible chelator-derived Al.
   
   Iron (Fe)
   Na₂Ca-EDTA (Fig. 11b): Two components (N=2 optimal):
   Component 1 (Iron-Deficiency Floor; Weight ≈ 0.05): Minor group where depleted body stores yield near-zero chelatable iron [8,9].
   Component 2 (Iron-Replete; Weight ≈ 0.95): Dominant group, forming an approximately log-normal
   excretion distribution.
   Zn-DTPA (Fig. 11e): Two components (N=2 optimal):
   Component 1 (Normal/Low-Normal Stores; Weight ≈ 0.75): Primary group clustering around moderate
   excretion reflecting the iron-replete population.
   Component 2 (High-Excretion/Potential Overload; Weight ≈ 0.25): Smaller, higher-concentration
   subgroup, possibly reflecting individuals with higher iron stores or iron overload responsive to DTPA [10].
   
   Uranium (U)
   Na₂Ca-EDTA (Fig. 11c): Single component (N=1 optimal), with no detectable iatrogenic U peak.
   Zn-DTPA: Insufficient detections for reliable GMM analysis of its population distribution.
   
   Together, these GMM results underscore that only Na₂Ca-EDTA introduces a distinct iatrogenic aluminum peak, whereas Zn-DTPA yields a purely endogenous aluminum distribution. Iron excretion patterns under both chelators mirror underlying iron status—deficiency, repletion, or overload—rather than chelator impurities, and uranium remains uniformly mobilized by Na₂Ca-EDTA with no chelator-driven modes and is too sparsely detected under Zn-DTPA to model.
   
   
5. Iatrogenic aluminum modulates urinary metal excretion: both environmental Al and that introduced by Na₂Ca-EDTA jointly govern the Al–U dose–response and alter inter-metal correlations in urine, with the iatrogenic fraction dominating at aluminum levels > 150 µg Al/g creatinine.
   
   
6. Zn-DTPA confirms endogenous aluminum mobilization: Provocation with cleaner Zn-DTPA also reveals persistent and robust Al–metal correlations (e.g., Al–Mn, Al–Fe), highlighting the contribution of endogenous aluminum mobilization and its biological interactions (Figs. 8E, 12, Excel 'Data_Sets_Imputation_Correlation').
   
   
7. Mitigating Iatrogenic Effects by QC: The variability of Na₂Ca-EDTA impurities—primarily in aluminum and, to a lesser extent, iron and manganese—poses significant confounding potential. This multi-modal validation study highlights the critical importance of awareness and management of chelator solution variability in provocation testing. By characterizing these artifacts and proposing mitigation approaches, we aim to provide a framework for more robust and reliable analysis of urinary metal excretion data in both clinical and environmental research settings. For valid downstream environmental epidemiology, mitigation strategies are crucial, such as excluding samples with urinary aluminum concentrations exceeding a threshold indicative of substantial iatrogenic load (e.g., > 140 µg/g creatinine). The datasets and detailed analyses supporting these conclusions are provided herein.
   
   
8. Biological significance of environmental aluminum: The steep initial Al–U dose–response (Fig. 1) and durable Al–metal correlations under Zn-DTPA (Figs. 8E, 12) reveal that low-level, environmentally deposited aluminum can mobilize other metals more powerfully than larger iatrogenic loads. Future provocation testing frameworks should therefore aim to quantify and correct for this baseline endogenous aluminum burden.

References

1) Blaurock-Busch E, Strey R, editors. Chronische Metallbelastungen – Toxikologie, Diagnose und Therapie [Chronic Metal Exposure: Toxicology, Diagnosis and Therapy]. Norderstedt (Germany): BoD – Books on Demand GmbH; 2017. p. 155. Available from: https://d-nb.info/1147825653.

2) Carmine TC. The Uranium Episode (March–May 2017) in Temporal Context: Associations with CEMET Uranium, Aluminum, and Local PM10 Exposure (2016–2019) [dataset]. figshare; 2024. Available from: https://doi.org/10.6084/m9.figshare.27435639.v5.

3) Rahimzadeh MR, Rahimzadeh MR, Kazemi S, Amiri RJ, Pirzadeh M, Moghadamnia AA. Aluminum Poisoning with Emphasis on Its Mechanism and Treatment of Intoxication. Emerg Med Int. 2022 Jan 11;2022:1480553. doi: 10.1155/2022/1480553. PMID: 35070453; PMCID: PMC8767391.

4) Fatemi SJ, Kadir FH, Moore GR. Aluminium transport in blood serum. Binding of aluminium by human transferrin in the presence of human albumin and citrate. Biochem J. 1991 Dec 1;280 ( Pt 2)(Pt 2):527-32. doi: 10.1042/bj2800527. PMID: 1747128; PMCID: PMC1130580.

5) Cochran M, Coates J, Neoh S. The competitive equilibrium between aluminium and ferric ions for the binding sites of transferrin. FEBS Lett. 1984 Oct 15;176(1):129-32. doi: 10.1016/0014-5793(84)80926-6. PMID: 6489514.

6) El Hage Chahine JM, Hémadi M, Ha-Duong NT. Uptake and release of metal ions by transferrin and interaction with receptor 1. Biochim Biophys Acta. 2012 Mar;1820(3):334-47. doi: 10.1016/j.bbagen.2011.07.008. Epub 2011 Aug 17. PMID: 21872645.

7) Golub MS, Han B, Keen CL. Aluminum alters iron and manganese uptake and regulation of surface transferrin receptors in primary rat oligodendrocyte cultures. Brain Res. 1996;719(1-2):72–77. doi:10.1016/0006-8993(96)00087-X.

8) Powell JJ, Burden TJ, Greenfield SM, Taylor PD, Thompson RP. Urinary excretion of essential metals following intravenous calcium disodium edetate: an estimate of free zinc and zinc status in man. J Inorg Biochem. 1999 Jun 30;75(3):159-65. doi:10.1016/S0162-0134(99)00054-9. PMID: 10474201.

9) Wahidiyat PA, Wijaya E, Soedjatmiko S, Timan IS, Berdoukas V, Yosia M. Urinary iron excretion for evaluating iron chelation efficacy in children with thalassemia major. Blood Cells Mol Dis. 2019 Jul;77:67-71. doi:10.1016/j.bcmd.2019.03.007. PMID: 30978615.

10) Barry M, Cartei G, Sherlock S. Quantitative measurement of iron stores with diethylenetriamine penta-acetic acid. Gut. 1970 Nov;11(11):891-8. doi:10.1136/gut.11.11.891. PMID: 5492246.