I Clin Pathol 1995;48:229-235

Histomorphometric classification ofpostmenopausal osteoporosis: implications for themanagement of osteoporosis

M T A Rehman, J A Hoyland, J Denton, A J Freemont

AbstractAims-To define and group static anddynamic iliac crest histomorphometricparameters in women with establishedosteoporosis.Methods-Iliac crest biopsy specimensfrom 146 white women were sectionedundecalcified and examined using imageanalysis.Results-Five distinct groups weredefined on the basis of histomorpho-metric changes in cell function: group 1,decreased osteoblastic and osteoclasticactivity; group 2, decreased osteoblasticand increased osteoclastic activity; group3, increased osteoblastic and osteoclasticactivity; group 4, no bone surface cellactivity; and group 5, apparently normalosteoblastic and osteoclastic activity.Conclusions-Five distinct subgroups ofpatients with postmenopausal osteoporo-sis can be defined based on changes inbone cell function. Defining cellular dys-function in this way may be important fortailoring treatment regimens to the needsof individual patients.(7 Clin Pathol 1995;48:229-235)

Keywords: Histomorphometric parameters, osteoporosis

UniversityDepartment ofMedicine, HopeHospital, SalfordM T A RehmanDepartment ofOsteoarticularPathology, MedicalSchool, University ofManchester,ManchesterJ A HoylandJ DentonA J FreemontCorrespondence to:Professor A J Freemont,Department ofPathological Sciences,Stopford Building,University of Manchester,Manchester M13 9PT.

Accepted for publication7 June 1994

Postmenopausal osteoporosis is the common-est type of osteoporosis in clinical practice.Despite increased awareness of the conditionand its importance, most women still presentwith the established disorder, resulting in allthe clinical problems of managing patientswith a symptomatic disorder but without a

satisfactory treatment regimen. Patients are

currently grouped on the basis of clinicalcharacteristics. Response to treatment isunpredictable perhaps because current treat-ment regimens are unable to normalise themicrostructural changes associated withosteoporosis. We hypothesise that the failureof single treatment regimens might also bebecause of the heterogeneous nature of thecellular changes in this group.The clinical syndrome of osteoporosis is a

common manifestation of a variety of disor-ders. At the cellular level, the syndrome can

be broadly regarded as an imbalance or

uncoupling of osteoblastic bone depositionand osteoclastic bone resorption.' Theoret-ically, this imbalance may take several formsranging from an increase in both osteoclasisand osteoblastic activity, with a greater

increase in the former,2 to a decrease in both,with a greater reduction in osteoblastic activ-ity.34

Studies of postmenopausal osteoporosissuggest that there is heterogeneity within thisgroup. Some studies have shown that "uncou-pling of bone" is caused by increased boneresorption.' This may lead to perforation ofbone trabeculae, occasionally with an increasein bone fragility greater than that expectedfrom the amount of bone lost.34 Other studiessuggest that the predominant abnormality isthe failure of osteoblasts to fill resorbed

6 7spaces. There is no overall consensus con-cerning the nature of bone remodellingchanges in postmenopausal osteoporosis.Whilst most studies on subjects with idio-pathic osteoporosis have demonstrated lowbone turnover,38-11 Recker et al I reported tra-becular microarchitectural changes associatedwith normal bone remodelling and Wands etal'2 demonstrated that women with "vertebralcrush syndrome" continue to have an inap-propriately prolonged phase of increased boneremodelling. Generally, the same principlesapply to cortical and trabecular bone loss.13 14There is no consensus as to the overall rela-tion between clinically defined patient groupsand the changes in bone cell function andmatrix that typify the syndrome.

This perplexing diversity of opinion mayhave arisen because the cellular changes areheterogeneous. Recognition of specific para-meters of bone turnover and the nature of thepredominant cellular abnormalities in an indi-vidual or group of patients may be a moreuseful way to define osteoporosis, permittingselection of optimum therapy to improve or atleast to reduce further deterioration of thebiomechanical integrity of the skeleton. Totest this hypothesis, we have used histomor-phometry to obtain data on bone cell activityand function in 146 postmenopausal womenwith established osteoporosis.

MethodsAll patients were recruited from either HopeHospital, Salford, or the Manchester RoyalInfirmary, Manchester. Women were regar-ded as- postmenopausal if they had beenamenorrhoeic for more than 12 months.Oestradiol and gonadotrophin concentrationswere measured to confirm menopausal status.Diagnosis of established osteoporosis wasdetermined radiologically by the presence ofat least one non-traumatic vertebral crush

229

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Rehman, Hoyland, Denton, Freemont

fracture or a lumbar spine (L2-L4) bonemineral density (BMD) less than 2 SD belowpeak bone mass (assessed by DXA). All sec-ondary causes of osteoporosis were excludedfollowing appropriate investigations: deter-mination of parathyroid hormone, serum cal-cium, phosphate, and alkaline phosphataseconcentrations, thyroid function tests, andimmunoglobulin electrophoresis.To distinguish between postmenopausal

and senile osteoporosis, an arbitrary upper agelimit was set as recommended by Riggs andMelton.'5 Only women between 45 and 75years of age were studied and only thosepatients with an adequate amount of bothcortical and trabecular tissue in the iliac crestbone biopsy specimen were included.

All patients were given oral dimethyl-chlor-tetracycline (10-15 mg/kg body weight) individed doses over 12 hours 15-18 daysbefore bone biopsy. This dose schedule wasrepeated after an interval of 10 days. Earlymorning urine specimens were collectedbefore each tetracycline dose schedule and aday later to confirm compliance and adequateabsorption of tetracycline. Biopsy was per-formed within four to seven days of the lasttetracycline dose. A full blood count andcoagulation profile were performed beforebiopsy to exclude any bleeding disorders.

Bone biopsy was performed under localanaesthesia. Full thickness transiliac bonebiopsy specimens were taken, 2 cm below andbehind the anterior superior iliac spine,'6using a Lalor trephine biopsy needle. None ofthese patients had previously undergonebiopsy at that site. Haemostasis was routinelysecured by a couple of superficial stitchesand a pressure bandage was applied for 12hours.

Biopsy specimens were fixed in absolutealcohol and processed three times in LR whiteresin monomer (London Resin Co.), the lasttwo taking place under reduced pressure.Resin polymerisation was carried outovernight at 60°C. Twenty seven 5 um stepserial sections were cut through each blockwith a Tungsten tipped knife on a LKBpowered microtome. Groups of three sectionswere stained with toluidine blue. (pH 4 2), or

using the modified Giemsa or Von Kossatechniques. For fluorescence microscopy,20,um unstained sections were cut at differentlevels throughout the block.A semiautomated VIDS II image analyser

was used to derive area, length, number, andtwo point analyses on each of nine toluidineblue stained sections by a single experiencedoperator. The mineral apposition rate(MAR), however, was measured using an eye-piece graticule in the fluorescence microscopeto perform two point analysis at four regularlyspaced positions along each front delineatedby double fluorescent label.'7 Microscopicmagnification varied from x40 for area mea-surements to x 160 for two point analyses andcell counting. Where data were expressed asratios of similar parameters, conversion ofmeasured two-dimensional to derived three-dimensional numerics required no change in

the ratio, as defined by the theorem ofDeLesse,'8 but where spherical geometry wasinvolved, the stereologic constant n/4 wasused to convert two-dimensional measure-ments into three-dimensional measure-ments. 19When tetracycline labelled sections were

used, the length of mineralising surface wasdivided into two components-those with twoand those with only one label. The latter rep-resents regions of the bone surface that eitherfinished or started mineralising in the intervalbetween administration of the two labels. Todefine certain parameters of bone formationmore accurately, a corrected value of thelength of the mineralising surface was con-structed, taking account of single labelling,using the standard method of defining thetrue active mineralising surface as the doublelabelled plus half the single labelled surfaces(dLs+ '/2sLs).

Toluidine blue stained sections were usedto identify the osteoid surface, defined as anunmineralised surface at least 3 ,um thick.Osteoid width (O.Wi) was then measured atfour equidistant sites on each osteoid seam.All measurements were based on the mean of20 separate fields using the standard constantfor section obliquity (1I2).

Parameters were measured using standardtechniques17 19-21 and defined according to theterminology proposed by Parfitt et al.22 Thestatic and dynamic parameters obtained bymeasurement or calculation are presented intable 1.

ResultsAll histomorphometric parameters are pre-sented as a mean z-score. A z-score for anindividual value is the number of standarddeviations by which the value differs from themean of normal age and sex matched controlsfor the local population.23 Using the z-score,data from different age groups with differentnormal values can be pooled. The mean z-score can be calculated for a group as well asan individual. A difference of more than 3 z-score units (that is, 3 SD above or below nor-mal) between the patient groups and controlsis regarded as abnormal; any parameter with az-score between 2 and 3 is classified as mod-erately abnormal and is only marginally orslightly abnormal if the z-score is between 1and 2. The parameters are regarded as normalif the z-score is less than 1.

Patients could be divided into five distinctsubgroups based on the pattern of overallchanges in certain histomorphometric para-meters (z-score, factor and cluster analysis).Osteoblast parameters included osteoblastsurface, osteoid surface (dLs+ ½/2sLs) minerali-sation lag time, bone formation rate, andosteoclast parameters included eroded sur-face, osteoclast number, osteoclast surface,and erosion period (figure). The numeric datafor all the patients are presented in tables 2to 5.

Group 1 comprised 85 (58%) patients.Trabecular bone in this group was charac-

230

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Histomorphometric classification ofpostmenopausal osteoporosis

Table 1 Histomorphometric parameters: methods of measurement or calculation

Parameters Abbreviation Methods

Activation frequency* Ac.f (days) 1/(EP+FP+QP)Adjusted appositional rate* Aj.AR (um/day) (MS/OS)XMARBone formation rate* BFRIBV (% year) (MARXMSxBS/BV) 100Bone surface BS (um) LMBone volume BV (%) AM (BV/TV) x 100Cortical thickness C.Th (um) 2PMXn/4Cortical volume CV (mm3) AM (CV/TV) X 100Erosion depth E.De (uum) 2PMActive erosion period* a.EP (days) (a.ES/OS) x FPErosion rate* ER (mm/year) E.De/a.EPX 1/1000X365Eroded surface p.(absolute) ES (um) LMActive eroded surface a.ES (um) LMEroded surface 4(relative) ES/BS (%) (ES/BS) x 100Formation period* FP (days) W.Th/Aj.ARLabel interval LP (days) time between labels

double label surface dLs (uum) LMinter label distance iLs (um) LMsingle label surface sLs (um) sLs+dLstotal label surface tLs (um) 2PM

Mineralisation apposition rate (cortical) MARc (um/day) iLd/LPMineralisation apposition rate (trabecular) MARt ((uum/day) iLd/LPMineralisation lag time* Mlt (days) O.Th/Aj.ARMineralising surface (corrected) MS.c (um) dLs+ /2sLsMineralising surface (double) Msd (%) dLs/tLsMineralising surface total (absolute) MSt (um) tLsMineralising surface absolute (bone

referent) MS/BS (%) MS/BS (dLs+'/2sLs)Mineralising surface (osteoid referent) MS/OS (%) MS/OS (sLs+dLs)Number of osteoclasts (cortical) N.Oc.c (/mm2) 10-2(NOc.c/TV.c)Number of osteoclasts (trabecular) N.Oc.t (/mm2) 10-2(NOc.t/TV.t)Number of osteoclasts (subcortical) N.Oc.s (/MrnM2) 10-2NOc.s/LMOsteoblast surface Ob.S (%) Ob.S/BSOsteoclast surface Oc.S (%) Oc.S/BSOsteoid surface (absolute) OS (um) LMOsteoid surface (relative) OS/BS (%) (OS/BS)x 100Osteoid thickness O.Th (um) OWiXir/4Osteoid width O.Wi (um) 2PMOsteoid volume OV (%) OV/BVQuiescent period* QP (days) (QS+OS)FPQuiescent surface* QS (um) BS-(ES+OS)Trabecular number TN (/mm2) numbers/areaTrabecular separation TS (um) 2PMTrabecular thickness T.Th (um) 2PMTissue volume TV (mm2) AMWall thickness W.Th (um) 2PMX-rr/4Wall thickness (cortical) W.Th.c (um) 2PMXir /4Wall width W.Wi (um) 2PM

*Calculated parameters; LM, length measurement; AM, area measurement;2PM, two point measurement.

terised by a low bone volume, reduced wallthickness, and a decreased appositional ratesecondary to diminished osteoblast cellnumber and activity. These parameters pro-gressively decreased with increasing age. Theosteoid surface, volume and thickness as wellas mineralising surface (corrected) (measuredas double label surface + ½/2 single label sur-face) were normal. The mineralising surface,whether measured as double label surface plussingle label surface (more than - 2 0) or as thesum of double and total label surfaces (morethan-3-5), was moderately to notablyreduced. Trabecular thickness and numberdecreased with increasing age in this groupand hence trabecular separation increasedwith age (+ 1 0 at 45-50 years to more than+3*0 at 50 years or over). Trabecular struc-ture was largely fragmented-that is, trab-eculae were perforated, and appeareddiscontinuous in sections (82%); however, a

Table 2 The characteristics ofpatients in the differenthistomorphometric groups

Group Total No. (%) Age (SD) in years

1 76 (56) 59 (9)2 34 (25) 67 (8)3 15 (11) 56 (10)4 4 (3) 48 (4)5 7 (5) 51 (8)

small proportion was either normal (4%) or ofthe filigree type (14%) that is, the trabeculaewere continuous but appeared generally thinand frail in tissue sections. Both the frag-mented and filigree types of trabeculae areabnormal and result in reduced bonestrength. Recognition of the abnormal patternof trabeculae has important therapeutic rele-vance. The structure of trabeculae is of primeclinical importance; in the fragmented typethe trabecular scaffold is discontinuous and,therefore, is theoretically less amenable tobone forming therapy and disproportionatelyincreases the risk of further fracture. On thewhole, patients in group 1 had a noticeablyreduced rate of bone formation (-1 9 to-3 8), as expressed by the ratio of bone for-mation to bone volume (BFR/BV), andalthough the bone formation period was nor-

mal in younger patients (z-score +0<1 forthose aged from 45 to 50 years), it decreasedprogressively with age (-2 1 for those agedfrom 60 to 65 years). Mineralising surfaceswere noticeably reduced but the mineralisa-tion lag time was only marginally increased(+0 4 to + 1 1), suggesting that the predomi-nant abnormality is impairment of the initia-tion of bone formation. The activationfrequency was also reduced.

Analysis of cortical bone revealed veryreduced wall thickness and a moderate

Table 1 abbreviations:

MAR= mineral apposition ratedLs = double labelled surfacesLs = single labelled surfaceTb.N = trabecular numberTb.Th = trabecular thicknessW.Th = wall thicknessTBV = trabecular bone volumeOS/BS = osteoid surfaceMARt = mineral apposition rate in trabecular boneBFR/BS = bone formation rate at bone surfaceBER/BV = bone formation rate to bone volumeAj.AR = adjusted apposition rateMlt = mineralisation lag timeFP = formation periodAc.f = activation frequencyES/BS = eroded surfacesOcS/BS = osteoclast surfaceEP = erosion periodCt.Th cortical thicknessCV = cortical volumeW.Thc = cortical wall thicknessMAR c = mineral apposition rate in cortical boneE.De = erosion deptha.Ep = active erosion periodER = erosion rateES = eroded surface (absolute)a.ES = active eroded surfaceES/BS = eroded surface (relative)LP = label intervaliLs = inter label distancetLs = total label surfaceMS.c = mineralising surface (corrected)MSd = mineralising surface (double)MSt = mineralising surface total (absolute)MS/BS = mineralising surface absolute (bone referent)MS/OS = mineralising surface (osteoid referent)NOc.c = number of osteoclasts (cortical)NOc.t = number of osteoclasts (trabecular)NOc.s = number of osteoclasts (subcortical)Ob.S = osteoblast surfaceOc.S = osteoclast surfaceOS = osteoid surface (absolute)OS/BS = osteoid surface (relative)O.Th = osteoid thicknessO.Wi = osteoid widthOV = osteoid volumeQP = quiescent periodQS = quiescent surfaceTN = trabecular numberTS = trabecular separationTV = tissue volume

231

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Rehman, Hoyland, Denton, Freemont

Table 3 Static histomorphometric parameters in trabecular bone

AgeGroup (years) No. TBV W.Th OSIBS OV O. Th Ob.S MS MS MSd ES OcN OcS ThTh TN TS

1 45-50 21 -1-6 -1-4 -0-2 -1 2 -0 7 -1-2 -0 4 -2-6 -4-1 -0-1 -0.2 +0 3 -0-2 -0-7 +1.050-55 33 -2-9 -2-5 -0 5 -1-2 -0-6 -1-2 -0-4 -2-4 -3-7 +0 3 +0.3 -0 5 -0 4 -1 9 +3-355-60 17 -3-1 -3 5 -0 4 -0.9 -0-6 -0 9 -0-9 -2-5 -3-1 +0 0 +0 1 -0 4 -0-8 -1 9 +4-160-65 4 -2-8 -33 -0 3 -009 -0.5 -0.9 -1 0 -2-3 -3 0 +0 1 0 0 -0-4 -0-8 -1 7 +3-165-75 1

2 45-50 150-55 3 -1-7 -2-3 +0 1 -1-4 -0.6 -1-7 -1 1 -1-5 -3-6 +2-7 +2-5 +1-4 -0 4 -0-6 +1 155-60 4 -2-4 -2-8 -0 3 -1-6 -0-8 -2-2 -1-5 -1-7 -3 0 +3-2 +2.0 +2-3 -1 0 -0-8 +3-460-65 7 -3-6 -2-8 +0 0 -0 7 -0.5 -2-1 -1 9 -1-8 -2-9 +3-1 +2-4 +2-1 -1-3 -1-3 +3-765-70 8 -3-8 -2-5 -0-2 -0 9 -0.9 -1-9 -1-3 -1 9 -3-1 +3 0 +3-3 +1-6 -1-1 -1 1 +3-670-75 11 -3-7 -2-4 -0-2 -1 0 -0 7 -2-0 -1-6 -1-8 -2-4 +3-4 +3-4 +1-8 -1-2 -1 0 +4-0

3 45-55 5 -1 1 +00 +1 0 +07 -0-2 +07 +0 1 -04 -04 +1-5 +09 +1-3 -07 +0.0 +1-355-60 5 -1-8 +0 3 +1-3 +0-2 -0-2 +0-8 -0-1 -0 4 -0 3 +1 1 +1-4 +1-3 -1-6 -0-7 +2-665-75 5 -1-9 -0-6 +1 1 +00 +0.0 +09 +04 -05 -0-6 +1-3 +29 +14 -1-5 -03 +2-9

4 45-50 4 -2-7 -1-6 -3 3 -3-8 -3-2 -0.7 +0-1 -1-2 -0-2 -1-8 -2-2 -1-5 -1-5 +0 0 +1-4

5 55-60 3 -0 1 +00 +0-2 +0 1 -0 1 +00 -04 -03 +0-2 +0-2 +01 +0-2 +03 -0.1 +00

Units = z-score, above (+) or below (-) the mean.

m.E

Basis f;-.. -lhctrir classificationiPatterns of significant histomorphometric parameters defining osteoporotic

increase in eroded surfaces, but the numbersteoclasts of osteoclasts, and cortical thickness, corticalrface volume, and osteoid surfaces were normal.e Although the number of osteoclasts in the

subcortex was moderately elevated, the ero-sion rate and the erosion period were withinthe normal range. It was technically difficultto assess the rate of activation frequency in thecortex.Group 2 comprised 32 (22%) patients.

These patients had reduced trabecular bonevolume and wall thickness, as in patients in

SXg//1 el group 1, but the number of osteoblasts was

moderately to noticeably reduced (-1.7 to-2.2). The rate of bone formation (BFR/BV)was reduced but less so than in group 1

patients and the formation period was withinro uP the normal range. In contrast to group 1,

however, group 2 patients had a mildly tomoderately prolonged mineralisation lag time(+1*3 to +2-1). The striking feature of tra-becular bone in group 2 patients was the pres-

ence of a noticeable increase in the number ofosteoclasts (+2-0 to +3-4) associated with a

similar increase in the number of eroded sur-

faces (+2-7 to +3.4), but the erosion cavitiesroLip ^ were relatively shallow (-1-0 to -1-2).

Although the erosion rate was marginallydecreased (-0-6 to -1 2), overall resorptionwas increased because the erosion period was

r/MMA. also marginally longer (+0 9 to +1I8). On thewhole, the rate of bone remodelling, asassessed by the activation frequency, was nearthe lower limit of the normal range. Thetrabecular pattern in most of the cases was

fragmented and was of the filigree type in theGroup 4 remainder.

Cortical bone thickness was slightlyreduced in group 2 patients when comparedwith those in group 1 and was associated witha mild to moderate increase in the number ofosteoclasts both in the cortex and subcortex.There was a moderate increase in the numberof eroded surfaces but the rate of erosion andthe erosion period were normal, except inpatients aged between 70 and 75 years, where

c subgroups. they were marginally increased.

OS/BS = Osteoid/bone surface OcN = Number of osiObS = Osteoblast surface OcS = Osteoclast surMS = Mineralising surface ES = Eroded surfac4MIt = Mineralisation log time EP = Erosion periocBFR = Bone formation rate

232

e_

i.- '..,..-7'

1-1.1 LLIT.'.

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Histomorphometnic classification ofpostmenopausal osteoporosis

Table 4 Dynamic histomorphometric parameters in trabecular bone

AgeGroup (years) No. MARt Aj.AR Mit FP BFRIBV E.De EP ER Ac.f1 45-50 21 -1-9 +1 4 +0 7 +0 1 -3-8 -1 0 -0-3 -0-5 -0-7

50-55 33 -2 1 -1-6 +0 4 -0-8 -2-6 -1-4 -0-7 -0 9 -1-655-60 17 -2 1 -1-6 +0 4 -1-7 -2.2 -11 -0-4 -0-6 -1-760-65 4 -2.4 -1-3 +0 5 -2-1 -1-9 -0.90 06 -0-8 -1 565-75 1

2 45-50 1 -1 0 +1 350-55 3 -1.6 -1.8 +2-1 -0-7 -2-0 -1-5 +0 9 -0.7 +0 255-60 4 -1 9 -2-2 +1 9 -0-3 -1-8 -1-7 +1 8 -0-6 -0.760-65 7 -22 -1 5 +20 -1-2 -1 5 -1 3 +09 -09 -0665-70 8 -1.9 -1.8 +1 3 -0-6 -1 0 -1 2 +1-4 -1-2 -0.570-75 11 -2-0 -2-1 +1-8 -0 1 -1 0 -0-6 -0 1

3 45-55 5 -1-2 -09 +0-8 +1.0 -0-6 +1 1 +05 +1 7 +0-655-60 5 -1-8 -07 +04 +1 0 -04 +1 2 +06 +1 5 +0465-75 5 -1-7 -0.7 +0 7 -0 1 -0-3 +0 7 +0 4 +1 9 +0 4

4 45-50 4 -1-6 -14 -1.7 +0 1 -1.1 -1-2 +03 -0-8 -1-7

5 55-60 3 +0 1 +03 +03 -0 1 +0 1 +0 1 +0 1 -0-2 -03

Units = z-score, above (+) or below (-) the mean.

Table 5 Static and dynamic histomorphometric parameters in cortical bone

AgeGroup (years) No. C.Th CV C.Wlh C.OS N.Oc.c ES.s N.Oc.s ER EP

1 45-50 21 -0-5 -0-6 -1-4 0.0 +0-6 +2-1 +2-9 +1-3 +0 250-55 33 -0 4 -0-7 -3-1 -0 3 +0 3 +1-8 +1-2 +1-4 +0 355-60 17 -04 -1 0 -2-9 -0-2 +0-2 +1-4 +1 1 +0-8 -0 160-65 4 -0 4 -0-5 -2-1 -0 1 +0.3 +1-5 +1-2 +0 7 +0-8

2 45-50 3 -1-5 -20 -20 -07 +09 +1 6 +24 +07 -0350-55 4 -1-8 -27 -22 -1 1 +13 +20 +1 9 +06 +0755-60 7 -2-4 -3-1 -2-1 -1 0 +0.8 +1-5 +1 9 +0 5 -0.460-65 8 -2-7 -3 0 -1 9 -1 0 +0.9 +1 7 +1 9 +0 1 -0-865-70 11 -2-3 -2-9 -2-0 -0-8 +1.0 +1-8 +1 6 1 0 +0 670-75

3 45-55 5 -0-4 -1-2 -0.4 +0-2 +1.0 +1-2 +1-2 +0 3 +0-655-60 5 -0-8 -1-7 -0-6 +0 5 +1.2 +1-6 +1-7 +0-2 +0 165-75 5 -0 9 -1-5 -0 5 +0 5 +1.2 +1-6 +1 7 +0 4 +0 1

4 45-50 4 +0 1 -0-6 -2-9 -2-2 -0.4 -1-3 -3-7 -1-6 -1-7

5 55-60 3 00 -0-2 -0-2 -03 -0 1 - - - -

Units = z-score, above (+) or below (-) the mean.

Group 3 comprised 16 (1 1%) patients whohad evidence of increased osteoid surface intheir trabecular bone which was associated witha moderately reduced appositional rate. Theosteoclast cell number and activity wereincreased in both the cortical and trabecularbone and, as such, the overall picture was similarto that of the common form of primary hyper-parathyroidism seen in middle-aged women.24 25Although the activation frequency in the trabec-ular bone was normal, the erosion rate wasmoderately increased. These patients all hadnormal serum parathyroid hormone, serumcalcium, phosphate, and alkaline phosphataseconcentrations.Group 4 comprised seven (5%) patients

and was characterised by a noticeable reduc-tion in both osteoblast and osteoclast numberand activity in trabecular bone. This was asso-ciated with a moderately reduced rate of acti-vation. Cortical wall thickness and theamount of cortical osteoid were reduced andwere associated with an equally profoundreduction in osteoclasis, particularly in thesubcortical region. The eroded surfaces, rateof erosion, and the erosion period in the cor-tex were moderately reduced but cortical vol-ume and cortical thickness were normal.

Group 5 comprised six (4%) patients whowere normal on histomorphometry despiteradiological evidence of osteoporosis.

DiscussionA semiautomated histomorphometric tech-nique was used to examine iliac crest bonesamples from a large cohort of patients (agedfrom 45 to 75 years) with established post-menopausal osteoporosis. These patients weredivided into five distinct subgroups based onhistomorphometric data.

Histomorphometric subgroups, basedmainly on the differential rate of boneturnover, have been described previously.9 26 27Most histomorphometric parameters havebeen described separately in previous studiesbut we believe that much of the debate on thetype of histomorphometric heterogeneity seenin postmenopausal osteoporosis results fromfailure to recognise the full spectrum ofchanges that can occur in one group ofpatients, an aspect highlighted by this study.The most important findings of this study arethe wide variation in the number and activityof bone surface osteoblasts and the lack of anycorrelation between osteoblastic and osteo-clastic activity. Variation in these parameterscan be illustrated by examining each individ-ual group in more detail. The five subgroupscan be generally defined as follows: group 1,decreased osteoblastic activity and normalosteoclasis in trabecular bone and decreasedosteoblastic activity and normal or focallyincreased osteoclasis in cortical bone;

233

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Rehman, Hoyland, Denton, Freemont

group 2, decreased osteoblastic activity andincreased osteoclasis in both trabecular andcortical bone; group 3, marginally increasedosteoblastic activity and increased osteoclasisin both trabecular and cortical bone; group 4,decreased osteoblastic and osteoclastic activ-ity in both trabecular and cortical bone; andgroup 5, no apparent cell abnormality.The histomorphometric changes in groups

1 and 2 correspond to those described in thetwo main clinical subclasses of osteoporosis,namely postmenopausal (type 1) and senile(type 2) osteoporosis,l-4 the major differencebeing that none of our patients fulfilled theclinical criteria for senile osteoporosis. Thisillustrates that the changes previouslydescribed as characteristic of senile osteoporo-sis are present in one subgroup of patientswith postmenopausal osteoporosis. Osteoblastactivity was reduced in both groups 1 and 23 4 6 7but osteoclastic bone resorption was increasedin group 2. Although not identical, group 5 inthis study corresponds in some respects to thepatients described by Recker and Kimmel'and group 3 is similar to those patientsdescribed by Melsen et al24 and Christiansenet al.25

This is, of course, a very general approachto analysis of the data. In cellular termsosteoblastic "activity" is a combination of anumber of parameters (cell number, individ-ual cell function, etc.) which can be looked atmore closely based on the data collected. Forinstance, if the bone changes in group 1 areexamined more closely, the histomorphomet-ric data on the 76 younger postmenopausalwomen in this group show that the number ofosteoblasts is reduced together with their abil-ity to synthesise matrix. There is also a reduc-tion in the amount of trabecular bone with anassociated loss of trabecular integrity, suggest-ing that a local increase in bone destructionoccurred leading to penetration of the trabecu-lar plates. Despite fragmentation of trabecu-lae, which occurred in 82% of patients, therewas no evidence of an increase in osteoclasticactivity, suggesting that the pattern of cellularimbalance was different originally and that thepattern of cell changes that ultimately lead toclinically established osteoporosis may varywith time.'0

Histomorphometric data are important forthe management of patients with post-menopausal osteoporosis. In group 1 patientsosteoblast number and activity are reduced.This is associated with a normal number ofpoorly functioning osteoclasts. Ideally, there-fore, therapy should be directed towardsincreasing osteoblast recruitment and activity,whilst initially trying to prevent a concomitantincrease in osteoclasis.A similar exercise can be applied to the

other groups. The extent of osteoblast dys-function in group 2 is similar to that in group1; however, group 2 patients have anincreased number of osteoclasts. Although, asin group 1, the activity of individual osteo-clasts was suppressed, osteoclasts were able toerode an increased proportion of trabecularsurfaces, by virtue of their large numbers. The

activation frequency was marginally reduced.Cortical bone parameters in group 2 patientsexhibit similar changes to those in trabecularbone.The most important consequence of these

observations is that, at presentation, everypatient can be regarded as having a specificcellular abnormality. Ideally, each patientshould then receive a tailored treatment regi-men. For instance, patients in group 1 arelikely to respond positively to therapeutic regi-mens which promote bone formation, whilethose in group 2 should respond morefavourably if potent antiresorptive agents areincluded.

These data should also be considered fromtwo other perspectives. Firstly, they areexpressed relative to age and sex matched"controls". Should fracture risk be an impor-tant end point in osteoporosis, all data(including age and sex matched normals)should be given relative to peak bone massand its associated parameters. Were this thecase, the negative z-scores in the osteoporoticpatients would be even larger. Age and sexmatched controls have been used here tostress how women with postmenopausalosteoporosis differ from women of the sameage without the syndrome because it could beargued that treatment might be aimed at"normalising" a defect rather than at trying toreattain the ideal, as represented by the peakparameters.

Secondly, the approach adopted here forthe analysis of osteoporosis is quite differentto that used by other authors. There is a long-standing discussion in the literature of thevalidity of data analysis in patients with type 1osteoporosis.28 These researchers classifiedtheir patients clinically (mainly based on ageand clinical characteristics) and thenattempted to identify a unified histomorpho-metric defect to explain bone loss. We believethat the approach adopted in this study is ofgreater clinical and therapeutic value.

1 Riis BJ, Rodbro 0, Christiensen C. The role of serum con-centration of sex steroids and bone turnover in the devel-opment and occurrence of postmenopausal osteoporosis.Calcific Tissue Int 1986;38:318-22.

2 Eriksen EF, Hodgson SF, Eastell R, Cedal SL, O'FallonWM, Riggs BL. Cancellous bone remodelling in type I(postmenopausal) osteoporosis: quantitative assessmentof rates of formation, resorption, and bone loss at tissueand cellular levels. 7 Bone Miner Res 1990;5:311-19.

3 Parfitt AM. Bone remodelling in the pathogenesis of osteo-porosis. Medical Times 1981;109:80-92.

4 Parfitt AM. Age-related structural changes in trabecularand cortical bone: cellular mechanism and biochemicalconsequences. a) differences between rapid and slowbone loss. b) localised bone gain. Calcific Tissue Int1984;36(Suppl):S 123-36.

5 Heaney RP, Recker RR, Saville PD. Menopausal changes inbone remodelling. J Lab Clin Med 1978;92:964-70.

6 Kleerekoper M, Villaneuvar AR, Stanciu J, Rao DS, ParfittAM. The role of three dimensional trabecular micro-structure in pathogenesis of vertebral compression frac-tures. Calcif Tissue Int 1985;37:594-7._

7 Vesterby A, Gundersen HJG, Melsen F. Star volume ofmarrow space and trabeculae of first lumbar vertebra:Sampling efficiency and biological variation. Bone 1989;10:7-13.

8 Parfitt AM. Bone remodelling: relationship to the amountand structure of bone, and pathogenesis and preventionof fractures. In: Riggs BL, Melton LJ, eds. Osteoporosis,etiology, diagnosis and management. New York: RavenPress, 1988:45-93.

9 Whyte MP, Bergfeld MA, Murphy WA, Avioli LV,

234

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from

Histomorphometric classification ofpostmenopausal osteoporosis

Teitelbaum SL. Postmenopausal osteoporosis. A het-erogenous disorder as assessed by histomorphometricanalysis of iliac crest bone from untreated patients. AmJMed 1982;72:193-202.

10 Meunier PJ, Sellami S, Briacorn D, Edouard C.Histological heterogeneity of apparently idiopathicosteoporosis. In: DeLuca HF, Frost HM, Jee WSS,Johnston CC, Parfitt AM, eds. Osteoporosis. Recentadvances in pathogenesis and treatment. Baltimore, MD:University Park Press, 1981:293-301.

11 Recker RR, Kimmel DB. Changes in trabecularmicrostructure in osteoporosis occur with normal boneremodelling dynamics. J Bone Miner Res 1991 ;6(Suppl):S225.

12 Wand JS, Green JR, Hesp R, Bradbeer JN, Sambrook PM,Smith T, et al. Bone remodelling does not decline aftermenopause in vertebral fracture osteoporosis. BoneMiner 1992;17:361-75.

13 Keshawarz NMS, Recker RR. Expansion of medullarycavity at the expanse of cortex in postmenopausal osteo-porosis. Metab Bone Dis Relat Res 1984;5:223-8.

14 Brown JP, Delmas PD, Arlot AM, Meunier PJ. Activebone turnover of the cortico-endosteal envelope in post-menopausal osteoporosis. j Clin Endocrinol Metab 1987;64:954-9.

15 Riggs BL, Melton U III. Evidence for two distinct types ofosteoporosis. Am JfMed 1983;75:899-901.

16 Melsen F, Moselilde L. The role of bone biopsy in diag-nosis of metabolic bone disease. Orthop Clin North Am1981;12:571-602.

17 Parfitt AM, Mathews CHE, Villaneuva AR, KleerekoperM, Frame B, Rao DS. Relationship between surface vol-ume and thickness of iliac trabecular bone in aging andosteoporosis. J Clin Invest 1983;72: 1396-409.

18 DeLesse A. Procede mecanique pour determiner la com-position des roches. Ann Mines 1884;13:339.

19 Frost HM. Histomorphometry: techniques and interpretation.In: RR Recker, ed. Boca Raton, FL: CRS Press,1983:109-32.

20 Eriksen EF, Hodgson SF, Eastell R, Cedal SL, O'FallonWM, Riggs BL. Cancellous bone remodelling in type I(postmenopausal) osteoporosis: quantitative assessmentof rates of formation, resorption, and bone loss at tissueand cellular levels. J Bone Miner Res 1990;5:311-19.

21 Compston JE, Mellish RWE, Croucher P, Newcombe R,Garrahan NJ. Structural mechanisms of trabecular boneloss in man. Bone Miner 1989;6:339-50.

22 Parfitt AM, Drezner MK, Glorieux FH, Kanis JA,Maluche H, Meunier PJ, et al. Bone histomorphometry:Standardisation of nomenclature, symbol, and units.J Bone Min Res 1987;2:595-610.

23 Rehman MTA, Hoyland JA, Denton J, Freemont AJ. Agerelated histomorphometric changes in bone in normalBritish men and women. Jf Clin Pathol 1994;47:529-34.

24 Melson F, Mosekilde L. The role of bone biopsy in diag-nosis of metabolic bone disease. Orthop Clin North Am1981;12:571-602.

25 Christiansen P, Steniech T, Brocketedt H, Mosekilde L,Hessove I, Melsen F. Primary Hyperparathyroidism:Iliac crest cortical thickness, structure and remodellingevaluated by histomorphometric methods. Bone 1993;14:755-62.

26 Arlot ME, Delmas PD, Faculte A, Chappard D, MeunierPJ. Impaired osteoblast function in osteoporosis: com-parison between calcium balance and dynamic histo-morphometry. Osteoporosis Int 1990;1:41 -9.

27 Kimmel DB, Recker RR, Gallaghar JC, Vaswani AS, AloiaJF. A comparison of iliac crest bone histomorphometricdata in postmenopausal osteoporosis normal subjects.Bone Miner 1990;11:217-35.

28 Parfitt AM. Bone remodelling in type I osteoporosis[letter]. J Bone Miner Res 1991 ;6:95-6.

235

on 11 May 2019 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.48.3.229 on 1 March 1995. D

ownloaded from