dihydrate crystals, pseudogout

Annals of the Rheumatic Diseases 1995; 54: 825-830

Evidence for a causal relationship between thestructure, size, and load of calcium pyrophosphatedihydrate crystals, and attacks of pseudogout

Angela Swan, Brigid Heywood, Barry Chapman, Heather Seward, Paul Dieppe

AbstractObjective-To investigate any relationshipbetween the nature, size, and numbers ofsynovial fluid (SF) calcium pyrophos-phate dihydrate (CPPD) crystals, andattacks ofpseudogout.Methods-Knee SF was aspirated fromnine selected patients, first during anattack of pseudogout (acute sample) andagain later when the attack had subsided(interval sample). CPPD crystals wereextracted, weighed, examined by highresolution transmission electron micro-scopy (HRTEM), and characterised bysize and crystal habit (monoclinic or tri-clinic). Structural analysis was carried outby x ray powder diffraction (XRD) and theproportions of monoclinic to triclinicCPPD were estimated from densitometricmeasurements ofselected key reflections.Results-The mean crystal size, byHRTEM, indicated that the crystals in theacute sample were larger than those in theinterval sample. The ratio of monoclinicto triclinic CPPD, whether estimatedfrom their morphological appearance byHRTEM, or from XRD, was greater in theacute than in the interval sample in allnine patients. The total amount ofextracted mineral varied, but in everypatient the concentration ofCPPD per mlof fluid, and the total mineral per joint,were greater in the acute sample than inthe interval sample.Conclusion-In this highly selected groupof patients, the large numbers of CPPDcrystals associated with attacks ofpseudo-gout included a greater proportion ofmonoclinic crystals, and larger crystals,than those present when inflammationhad subsided. A special, phlogisticpopulation of crystals may exist, origi-nating in different joint tissues, or clearedin a different manner, than the morecommon populations of smaller crystalswith a greater proportion of triclinicCPPD, seen in chronic disease.

(Ann Rheum Dis 1995; 54: 825-830)

The deposition of calcium pyrophosphate di-hydrate (CPPD) crystals in joints is a common,age-related phenomenon.' Large deposits canbe detected radiographically (chondrocalcino-sis (CCA)) and are particularly common in theknee joint, affecting both fibrocartilage menisci

and hyaline articular cartilage.2 Crystallo-graphic analysis of these deposits has invariablyshown them to contain mixtures of monoclinicand triclinic CPPD crystals of varying sizes.'People with radiographic evidence of CPPDdeposits in their knee joints may have noevidence of joint disease; however, there aretwo well established clinical associations:attacks of acute, self limiting synovitis (pseudo-gout), and osteoarthritis (OA).4 5 Both inpatients with pseudogout and in patients withOA and CCA, CPPD crystals are generallypresent in the synovial fluid (SF). Severalinvestigators have shown that CPPD crystalscan cause inflammation6 7 and it is widelyaccepted that these crystals cause the inflam-mation of pseudogout. However, one of theenigmas of the CPPD related disorders, as withother forms of crystal arthropathy, is the factthat the crystals appear to cause attacks ofsevere synovitis, but are also frequently foundfree in the joint fluid between attacks, and inpatients with OA in whom there is very littlesynovitis.There are several possible explanations for

this enigma. The crystal load in the joint maybe critical, and it has been reported previouslythat fluids from patients with pseudogout tendto contain more CPPD than those from OApatients.8 Alternatively, either the size orstructure of the crystals might be important.The size of CPPD crystals identified in jointtissues and fluids varies considerably, and theyare of two types, either monoclinic or triclinicin lattice structure. Using the right hand screwconvention for the labelling of axes and inter-axial angles of crystals, in the monoclinicsystem the axial restriction is that a = p = 90degrees; in the triclinic system there are nosymmetry elements. For CPPD, the usualvisual expressions of these symmetries (crystalhabits) are elongated rod-like for monoclinic(M-)CPPD, and 'squat' rhomboid or cuboidfor triclinic (T-)CPPD. It follows frommathematical principles that the surfacearea:volume ratio of the longer, rod-likeM-CPPD is greater than that of T-CPPD.There is some evidence in the literature thatsmaller crystals can be more phlogistic thanlarger ones,9 that the structure of crystals, inaddition to their surface identity, is importantin determining inflammatory response,'0 andthat monoclinic CPPD may be more activethan the triclinic form." Finally, the proteincoating of the crystal might be the dominantfactor.'2 13

Rheumatology Unit,Bristol RoyalInfirmary,Bristol BS2 8HW,United KingdomA SwanH SewardP DieppeDepartment ofChemistry,University of Salford,Salford,United KingdomB HeywoodDepartment ofPhysics,University ofBath,Bath,United KingdomB ChapmanCorrespondence to:Angela Swan.Accepted for publication23 May 1995

825

on April 5, 2022 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.54.10.825 on 1 October 1995. D

ownloaded from

http://ard.bmj.com/

Swan, Heywood, Chapman, Seward, Dieppe

In order to examine this, we have studied thesynovial fluid crystal load and the size andcrystallographic identity of the individualcrystals in patients with pseudogout, bothduring an attack, and at a time in which therewas little or no evidence of synovitis.

Patients and methodsPATIENTS

Patients presenting with pseudogout of theknee joint were included in this study. Eachpatient presented with the sudden onset of painand swelling of the knee joint, radiographicevidence of CCA and a synovial fluid aspiratewhich had a high count of polymorphonuclearleucocyte (PMN) cells, and CPPD crystals onpolarised light microscopy. The patients weretreated by aspiration of as much synovial fluidas possible from the knee, followed by injectionof a long acting intra-articular steroid prep-aration. The patients were followed up over thenext six months, and in those patients in whomthe affected joint became cool and painless, butshowed clinical evidence of an effusion, asecond SF sample ('interval' sample) wasobtained, the operator again attempting toaspirate as much fluid as possible from thejoint. Sequential SF samples suitable for theextraction and analysis of crystals wereobtained from nine subjects. In each case, kneeradiographs were obtained during the sixmonths of the study and examined by a singleobserver who recorded the presence of CCAand graded it as minor (1), moderate (2) orsevere (3) in amount, in addition to recordingthe grade of OA on the Kellgren and Lawrencescale. 14

SYNOVIAL FLUID EXAMINATION AND

TREATMENT

The synovial fluids were aspirated via a 21gauge (0-8 X 40 mm) needle, into syringeswhich were immediately capped; SF wereexamined, and usually processed within twohours of joint aspiration. When crystal extrac-tion could not be carried out immediately, theSF samples were stored uncentrifuged, insterile tubes with minimal air space, at -70°Cfor up to two weeks before processing; it hasbeen shown that storage under theseconditions, for this time, does not significantlyalter the CPPD content of SF samples."5The total fluid volume was measured and a

small aliquot examined for total leucocytecount using a haemocytometer; a differentialleucocyte count was performed on an air driedsmear. The presence of CPPD-like crystals wasconfirmed by polarised light microscopy usinga Zeiss Axioplan microscope, incorporating apolarised light facility, with first order redquartz plate. The remainder of the sample wasprocessed for extraction of mineral as pre-viously described and validated.'6 Briefly, thefluids were incubated at 37°C with papain1 mg/ml SF w/v (Sigma P4762), then withfresh sodium hypochlorite solution (final con-centration 0/5% v/v). The extracted depositwas washed five times in sterile deionised dis-

tilled water, and finally washed in 98% v/vethanol and dried at 37°C. The small pellet ofmineral deposit was then weighed on a Cahn21 microbalance, before microscopic andanalytical examination.

EXAMINATION OF THE EXTRACTED CRYSTALS

Samples of the extracted crystal pellets wereexamined by polarised light microscopy, highresolution transmission electron microscopy(HRTEM) and x ray diffraction (XRD).HRTEM was carried out using a JEOL

200CX high resolution transmission electronmicroscope, operating at 100 KV, and crystalswere examined under conditions of minimalillumination. For statistical analysis, thecrystals were characterised on the basis of size,and structure (monoclinic and triclinic); thepopulation minimum was 100 crystals andmeasurements were taken from at least threerandomly selected micrographs of eachsample.

Structural analysis by XRD used a PW1010/80 Phillips x ray generator: Cu Kot radia-tion plus nickel filter, with a Debye-Scherrerpowder camera (1 14 83 mm diameter). Therelative intensity of the XRD powder lines inany one film was measured by means of aJoyce-Loebl double beam recording micro-densitometer, Mk I 1 1 C. Principal XRD reflec-tions were used to distinguish between the twomorphological forms of CPPD; these wereat d-spacing 8-04 A (intensity I/Il 100)(reflection A) for T-CPPD, and at d-spacing7-37 A (intensity I/I1 100) (reflection B) and4-65 A (intensity I/I1 90) (reflection C) forM-CPPD. Because the small quantities of bio-genic material available restricted directquantitative analysis of the composition byXRD, a compromise method was based ondensitometric measurements of peak height ofthe selected reflections (peaks A, B, and C).From these data, estimates of the relative pro-portion of M-CPPD and T-CPPD could bedetermined for each sample, with variations inthis ratio indicating changes in the crystalpopulation over time.

ResultsTable 1 shows details of the nine patients.There were five women and four men (meanage 75-4 (range 55-86) years). Seven hadevidence of OA, in addition to CCA on theknee radiograph. The interval betweencollection of the fluid during an acute attackand aspiration of the second (interval) samplevaried from four to 24 weeks (mean interval10 weeks). The volume of SF obtained variedfrom 3 to 50 ml, and the leucocyte count from0-1 to 160 X 105/ml SF, but in each of thepatients both the total volume of fluid aspiratedand the PMNT cell count were greater in thefluid obtained during the acute attack than inthe interval sample.

All nine patients had radiological evidence ofCCA; there was no apparent relationshipbetween the amount of CCA or the amount ofOA change seen on radiographs, and any of the

826



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Causal relationship between CPPD crystals and pseudogout

Table 1 Patient and synovialfluid (SF) data

Patient Age Sex x Ray SF volume (ml) Leucocyte count(yr) (X 105)

CCA K&L Acute Intervalgrade score Acute Interval

1 81 M 2 0 30 15 10 0-12 86 M 2 3 25 9 20 53 78 F 1 3 5 5 10 14 80 F 2 4 5 4 10 0-15 55 F 2 2 50 6 20 0-16 75 F 1 3 20 6 20 27 69 M 1 0 15 5 27 <18 72 M 1 1 5 3 21 <19 83 F 1 2 29 21 160 5

CCA = Chondrocalcinosis-grades: 0= none; 1 = mild; 2 = severe.K & L = Kellgren and Lawrence score for OA (0= none ... 4 = severe).

SF findings (table 1). Table 2 shows theamount and nature of the mineral recoveredfrom each SF sample, the crystal size data, andthe ratio of monoclinic to triclinic crystals ineach sample (by HRTEM and by densitometryof XRD principal reflections). The totalamount of mineral varied widely, from 10 to13000 ,ug, but in each individual there was

considerably more mineral per millilitre of SF,and more total mineral per joint, in the sampleobtained during an acute attack than that takenin the interval period. Table 2 shows thatcalcium hydroxyapatite (HAP) crystals were

also detected in both acute and interval SFfrom four patients; in one patient HAP crystalswere detected only in the acute sample;another patient had detectable HAP crystalsonly in the interval sample. In three patients,neither the acute nor the interval sample con-

tained detectable HAP. Thus there was no

pattern of association of SF HAP crystals withinflammation. The amount of HAP detectedwas invariably very small (just a few HAPcrystals detected by HRTEM); there was no

apparent reduction in HAP crystal numbersduring interval episodes, but the very low con-

centrations made quantification impossible.Additional information on approximate

crystal surface area was calculated frommeasurements of aspect ratios of extractedbiogenic M-CPPD and T-CPPD crystals fromthe electron micrographs. The results providedsome evidence that the mean surface area ofmonoclinic CPPD (40-20 (SD 75 70) p.m2;range 0-94-283-30 p.m2 (n = 20)) was greaterthan that of triclinic CPPD (27-38 (37 60)p.m2; range 0-78-165-12 p.m2 (n = 20)). Theresults clearly showed the major contributionmade to total available surface area by the

larger crystals in this sample though, weight forweight, smaller crystals will always provide a

greater specific surface area than large ones.

Unfortunately, insufficient biogenic materialwas available for the far more accuratemeasurement of crystal surface area by gas

absorption methods.Electron micrographs of CPPD crystals

extracted from SF from patient 6 illustrate thepresence of larger, monoclinic CPPD in theacute sample (fig 1) compared with the intervalsample (fig 2) which contains smaller, mainlytriclinic CPPD. Figures 3 and 4 illustrate theXRD films of these two populations ofextracted CPPD crystals from sample 6,together with the corresponding densitometrictraces. The principal reflections A, B, and C,from which the proportion of monoclinic totriclinic CPPD was calculated for all thesamples, are indicated on the traces.The estimates of mean crystal size, based on

examination by HRTEM of a minimum of 100representative crystals per sample, indicatedthat the crystals in the acute attack samplestended to be larger than those found in theinterval fluids. In addition, the ratio of mono-clinic to triclinic crystals, whether estimatedfrom their morphological appearance byHRTEM, or by densitometry of XRD keyreflections, showed a consistent pattern of a

much greater proportion of monoclinic crystalsin the acute than in the interval SF samples.

DiscussionThe objective of this study was to examine thehypothesis that acute attacks ofpseudogout are

associated with a number and type of CPPDcrystals in the SF different from those found inpatients presenting with little or no synovitis.The hypothesis was based on two previousobservations: the cross sectional data to indi-cate that more CPPD is found in SF frompatients with pseudogout than in those withOA and chondrocalcinosis,8 and the suggestionthat monoclinic CPPD might be more phlo-gistic than the triclinic form." In order to testthe hypothesis, it is clearly desirable to examinesequential SF samples rather than cross

sectional samples, because of the variationboth in patients with CPPD related disordersand in their SF findings. Patients with pseudo-gout usually recover quickly from an acuteattack of synovitis, particularly if treated with

Table 2 Extracted crystal data

Concentration in HRTEM M-CPPD M:TCPPD ratioSF and total mean length (,um)

HRTEM XRDAcute Interval Acute Interval

Acute Interval Acute IntervalConcn Total Concn Total(,ug/ml) f,ug) (,ug/ml) (,ug) B.A C.A B.A C.A

1 100 3000 22 330 3-8 2-8 4-0 2-0 1-6 0-6 1-0 0-32 392 9800 60 540 9-2* 4-9* 6-2 2-4 0-9 0-4 0-3 0-23 2600 13000 2 10 8-4* 5-3* 6-0 2-0 0-8 0-3 NA NA4 110 550 32 128 3-8* 2-4* 4-3 2-7 1-5 1-3 0-9 0-45 40 2000 10 60 3-9* 2-3* 3-6 2-1 1-1 0-6 0-5 0-46 13 234 5 30 5-6 3-8 6-0 2-7 8-0 3-8 1-0 0-57 84 1260 12 60 4-6 3-6* 6-6 2-2 1-6 0-8 0-6 0-38 36 180 40 120 6-4* 3-3 5-2 2-4 1-6 1-2 0-7 0-59 110 3190 13 273 6-4 3-6 7-0 2-4 1-2 0-8 0-3 0-2

*Small numbers of calcium hydroxyapatite crystals also detected by HRTEM.

827



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


Figure 1 Electron micrograph ofcalcium pyrophosphate dihydrate crystals extractedfromacute synovialfluid sample from patient 6. Totalyield of crystals 234 ,g; almost entirelymonoclinic (rod shaped), with a mean length of5 6 ,um.

intra-articular steroids,17 but may have persist-ent small effusions containing CPPD crystalsbetween attacks. These features make thesepseudogout patients an ideal group for investi-gation, and a group of nine patients withclassical diagnoses of pseudogout (based uponclinical, radiographic, and SF findings) was

selected for this sequential study. It should benoted, however, that this was a highly selectedgroup, because many of those who were suc-

cessfully treated for pseudogout had either noeffusion to aspirate, or too little SF for mineralextraction at the follow up visits. It is also clearthat the group selected for study did have someintra-articular inflammation at the time at

which the interval sample was aspirated; theyhad, by definition, an effusion, and the SFanalysis showed slightly greater leucocytecounts than are found in normal SF.18 In spiteof these reservations, the strength of the studyis that it allowed comparison of two sequentialsamples from each patient, and that in every

/_./

Figure 2 Electron micrograph ofcalcium pyrophosphate dihydrate (CPPD) crystalsextractedfrom the interval synovialfluid sample from patient 6. Totalyield of crystals 30 ,ug:a mixture ofM-CPPD (mean length 3-8 ,tm) and T-CPPD (rhomboid/cuboid shaped).

case there was evidence of much more severeinflammation in the acute than in the intervalsample.Our results clearly show that in the acute

episode samples there was a greater crystalload, larger crystals, and a greater ratio ofmonoclinic to triclinic CPPD crystals.Although the number of patients was relativelysmall, the consistency of the data, which showthe same trend in each patient, and the size ofthe differences noted between the two samples,are impressive. Furthermore, the HRTEM andXRD analyses, both of which were carried outblind.to the clinical data, and each of whichused a quite different approach, showed agreater ratio of monoclinic to triclinic crystalsin the acute episode SF compared with theinterval fluid for every patient. The exactamount of monoclinic and triclinic material ineach sample was difficult to compute. The dataobtained by HRTEM are based on a morpho-logical assessment of crystal habit. Asdiscussed earlier, the XRD technique used isonly semiquantitative; the relative intensities ofthe reflections for M- and T-CPPD are notrelated solely to the concentrations of eachcrystal form within a mixture. However, frommeasurements of the relative intensities of thekey XRD reflections for M- and T-CPPD, anindication of the relative abundance of the twoforms was obtained. Changes in this ratio wereobserved in all nine patients, each acute samplehaving a greater ratio than the correspondinginterval sample. This does not indicate that theconcentration of M-CPPD was greater thanthat of T-CPPD in the acute SF sample, butrather that the ratio shifted in favour ofM-CPPD. The morphology of a crystal doesnot always reflect its structure (although this isusually the case). This fact, together with thesemiquantitative nature ofXRD analysis, mayexplain the difference in monoclinic to triclinicratios reported with each technique, thosereported by HRTEM being greater thanthose recorded from XRD data. In contrast,with CPPD, there was no apparent cor-relation between the presence of smallnumbers of coexisting calcium HAP crystalsand inflammation.

Previous data on the CPPD crystal loads inSF have also suggested the presence of agreater amount in pseudogout than in lessinflammatory conditions.8 However, these datawere based on cross sectional rather than longi-tudinal samples, and used chemical rather thangravimetric methods of analysis. The differ-ence in methodology may explain the discrep-ancies in the total amount of mineral found.Each acute SF sample contained a greaterconcentration of CPPD, and a greater totalcrystal load, than the corresponding intervalSF. However, the wide variation in concentra-tion of SF CPPD associated with pseudogoutsuggests that there is not a clear thresholdconcentration for CPPD above which aninflammatory response is triggered. Athreshold concentration might be associatedwith some other feature of the crystalpopulation. Crystal habit (monoclinic versustriclinic) or particle size may be significant.9 11

828



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Causal relationship between CPPD crystals and pseudogout

B7.37

3.03

2.67

3-25 C

4-65

A8.04

A;|Z..~~ ~~~~~~ (A) 7 37

Figure 3 Diffraction pattern (below), and densitometric trace ofx ray powder dijfilm, of extracted calcium pyrophosphate dihydrate (CPPD) crystals from the acuisynovialfluid samplefrom patient 6: d-spacings (in dngstroms (A)) ofprincipal rfor T-CPPD and M-CPPD. Estimates of the ratio ofM- to T-CPPD were obtainthe ratios ofpeak height (mm) ofB.A, and ofC.A. Reflections indicate that thissample consists ofpredominantly M-CPPD crystals.

The data reported here, showing a greaterratio of monoclinic to triclinic CPPD crystalsin acute than in less inflammatory samples, anda greater mean crystal size, are both entirelynew findings. These data raise two questions:first, why is there a difference in the type ofcrystal seen in the two samples, and second,does the crystal type dictate the synovialinflammatory response?The crystals found in pseudogout are

believed to arise through crystal shedding froma preformed deposit in cartilage; data tosupport this include the initiation of pseudo-gout attacks after solubilisation of deposits

B

3.11

3.052-67 3.45

6.94 7-37A

4.65 1

2.56

d (A)=804,,,. os-*s - -g7.37

6904..... 4 65

Figure 4 Diffraction pattern (below), and densitometric trace ofx ray powder diffractionXRD film, ofextracted calcium pyrophosphate dihydrate (CPPD) crystals from the intervalsynovialfluid sample from patient 6: d-spacings (in dngstroms (A)) ofprincipal reflectionsfor T-CPPD andM-CPPD. Estimates of the ratio ofM- to T-CPPD were obtainedfromthe ratios ofpeak height (mm) ofB.A, and ofC.A. Reflections indicate that this sampleconsists ofa mixture ofM-CPPD and T-CPPD crystals.

with magnesium lavage,19 and the dis-appearance of radiographic deposits duringacute attacks.20 Our results indicated thatlarger crystals were associated with inflamma-tory SF. Phlogistic potential is related to thecrystal surface available for interaction with cellmembranes or proteins, however, and smallercrystals provide a greater specific surface areathan large ones; in addition, their size mayfacilitate phagocytosis more readily, whilesome crystals may be too large to be phago-cytosed. The evidence in the literatureindicates that small CPPD crystals are moreinflammatory,9 though the lower limit of thisassociation is not known. It is possible that theacute samples also contained a number of verysmall crystals, difficult to identify, but ofimportance in the inflammatory response.Careful examination of these extracted CPPD,by HRTEM using magnifications that wouldallow easy visualisation of crystals as small as100 nm, failed to reveal identifiable M-CPPDcrystals of length <0-08 ,um. The absence ofsuch small crystals may be explained by thephenomenon of Ostwald ripening, in whichsmaller crystals are forming continuously in asupersaturated solution, and immediatelydissolving, their substance 'adding to' alreadyformed larger crystals; thus, the crystalpopulation contains very few very smallcrystals.

It may be that acute attacks are related to theincreased proportion of monoclinic to triclinicCPPD, also observed in inflammatory SF.Given the mechanism described above, ourdata could be explained in one of two ways.The monoclinic crystals found in pseudogoutattacks may be removed more quickly thantriclinic crystals. Alternatively, they may bederived from a different type of deposit thanthe triclinic crystals which predominated in theinterval fluids. The possibility of differences inthe origins of the crystals, which might alsoexplain the observed size differences, iscurrently under investigation, by examinationof deposits which are formed in different partsof the joint. The third possibility is thatM-CPPD crystallises during acute inflamma-tion, but the thermodynamics of aqueousCPPD formation in vitro make this veryunlikely.2'

In vitro, M-CPPD converts, over time, toT-CPPD,2' suggesting that it is more soluble.The reduction in the monoclinic to triclinicratio and the overall decrease in mean crystalsize in the interval samples may then reflectdissolution of the M-form, or phase trans-formation to the T-form as the acute episodesubsides. If both M- and T-CPPD reach theSF from joint tissues in large numbers, thusprecipitating an attack ofpseudogout, the moresoluble M-CPPD may dissolve or revert toT-CPPD by the time the attack subsides,though it is unlikely that this could happenin vivo in inflammatory SF.Whatever the reasons for the observed shifts

in the proportions of M-CPPD and T-CPPD,our results raise the possibility that monoclinicCPPD crystals are more inflammatory thantriclinic CPPD. Certainly, weight for weight,

829



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/


M-CPPD have a greater surface area thanT-CPPD, and it is the inorganic surface thatmediates interactions with cells and proteins.Indeed, crystal surfaces are known to vary intheir ability to bind proteins, including macro-molecules associated with cell activation andother mediators of crystal induced inflamma-tion. 2 13 22 Structural differences in the geo-metric and stereochemical groupings of atomswithin the crystal lattice generate variations inthe exposed atomic motif and net surfacecharge of crystals. Inflammatory crystals suchas CPPD and monosodium urate mono-hydrate, for example, have a high negative zetapotential. It is anticipated that molecularmodelling studies will highlight differences inthe surface chemistry of T-CPPD andM-CPPD and provide further insight into theimportance of surface structure in mediatingpseudogout.Another factor which affects the phlogistic

potential of crystals is the adsorption of SFproteins onto surfaces. IgG, for example,adsorbs onto crystal surfaces in a chargemediated process; IgG complexes arepositively charged under physiological con-ditions, and readily adsorb onto negativelycharged surfaces. The adsorption mechanismis such that the Fc portion of the IgG moleculeremains exposed and available for recognitionby Fc receptors on PMN cells.'2 It may beargued, therefore, that crystals with a netnegative surface charge have a greater affinityfor IgG and are more inflammatory. Sequentialbinding of other SF proteins may limit theinteraction of those species capable ofinducingan inflammatory response.23 Equally, theadsorption of organic material in such amanner as to limit the exposure of putativerecognition sites would also affect this process.This surface binding may be different fordifferent crystal forms.There exist some in vitro data to support the

theory that monoclinic CPPD is more inflam-matory,'0 " though the difficulties experiencedin preparing pure monoclinic or tricliniccrystals of the same crystal size, and of a sizesimilar to that of biogenic crystals, has made itdifficult to establish this more definitively.M-CPPD crystals injected into rat pleuralcavity have been found to induce a greater,sustained inflammatory response thanT-CPPD, as measured by exudate volume orPMN counts."' Watanabe et al," using a rat airpouch model, found that CPPD crystals (andespecially M-CPPD) induced greater numbersofpouch exudate PMN cells, and greater levelsof protease, prostaglandin E2 and tumournecrosis factor than calcium HAP, and thatmixed crystals (CPPD/calcium HAP) were nomore inflammatory than each crystal alone.

In conclusion, we have shown that theCPPD crystals associated with attacks ofpseudogout include a greater proportion ofmonoclinic crystals, and larger crystals, thanare seen in less inflammatory conditions. This

suggests that there may be a special, phlogisticpopulation of CPPD crystals, that originatesin different tissues, or is cleared in a differentway, than the more common populations ofsmaller crystals having a greater proportion ofT-CPPD.

This study was funded by the Arthritis and RheumatismCouncil.

1 Felson D T, Anderson J J, Naimark A, Kannel W, MeenanR F. The prevalence of chondrocalcinosis in the elderlyand its association with knee osteoarthritis; theFramingham study.I Rheumatol 1989; 16: 1241-5.

2 Resnick D, Niwayama G, Georgan T G, et al. Clinical,radiographic and pathological abnormalities in calciumpyrophosphate dihydrate deposition disease: pseudogout.Radiology 1977; 122: 1-15.

3 Mandel N, Mandel G. Calcium pyrophosphate crystaldeposition in model systems. Rheum Dis Clin North Am1988; 14: 21-40.

4 Kohn N N, Hughes R E, McCarty D J, Faires J S. Thesignificance of calcium pyrophosphate crystals in thesynovial fluid of arthritic patients; the pseudogoutsyndrome. 11. Identification of crystals. Ann Intern Med1962; 56: 738-45.

5 Gibilisco P A, Schumacher H R, Hollander J L, SoperK A. Synovial fluid crystals in osteoarthritis. ArthritisRheum 1985; 28: 511-5.

6 Terkeltaub R A, GinsbergM H. The inflammatory reactionto crystals. Rheum Dis Clin North Am 1988; 14: 353-64.

7 Schumacher H R, Gordon G, Paul H, et al. Osteoarthritis,crystal deposition and inflammation. Semin arthritisRheum 1981; 11: 116-9.

8 Halverson P B, McCarty D J. Identification of hydroxy-apatite crystals in synovial fluid. Arthnitis Rheum 1979; 22:389-95.

9 Schumacher H R, Fishbein P, Phelps P, Tse R, Krauser R.Comparison of sodium urate and calcium pyrophosphatecrystal size and other factors. Arthritis Rheum 1975;(suppl) 18: 783-93.

10 Roch-Arveiller M, Legros R, Chanaud B, et al.Inflammatory reactions induced by various calciumpyrophosphate crystals. Biomed Pharmacother 1990; 44:467-74.

11 Watanabe W, Baker D G, Schumacher H R. Comparisonof the acute inflammation induced by calciumpyrophosphate dihydrate, apatite and mixed crystals inthe rat air pouch model of a synovial space. J Rheumatol1992; 9: 1453-7.

12 Kozin F, McCarty D J. Protein binding to monosodiumurate monohydrate, calcium pyrophosphate dihydrateand silicon dioxide crystals. 1. Physical characteristics.JTLab Clin Med 1977; 89: 1314-25.

13 Gordon T P, Ahem M J, Reid C, Roberts-Thompson P J.Studies on the interaction of rheumatoid factor withmonosodium urate crystals and case report of coexistenttophaceous gout and rheumatoid arthritis. Ann Rheum Dis1985; 44: 384-9.

14 Kellgren J H, Lawrence J S. Radiological assessment ofOsteoarthritis. Ann Rheum Dis 1957; 16: 494-501.

15 McGill N W, Swan A J, Dieppe P A. Survival of calciumpyrophosphate dihydrate crystals in stored synovial fluids.Ann Rheum Dis 1991; 50: 939-41.

16 Swan A J, Heywood B R, Dieppe P A. Extraction ofcalcium-containing crystals from synovial fluids andarticular cartilage. J Rheumatol 1992; 11: 1763-72.

17 Howell D S. Diseases due to the deposition of calciumpyrophosphate and hydroxyapatite. In: KellyW N, HarrisE D, Ruddy S, Sledge C B, eds. Textbook of rheumatology.Philadelphia: W B Saunders, 1981; 1438-54.

18 Gatter R A. The importance of joint fluid analysis. In:Gatter R A, Schumacher H R, eds. A practical handbookof joint fluid analysis, 2nd edn, Vol 1. Philadelphia &London: Lea & Febiger, 1991; 1-5.

19 Bennett R M, Lehr J R, McCarty D J. Crystal shedding andacute pseudogout. An hypothesis based on a therapeuticfailure. Ann Rheum Dis 1976; 19: 93-7.

20 Doherty M, Dieppe P A. Acute pseudogout; "crystalshedding" or acute crystallisation? Arthritis Rheum 1981;24: 954-7.

21 Hearn P R, Russell G G. Formation of calciumpyrophosphate crystals in vitro; implications for calciumphosphate crystal deposition disease (pseudogout). AnnRheum Dis 1980; 39: 222-7.

22 Guerne P A, Terkeltaub R, Zuran B, Lotz M. Inflammatorymicrocrystals stimulate interleukin-6 production andsecretion by human monocytes and synoviocytes. ArthritisRheum 1989; 32: 1443-52.

23 Terkeltaub R, Curtiss L K, Tenner A J, Ginsberg M H.Lipoproteins containing apoprotein-B are a majorregulator of neutrophil responses to monosodium uratecrystals. _J Clin Invest 1984; 73: 1719-30.

830



j.com/

Ann R

heum D


ownloaded from

http://ard.bmj.com/

Documents

dihydrate crystals, pseudogout