Technical method

Technical methodA histochemical method for thedetermination of titanium in tissuessurrounding implants

D. F. WILLIAMS and D. ADAMS School of DentalSurgery, University of Liverpool

Titanium is now accepted as a biocompatible implantmaterial and is extensively used in orthopaedics, oralsurgery, and neurosurgery. It is generally found to bewell tolerated by both soft and hard tissues. It isknown, however, that the tissue adjacent to titaniumimplants does contain some titanium after a periodof implantation. Laing et al (1967) and Meachim andWilliams (1973) found, using different techniques ongross tissue samples, that the titanium concentrationin tissues adjacent to titanium implants was sig-nificantly higher than in control specimens.

Histological examination of the tissue frequentlyrevealed particulate matter which was morpho-logically indistinguishable from haemosiderin andgave a positive Perls stain for ferric iron. Some,however, gave a negative stain for ferric iron andappeared dark-brown or black in haematoxylin andeosin stained sections. This material is responsiblefor the discolouration often noted on naked-eyeexamination of the tissue samples.

It is not known what is the nature of these particlesalthough it is natural to assume that they containtitanium. The metal itself does not corrode in thebody because an ultrathin protective layer of oxideforms on the metal surface, preventing furtherinteraction between the titanium and the environ-ment. It is possible, however, for there to be a veryslow diffusion of titanium ions through the oxidelayer, and also a very slow dissolution of the outersurface of the layer (Williams, 1976). In those caseswhere the titanium is used for the replacement ofhard tissue, some wearing of the surface, againsteither bone or another prosthetic component, mayalso take place.The usual histological stains do not provide any

evidence concerning the nature of the particles inthe tissue surrounding titanium implants. Hithertoit has been possible to analyse the gross tissue sampleonly for total titanium content. It is the purpose of

Received for publication 24 February 1976

this paper to describe a modification of a histo-chemical method, originally described by De Vriesand Meijer (1968), which is able to confirm whetherthese particles do contain titanium and which maybe used in a routine pathology laboratory.

Material and methods

Titanium powder and titanium dioxide powder wassurgically placed in the leg muscles of 10 adultblack-and-white hooded rats anaesthetized withveterinary nembutal. At weekly intervals between8 and 12 weeks, two animals were sacrificed, andblocks of muscle were removed from the area ofimplantation, fixed in formol saline, and blocked inparaffin wax. As a control for the method, salts ofsix other metals, which may commonly be found inimplant materials, were inserted, and sections fromthese animals were subjected to the technique. Thesalts were those of aluminium, cobalt, copper,chromium, iron, and nickel and were identified asbeing present in the sections by standard inorganicspot tests (Vogel, 1962) which are outlined in thetable. In vitro tests were also carried out to confirmthat the above salts did not interfere with the identi-fication of titanium. Once the technique had beenstandardized using animal tissue it was then usedon human tissue from around a titanium implantwhich had been removed surgically. Human controlmaterial was provided by tissue known to containgraphite particles which by routine histology gavean appearance similar to the titanium sections.

THE MICROINCINERATION FURNACEAn electric furnace was constructed for the minerali-zation of the sections. Figure 1 shows a diagram of

Salt Confirmatory Test

Aluminium One drop of saturated alcoholic alizarin put on sec-tion and slide held in ammonia fumes; a red/violetcolouration results

Cobalt Slide held in ammonia fumes and then one drop of1% solution of rubeanic acid in alcohol added; ayellow/brown precipitate results

Copper As cobalt; a black precipitate resultsChromium A green precipitate with the addition of sodium

phosphate solutionIron Dark blue precipitate with potassium ferricyanideNickel As cobalt; a blue precipitate results

Table Outline of tests performed to confirm presenceof metals in specimens

657

657

on January 30, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.29.7.657 on 1 July 1976. D

ownloaded from

658

Varibc

Fig 1 Diagram ofmicroincinerationfurnace. For explanationsee text.

U\J

the apparatus and the essential parts. The tubesare made of Pyrex-glass.A Tube measuring 37 x 25 mm on which the

slides carrying the sections are placed back toback

B Tube measuring 305 x 25 mm and closed atone end. An indentation on the lower side oftube B fits over the end of a steel rod L. Thetube is also provided with four small glassbulbs to keep it concentric with respect to atube G. The thickened upper side of the tubeis polished smooth. During the heating pro-cess this tube can be covered with

C A plate of Pyrex-glassD A stainless steel weightE Rubber stopper with two holes, the central one

Technical method

receiving the steel rod L, the other admittingair via tube J

G Tube measuring 570 x 34 mm, around whichis wound the wire from a l-kW electric fireelement secured at each end by a brassbracket. The connecting wires are fibreglassinsulated and come out between tubesG andH. Tube G is kept centrally within H byasbestos strips at the top and by the stopper Eat the bottom. The furnace is connected tothe mains via a Variac; thus by placing ther-mocouple leads between tubes H and K asetting may be obtained which maintains atemperature of 450°C within the furnace

H Tube measuring 630 x 55 mm in which therubber stopper is mounted

J Tube to admit compressed airK Tube measuring 650 x 70 mm resting on a

metal mantel made of aluminium sheet andserving to attach the oven to the wall. TubesH and K provide for heat insulation and as asafety measure in case of breakage of tubesB or G

L Steel rod, 5 mm in cross-section, by whichthe height of tube B can be controlled.

SOLUTIONS10% aqueous solution of sodium metasilicate filteredthrough 42 filter paper10% aqueous solution of tannic acid filtered through42 filter paper

PROCEDURE1 Six sections are deparaffinized in xylene, rinsed

in absolute ethanol, and air dried.2 After a 10-second immersion in the sodium

metasilicate solution the slides are centrifugedfor several mintes at 2000 rev/min and are thencarefully cleaned.

3 Two drops of concentrated sulphuric acid areplaced in tube B, and the slides are placed in itback to back in pairs and at right angles to eachother.

4 The height of tube B is adjusted by means of rodL until the slides are within the heating coil butthe bottom of the tube containing the sulphuricacid is without. The furnace is switched on andleft for at least 30 minutes.

5 After this time tube B is raised until the sulphuricacid is just seen to boil.

6 After a further 30 minutes tube B is lowered toits original level, the power is turned off, andcompressed air is introduced.

7 As soon as the slides are cool enough they areremoved with a pair of long tweezers and placedin the tannic acid solution.

on January 30, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.29.7.657 on 1 July 1976. D

ownloaded from

Technical method 659

'A ZR>.4.

V._ j\. vv C

ri , ,/ &S 'S'\!s 'Dk

Fig 2 A section of tissue from around a titanium implant. Transmitted light.Haematoxylin and eosin x 36.

,*~~~~x, -Na

7 <

vw'' +''¶ -ioS4rFig 2 A section of tissue fromtarondngaptitanuemln.Transmitted light.HanEx36

Haeatxyinan esinx 6

~~~: r~~~~~, *~~~~~~~~s~~~S *~jA

W t4)OU7IF~~~~~~~~~~~~~~~~~~~~~~~~ *

I* *av I~~~~~~~~~~~~~~~~~~~~~~~~~~

4~~~~~~~~~~J A'~~~~~~~~~4~

SN .*',~~~~~~~~~~~~~*u *~ ~~~LFig4Asetio o tisu cotiiggahte rnmte igt n 6

on January 30, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.29.7.657 on 1 July 1976. D

ownloaded from

Technical method

Fig 3 A section of tissue from around a titanium implant. Reflected light. Tannicacid x 80.

Fig 5 A section of tissue containing graphite. Reflected light. Tannic acid x 80.

660

on January 30, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.29.7.657 on 1 July 1976. D

ownloaded from

Technical method

8 After five seconds' staining they are removed andbriefly passed through water, 60% acetone, andacetone and then air dried.

9 The slide may be viewed using bright field trans-mitted light when titanium-containing areasappear yellow. Better visualization is achievedusing reflected light with dark field illumination.

Results

Immersion of the section in sodium metasilicateensures good reproduction of the original tissuemorphology as this forms insoluble silicium dioxidewhen heated. The initial heating period convertsany metal present to its oxide which in turn is con-verted to the sulphate by the sulphuric acid. Titanylsulphate then forms a yellow precipitate with thetannic acid solution.The series of controls showed that:

1 This sequence of chemical changes takes placewith titanium salts in vitro and no interference isexperienced when aluminium, cobalt, copper,chromium, iron or nickel are present.

2 A yellow precipitate is not produced by any ofthese other salts using the same technique.

3 Titanium may be identified in animal tissue afterimpregnation of the tissue with titanium saltsusing this technique.

Figure 2 shows a section of human tissue obtained

from the location of a titanium implant that hasbeen stained with haematoxylin and eosin while fig 3shows a further section stained using the techniquedescribed. Figures 4 and 5 show sections of humantissue containing graphite that have been stained bythe same methods, illustrating the similarity of thetwo haematoxylin and eosin stained sections but thedifferences between titanium- and non-titanium-bearing sections when stained with tannic acid.

Our thanks are due to Mr G. Abbott for construc-tion of the microincineration furnace, to Miss E.Mort for technical assistance, to Mr J. S. Bailieand the Department of Medical Illustration for theillustrations, and to Dr G. Meachim for advice andprovision of the graphite-containing material.

References

De Vries, G. and Meijer, A. E. F. H. (1968). Histochemicalmethod for identification of titanium and iron oxides inpulmonary dust deposits. Histochemie, 15, 212-218.

Laing, P. G., Ferguson, A. B. Jr., and Hodge, E. S. (1967).Tissue reaction in rabbit muscle exposed to metallicimplants. J. biomed. Mater. Res., 1, 135-149.

Meachim, G. and Williams, D. F. (1973). Changes in non-osseous tissue adjacent to titanium implants. J. biomed..Mater. Res., 7, 555-572.

Vogel, A. I. (1962). A Text-book of Macro and SemimicroQualitative Inorganic Analysis. Longmans, London.

Williams, D. F. (1976). Corrosion of implant materials.Ann. Rev. Mater. Sci., 6 (In press).

Letters to the Editor

No mouse model for H. influenzae infec-tion

In a recent series of experiments in-vestigating the toxic components ofH. influenzae type b, an animal model wasrequired for the demonstration of possiblein vivo effects.There are several reports on the effects

of injecting or inoculating rabbits and ratswith H. influenzae (Schneerson andRobbins, 1971; Smith et al, 1973; Moxonet al, 1974). As in humans, rabbits showan age-dependent susceptibility (with amaximum susceptibility at three weeks),which is explained by the level of serumanti-type b antibody.

Rats also show an age-dependentsusceptibility to H. influenzae (only youngrats are susceptible) but this does not

appear to be influenced by the level ofserum antibody (Smith et al, 1973).As the mouse would have been the most

suitable animal from the point of view ofavailability and general convenience it wasdecided to investigate the possibilitiesof a mouse model for H. influenzaeinfections.

There are two reports of infection inyoung mice following injection of H.influenzae: one describes injection of amucin-enhanced suspension (Alexanderet al, 1944) and the other injection ofinfected CSF from children suffering frommeningitis (Wollstein, 1911). No report ofthe effect of injection of a known viablebacterial count of H. influenzae in micewas found.We carried out two experiments on 11

litters of strain CD-1 mice (Charles

Rivers (UK) Ltd) from 0 to 10 days oldeach litter containing between 8 and 15mice). H. influenzae type b (NCTC 8467)was grown in Brain Heart Infusion (forti-fied with haemin and NAD), washed,counted, and resuspended in PBS andinjected one hour later. Each mouse wasinjected (in the first experiment sub-cutaneously and in the second intra-peritoneally) with a 0 05 ml suspensioncontaining 107 viable organisms. Themice never looked ill and no deaths oc-curred in either series of experiments.The animals were killed at five days, theblood expressed from the head was cul-tured, and H. influenzae was isolated fromtwo mice only, both in the intraperitonealinjection experiment-one from the new-born group and one from the one-daygroup. In a third experiment a massive

661

on January 30, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.29.7.657 on 1 July 1976. D

ownloaded from