ORIGINAL ARTICLE

Comparison of Gd DTPA-BMA (Omniscan) versus GdHP-DO3A (ProHance) Retention in Human Bone Tissue byInductively Coupled Plasma Atomic Emission Spectroscopy

Wendell A. Gibby, MD,* Krissa A. Gibby,† and W. Andrew Gibby‡

Rationale and Objectives: Human bone tissue was collected fol-lowing administration of a clinical dose of gadolinium chelate (0.1mmol per kg) to patients undergoing hip joint replacement surgeryto determine if measurable differences in Gd deposition occurbetween 2 widely available magnetic resonance contrast agents.Materials and Methods: Gd HP-DO3A (ProHance), Gd DTPA-BMA (Omniscan), and an age-matched control population withouthistory of gadolinium chelate administration were compared. Bonesamples were collected fresh, placed in refrigeration, and subse-quently frozen. Tissue digestion was performed using a microwavetissue digester and concentrated nitric acid. A method of tissueanalysis was created for gadolinium using inductivity coupledplasma atomic emission spectroscopy (ICP-AES).Results: Tissue retention was 1.18 � .787 �g Gd/g bone (N � 10)for Omniscan and 0.466 � .387 �g Gd/g bone (N � 8) for ProHancemeasured by ICP-AES.Conclusion: Omniscan (Gd DTPA-BMA) left 2.5 times more Gdbehind in bone than did ProHance (Gd HP-DO3A).

Key Words: human bone gadolinium retention, gadolinium ICP-AES, Gd HP-DO3A, Gd DTPA-BMA

(Invest Radiol 2004;39: 138–142)

Gadolinium (Gd) chelates are widely used as contrast-enhancing agents for magnetic resonance imaging. The

structure of the chelate in part determines the kinetic labilityof the metal ligand and, in turn, can affect the release of themetal in vivo. Numerous in vitro studies have indicated

differences in Gd chelate affinities and in Gd chelate kineticlabilities. These can be affected by a variety of factors,including the pH, the availability of other metal ions thatcompete with the Gd on the chelate, and the structure of thechelate. Gd DTPA-BMA is a substituted open-chain chelatewhich, although having somewhat similar thermodynamicstability1 to the ring compound Gd HP-DO3A, has vastlydifferent kinetic stabilities as a result of the relatively rigidring structure of HP-DO3A.2

Animal studies have indicated repeatedly an increase inGd retention for the more labile Gd DTPA-BMA materialboth as a chemical entity and as formulated for clinical use.However, there have been no human studies to date that haveattempted to demonstrate a significant difference in Gd re-tention. The retention of Gd could be important clinically,because Gd is not a naturally occurring biologic constituentand, once within the tissues of animals, persists for longperiods of time.3 It has significant toxicities, both in in vitroand in vivo experiments.4–6 For example, it is the most potentcalcium antagonist known.7,8 Gd has the potential of leechinginto membranes, bone, and enzymatic structures, causingas-yet undetermined long-term consequences. Therefore, therelease of Gd into the human body is of significant clinicalinterest.3,9,10

This study was undertaken to compare 2 U.S. Food andDrug Administration-cleared, commonly used Gd-basedmagnetic resonance imaging chelates, Gd DTPA-BMA andGd HP-DO3A, in their retentive properties in human bonetissue. Human bone tissue was selected for 2 reasons: 1) it isavailable in certain orthopedic procedures, whereas othertissues such as the liver, spleen, and so on, are not readilyacquired; and 2) bone is one of the target organs in which Gdretention occurs.

METHODSPatients undergoing a total hip arthroplasty with re-

moval of the femoral head were enrolled after informedconsent. Gd DTPA-BMA or Gd HP-DO3A was injected

Received August 29, 2003 and accepted for publication, after revision,November 23, 2003.

From the *Riverwoods Advanced Imaging Center, Provo, Utah; the †De-partment of Tumor Biology, Georgetown University, Washington, DC;and ‡Magnetic Research Inc., Provo, Utah.

Reprints: Wendell A. Gibby, MD, Riverwoods Advanced Imaging Center, 280W. Riverpark Dr., Ste. 100, Provo, UT 84604. E-mail: wgibby@novared.net

Copyright © 2004 by Lippincott Williams & WilkinsISSN: 0020-9996/04/3903-0138DOI: 10.1097/01.rli.0000112789.57341.01

Investigative Radiology • Volume 39, Number 3, March 2004138

intravenously at a dose of .1 mmol/kg not less than 3 days andnot more than 8 days before surgery. The study was per-formed under the auspices of the Institutional Review Board.Table 1 indicates the age distribution and timing of the Gdinjection between the 2 groups. An age-matched controlpopulation undergoing hip replacement was also obtained.The femoral heads were cut in half. Half of the tissue samplewas sent to Magnetic Research Incorporated, Provo, Utah,for analysis. The other half was sent to Bracco Research forindependent analysis.

Experimental ICP-AESEquipment and Materials

Argon from Praxair: 5.0 ultra-high purity. Composition:hydrocarbons less than 1 part per million, oxygen less than 2parts per million, and moisture less than 3 parts per million. ICPAES Instrument: Varian Liberty Series 2, Plasma 96, Fischer-brand-approved pump tubes. Internal diameter: 0.051. Manualsample tube aspiration connected to a short piece of Teflontubing. Software version 1.12. Operating conditions: Nebulizerset at 180 psi. Gas inlet pressure set at 100 psi. Integration time:1 second. Peak tracking window: 0.080 nm. Replicates: 3.Grating order: 1. Power: 1 kW. Viewing height: 10 mm. Generalsettings: Scan window first order 0.12 nm. Photo multiplier tubevoltage: 640. Plasma flow: 10.5 L/min. Auxiliary flow: 0.75L/min. Introduction settings: Sample uptake delay: 10 seconds.Pump rate: 15 RPMs. Instrument stabilization delay: 5 seconds.Rinse time: 300 seconds. Elemental analysis parameters for Gdwere set at 1-second integration time with a polynomial-plottedbackground correction and an analog wavelength of 342.247nm. The parameters for Yttrium were 1-second integration timewith automatic background correction and an analog wavelengthof 371.030 nm. Wavelength calibration and torch alignmentwere performed before each run. The 4-Gd standards were alsorun at the beginning and end of each run. Gd linearity and limitof detection was measured for standard solutions linear andreproducible to .1 �g Gd/g solvent (Fig. 1).

ReagentsThe following reagents were used: micron-filtered dis-

tilled water. Nitric acid, 70% (vol/vol): with heavy metalsless than 0.2 parts per million. Frozen, unprocessed humanbone tissue cut with a standard hacksaw in the pathology

laboratory. Yttrium standard for ICP Y2O3 in 2% HNO3;1000 �g/mL, gadolinium standard for ICP both from E. M.Science, Gibbstown, NJ. Yttrium stock solution was preparedby diluting 1:100 with 5 mL of the Yttrium stock at 10 �g/gH20 solution added to 495 mL of distilled water, creating a10.000 �g Yt/g H20 solution. Gd standard calibration solu-tions were prepared at: 10.036, 1.0043, 05031, and .2004 �gGd/g H20 in a washed 500 mL Pyrex volumetric flask at23°C; along with 36.68 g CaCl2 � 2 H2O; 0.5000 g Yttriumstandard solution (1000 �g/mL) and sufficient water to con-stitute 500 mL of stock solution. Standard solutions alsocontained 100 mL of 70% nitric acid and 36.68 g CaCl2 per500 mL H2O. Yttrium and Gd standards were weighed withan OHAUS Explorer analytical balance capable of measuringto 0.0001 g weighed in small plastic Dixie cups. Each boneassay was performed in triplicate.

Tissue DigestionBone tissue was digested by removing 1 g (� 0.001 g)

from the hip bone samples. Bone samples included a slice ofcortex and medullary bone, which contained marrow. Weavoided the dome of the femoral bone where cartilage anddegeneration were present. This was placed in a Pyrex StarSystem CEM Digestion Flask. Twenty milliliters of the nitricacid reagent was automatically pipetted by the CEM instru-ment (software version 86005/1.03). Ramp time was 5 min-utes. Target temperature was 95°C and the time at tempera-ture was 15 minutes. The bone samples were watchedvisually until nearly all the nitric acid boiled away. Whenapproximately 2.0 to 3.0 mL of residual acid and dissolvedbone are present, the sample is removed, 1 mL of the dilutedYttrium solution (10 �g/g H20) is added, and the volume ofthe sample is then adjusted to 10 mL with distilled water. (Ifthe sample is not caught before complete loss of the nitricacid, significant charring occurs and the sample must bediscarded.)

Recovery of Gd From Spiked BoneThe recovery analysis and limit detection of human

bone tissue was performed using bone tissue spiked withvarying concentrations of Gd from a single control bonebetween .1 and 1.0 �g Gd/gm bone (Fig. 2). The slope error

TABLE 1. Age Distribution and Gadolinium Timing

Control (N � 7) Omniscan (N � 10) ProHance (N � 8)

Average patient age* (years) 55.8 � 20.2 67.3 � 12.8 63.6 � 9.7Injection to bone harvest* N/A 4.4 days � 1.4 4.5 days � .9

*There was no significant difference between groups.

Investigative Radiology • Volume 39, Number 3, March 2004 Comparison of Gd DTPA-BMA vs Gd HP-DO3A Retention

© 2004 Lippincott Williams & Wilkins 139

was 16%. The recovery of Gd in the spiked control sampleswas 54%. The coefficient of determination (r2) was .8845.

RESULTSThe �g/Gm of Gd in the bone sample is equal to

measured �g/g of bone divided by the internal Yttriumstandard, divided by the bone weight, times 10 mL of solutionper 1 g bone, times the density correction factor of water. The

human data was then normalized for Gd recovery usingspiked bone samples of Figure 2. The formula used was:

Actual Concentration Gd �Measured Gd

.54� .0075

The detection limit of Gd for this system is greater than0.1 but less than 0.2 �g/g bone.

FIGURE 2.

FIGURE 1.

Gibby et al Investigative Radiology • Volume 39, Number 3, March 2004

© 2004 Lippincott Williams & Wilkins140

Table 2 represents the normalized, measured Gd con-tent in micrograms of gadolinium per gram of bone.

We did not attempt to determine any relationship be-tween the time after Gd injection and residual trace metalpresent. This was because the numbers of patients was lim-ited, and we tightly controlled the time from injection to boneharvest; 4.4 � 1.4 days in the case of Gd DTPA-BMA and4.5 � .9 days for Gd HPDO3A.

DISCUSSIONDespite claims to the contrary that “there seems to be

no dissociation of Gd within the body,”11 all of the currentlyused magnetic resonance contrast agents have the potentialfor transmetallation in vivo. Physicians injecting such mate-rial into the body should be aware of the potential deleteriouseffects of free Gd. Gd is one of the most potent inorganiccalcium antagonists known.6,7 High concentrations of heavylanthanide metals in animals are known to be toxic and cancause, among other things, fatty degeneration of the liver,changes in nucleic acid, lipid and carbohydrate metabolism,writhing, ataxia, labored respiration, sedation, hypotension,and death.3 Gd metal injected intramuscularly induces sarco-mas4 and “granulomatous neoplasms.”5 These agents are alsoknown to interfere with coagulation.8,11 Unlike metals with aknown biologic function, Gd does not have a known pathwayfor excretion from within the body. Once within the tissues,it can persist for long periods of time,6 primarily within theliver and bone. In small quantities, free Gd initially goes tothe liver and is then transferred to bone with negligibleelimination over the next 3 weeks.12 In the dosages givenclinically, transmetallation is not known or suspected to be anacute problem, although it could be the cause of the well-known elevated serum Fe levels reported initially with theintroduction of Gd DTPA. These were subsequently cor-rected to a large extent by adding more free ligand toMagnevist’s formulation. However, the potential of transmet-allation should be considered for long-term sequelae, andreasonable efforts should be made to select ligands thatreduce this effect. One of the primary causes of transmetal-lation in vitro is the competition by other bioavailable metals

such as zinc and copper that attack the chelate complex.Some chelates such as Gd DTPA-BMA formulate with 5%excess calcium ligand.13 The excess chelate scavenges thebioavailable metals. Studies with human volunteers haveshown that a single dose of Gd DTPA-BMA removes ap-proximately 32% of total plasma zinc (albeit a small fraction[.09%] of the total zinc pool in the body).14 However, withrepeated high dosages in subacute toxicity studies in animals,monkeys injected with Gd DTPA-BMA demonstrate all ofthe signs of zinc deficiency; including testicular atrophy, skinlesions with ulceration, and gastritis.15

One of the mistakes made in evaluating chelates for invivo use is to compare thermodynamic stability constants.Thermodynamic stability is measured by titrating the metal tothe chelate at a pH of approximately 11. At this pH, there areno competing hydrogen ions for the chelate and a theoreticalmaximum stability for the chelate is obtained. However, themore relevant stability constant for in vivo use is the condi-tional stability constant, which is calculated at a pH of 7.4. Inthe milieu of the body, there are metals that will attack thechelate and displace the Gd. The concentration of thesemetals depends on the microenvironment for the chelate.Furthermore, in the body, there is no equilibrium. The ther-modynamic equilibrium constant calculated in vitro is simplynot germane. Once a Gd ion is released, it is immediatelycarried away by other biologic material, being bound tomembranes, enzymes, bone substrate, or other weak chelatessuch as citrates and phosphates. It might never “see” thechelate again. The more important aspect of chelate safety isthat of kinetic stability. For example, consider 2 chelates ofnear-equivalent conditional thermodynamic stability: Gd HP-DO3A and Gd DTPA. Thermodynamic stability measure-ments at a pH of 7.4 would indicate that only 1 atom of 1017

(a very small number) of Gd would be loose from the chelateat any given time. Yet a variety of studies, including in vivotransmetallation11,16–19 and Gd retention in animals20,21 andhuman case reports,22–24 suggests that there is a significantdifference between the propensity of these 2 chelates to giveup Gd in vivo. Both chelates have the same number of coordi-nating sites. However, HP-DO3A is a ring compound that is

TABLE 2. Results of Gadolinium Bone Retention

Control Gd DTPA-BMA Gd HP-DO3A

Average �g/g fresh bone tissue Average �g/g fresh bone tissue Average �g/g fresh bone tissue�0.117 1.18 .466N � 7 N � 10 N � 8Standard Deviation Standard Deviation Standard Deviation.227 .787 .387t Test Control v ProHance t Test Control v Omniscan t Test Omniscan v ProHance0.0021 0.0004 0.0165

Investigative Radiology • Volume 39, Number 3, March 2004 Comparison of Gd DTPA-BMA vs Gd HP-DO3A Retention

© 2004 Lippincott Williams & Wilkins 141

quite rigid.25 For the Gd to break free, it must simultaneouslybreak 5 to 6 coordination sites. On the other hand, Gd chelatedto DTPA can break free sequentially. Thus, it is much less likelyfor the HP-DO3A to release its Gd.

An agent such as Gd DTPA-BMA, which not only hasa much lower conditional stability constant (14.9), also hasless kinetic stability than HP-DO3A or Gd DTPA. It istherefore more likely to leave more Gd behind as it dissoci-ates in the body. The half-life in the blood for an extracellularcontrast agent is approximately 20 minutes. Given that, vir-tually all of the chelate that could be filtered out of the tissuesshould be gone by 4 to 5 days. Furthermore, these agents arenot taken up by cells. Transmetallation experiments haveshown that Gd is readily displaced from weak open chainchelates by competition with other metal ions (especially Znand Cu) or weak chelates such as phosphate or citrate.18

Furthermore, animal experiments have shown dissociation invivo.16 Based on this, it is likely that most of the Gd detected3 to 8 days after administration is released from the originalchelate. It is possible, although not likely, that the injected Gdchelate could be responsible for at least part of the residual inthe bones.

This is the first article in humans to convincingly showa significant difference in the Gd deposition between 2commonly used magnetic resonance contrast agents. Further-more, we have provided methodology for the measurement ofresidual trace metals in human bone. The potential risk fromGd release and long-term retention rises with higher dosageand increased frequency of use. This information is importantto consider in certain patient populations in whom excretionof contrast is reduced (eg, renal failure or congestive heartfailure) or in patients who will receive repeated exposures toGd chelates (eg, pediatric brain tumor, patients with multiplesclerosis) or in patients in whom very little risk can betolerated (eg, pregnant patients).

REFERENCES1. Tweedle MF, Wedeking P, Kumar K. Biodistribution of radiolabeled,

formulated gadopentetate, gadoteridol, gadoterate and gadodiamide inmice and rats. Invest Radiol. 1995;30:372–380.

2. Tweedle MF. The ProHance story: the making of a novel MRI contrastagent. Eur Radiol. 1997;7:S225–230.

3. Tweedle MT. Physiochemical properties of gadoteridol and other mag-netic resonance contrast agents. Invest Radiol. 1992;27(Suppl 1):S2–S6.

4. Rocklage SM, Worah D, Kim S-H. Metal ion release from paramagneticchelates: what is tolerable? Magn Reson Med. 1991;22:216–221.

5. Costa M. Metal carcinogenesis in animals. Metal Carcinogenesis Test-ing Principles In Vitro Methods. 1980:33–34.

6. Talbot RB, Davison FG, Green TW, et al. Effects of S. Q. implantmentof rare metals. USA EC Report. 1965;1170.

7. Biagi BA, Enyeart JJ. Gadolinium blocks low- and high-thresholdcalcium currents in pituitary cells. Am J Physiol [Cell Physiol]. 1990;259:C515–C520.

8. Molgo J, del Pozo E, Banos JE, et al. Changes of quantal transmitterrelease caused by gadolinium ions at the frog neuromuscular junction.Br J Pharmacol. 1991;104:133–138.

9. Carr JJ. Magnetic resonance contrast agents for neuroimaging. NeuroClin North Am. 1994;4:43–54.

10. Beliles. Trace elements in plant foodstuffs. In: Toxicity of Heavy Metalsin the Environment. New York: Marcel Dekker; 1978.

11. Weinmann H-J, Brasch RC, Press W-R, et al. Characteristics of gado-linium-DTPA complex: a potential NMR contrast agent. AJR Am JRoentgenol. 1984;142:619–624.

12. Hals PA, Hogset A. Deposition of gadolinium after high and low dosesof gadolinium chloride to mice: organ distribution, elimination andsubcellular localization in liver cells. SMRM; 1990; WIP:1199.

13. Cacherris WP, Quay SC, Rocklage SM. The relationship betweenthermodynamics and the toxicity of gadolinium complexes. Magn ResonImaging. 1990;8:467–481.

14. Puttagunta NR, Gibby WA, Smith GT. Human in vivo comparativestudy of zinc and copper transmetallation after administration of mag-netic resonance imaging contrast agents. Invest Radiol. 1996;31:739–742.

15. Harpur ES, Worah D, Hals P-A, et al. Preclinical safety assessment andpharmacokinetics of gadodiamide injection, a new magnetic resonanceimaging contrast agent. Invest Radiol. 1993;28:S28–S43.

16. Kasokat T, Urich K. Quantification of dechelation of gadopentatedimeglumine in rats. Arzeimittelforschung. 1992;42:869–876.

17. Elst LV, VanHaverbeke Y, Goudemant J-F, et al. Stability assessment ofgadolinium complexed by P-31 and H-1 relaxometry. Magn Reson Med.1994;31:437–444.

18. Tweedle MF, Hagan JJ, Kumar K, et al. Reaction of gadolinium chelateswith endogenously available ions. Magn Reson Imaging. 1991;9:409–415.

19. Mann JS. Stability of gadolinium complexes in vitro and in vivo. JCAT.1993;17(Suppl 1):S19–S23.

20. Wedeking P, Kumar K, Tweedle MF. Dissociation of gadolinium che-lates in mice: relationship to chemical characteristics. Magn ResonImaging. 1992;10:641–648.

21. Tweedle MT. Physiochemical properties of gadoteridol and other mag-netic resonance contrast agents. Invest Radiol. 1992;27(suppl 1):S2–S6.

22. Nordenbo AM, Somnier FE. Acute deterioration of myasthenia gravisafter intravenous administration of gadolinium DTPA [Letter]. Lancet.1992;340:1168.

23. Hattner R, White DL. Gallium-67/stable gadolinium antagonism: MRIcontrast agent markedly alters the normal biodistribution of gallium-67.J Nucl Med. 1990;31:1844–1846.

24. Tien RD, Brasch RC, Jackson DE, et al. Cerebral Erdheim-Chesterdisease: persistent enhancement with Gd-DTPA on MR images. Radi-ology. 1989;172:791–792.

25. Tweedle MF. Physicochemical properties of gadoteridol and other mag-netic resonance contrast agents. Invest Radiol. 1992;27(suppl 1):S2–S6.

Gibby et al Investigative Radiology • Volume 39, Number 3, March 2004

© 2004 Lippincott Williams & Wilkins142