J Clin Pathol 1981 ;34:1295-1 307

Osteomalacia and chronic renal failureJOHN A KANIS

From the Department of Human Metabolism and Clinical Biochemistry, University of Sheffield MedicalSchool, Beech Hill Road, Sheffield S1O 2RX

Dialysis and transplantation have revolutionised themanagement of chronic renal failure. Many patientswith end-stage chronic renal failure are surviving fora long time, but there are some penalties attached tothis survival, namely the complications associatedwith prolonged renal failure as well as the sideeffects of the treatments (drugs, dialysis and trans-plantation). Major factors which influence themorbidity and mortality of this population includeanaemia, accelerated cardiovascular disease, dis-turbances in gonadal function, hypertension andabnormalities in lipid and skeletal metabolism.

Significant progress has been made in the past fewyears concerning the pathophysiology and treatmentof renal bone disease. This has been due to advancesin our understanding of hormone metabolism,particularly that of vitamin D, and of skeletalphysiology. Tn addition, large populations ofdialysis-or transplant-treated patients have now beenstudied for more than a decade with the result thatthe natural history of this bone disease is moreclearly understood.

Renal bone disease is a constellation of skeletalabnormalities (Table 1), none of which is specific forchronic renal failure. These disorders are frequently,though not invariably, found in combination, andconsideration of one of these in isolation has obviouslimitations. Nevertheless, the emphasis of thisreview is on the aetiology and management ofosteomalacia, but several more thorough andintegrated approaches have appeared recently asreviews or proceedings of symposia.'-5

HISTOLOGICAL DEFINITIONS OF OSTEOMALACIAOsteomalacia can have a variety of meanings. Theclinician may apply the term to the syndromeassociated with vitamin D deficiency, whereas to thebone histologist it implies defective mineralisation ofbone. There is, however, a need for precision, even inhistological definitions since, for example, a histo-logical characteristic of vitaminD deficiency in man isan increase in the amount of unmineralised bonematrix; but increased amounts of osteoid are alsoassociated with many other disorders with normal oraugmented rates of mineralisation. These include

Table 1 Features of renal bone disease and some clinicalmanifestations of disturbed calcium and phosphatemetabolism in chronic renal failure

Feature Clinical consequence

I Hyperparathyroidism and osteitis Skeletal deformity, bonefibrosa pain, pruritis? anaemia,

impotence, neuropathy etc

2 Osteomalacia and decreased Skeletal deformity, boneavailability of vitamin D, pain and tenderness,calcium, and phosphate pathological fracture,

proximal myopathy,haemolytic anaemia

3 Osteoporosis Pathological fracture,skeletal deformity

4 Osteonecrosis Joint pain

5 Osteosclerosis and periosteal None knownnew bone formation

6 Extraskeletal calcification Depends on site-skinulcers, vascular disease,cardiac failure etc

Paget's disease, hyperparathyroidism, fracture repairand hyperthyroidism, indicating that an increase inosteoid volume alone (hyperosteoidosis) is aninadequate criterion for osteomalacia.6 7The amount of osteoid present in bone depends

upon several factors including the rate of appositionof osteoid by osteoblasts, upon the rate of itscalcification, and upon the area of the bone surfaceinvolved. When considering osteomalacia it istherefore important to distinguish hyperosteoidosisdue to an increase in the proportion of bone surfaceson which osteoid is being deposited, as seen inhyperparathyroidism, from that due to a delay orabsence of mineralisation-particularly in chronicrenal failure where osteomalacia and hyperpara-thyroidism often coexist. Much confusion andapparently conflicting reports exist in the publishedreports concerning the pathophysiology of osteo-malacia and its response to various treatments. Anotable factor contributing to this confusion hasbeen the lack of precision in defining osteomalacia,particularly in appreciating the significance of

1295

1296

hyperosteoidosis in hyperparathyroid bone disease(Table 2).

Table 2 Some of the quantitative measurements madeon bone biopsies which have been used (often erroneously)in the assessment of osteomalacia

I Surface osteoid

2 Osteoid volume

3 Osteoid index

4 Maximum number ofosteoid lamellae

5 Calcification front

6 Appositional rate

7 Bone formation rate

proportion of trabecular surfacescovered by osteoid (depends onresolving power of microscope)

proportion of bone matrix (orsectional area) occupied bynon-mineralised bone matrix

mean osteoid seam width: may bederived indirectly from osteoidvolume and surface osteoid

semiquantitative

the proportion of the trabecularsurface (or osteoid surface)undergoing calcification as

judged by stains in vitro-forexample, toluidine blue or in vivo-for example, tetracycline

mean thickness between twotetracycline markers laid downper unit time

derived from 5 and 6

The definition of osteomalacia as impairment ofmineralisation indicates that evidence for osteo-malacia should be obtained by indices of mineralis-ation rather than by the amount of osteoid present.The extent of the bone surface undergoing calcifica-tion can be measured by suitable staining techniquessuch as toluidine blue in vitro or tetracyclinelabelling in vivo. A decrease in the calcificationfront is commonly thought to reflect a decreased rateof mineralisation and hence osteomalacia. There areseveral objections to this view. Thus, the calcificationfront does not measure the rate of mineralisationbut only its extent. Also the calcification frontcorrelates well with the proportion of bone surfacecovered by active-looking osteoblasts in patientswith chronic renal failure, and is therefore a betterindex of the number of functional osteoblasts thanof the adequacy of bone mineralisation.8 Moreover,calcification fronts are commonly expressed as aproportion of the non-mineralised (osteoid) bonesurface, clearly an inadequate baseline in renalosteodystrophy, where surface osteoid may beincreased because of augmented bone turnover dueto hyperparathyroidism. Indeed, reports thatosteomalacia has improved in response to treatmentsuch as vitamin D may be misleading where the"improvement"-that is, reduction in the calcifica-tion front, is due to a decrease in the amount of thebone trabecular surface occupied by osteoid.

Kanis

The double labelling of bone with tetracyclinesprovides a direct method for measuring mineralis-ation in man,7 9 although the technique is timeconsuming to perform. Nevertheless, care must betaken in the interpretation of findings when hyper-parathyroidism is present. Thus, woven osteoid,*which may take up tetracycline more easily thanlamellar bone,10 is laid down in hyperparathyroidismso that normal mineralisation measured in wovenbone may coexist with metabolic abnormalitieswhich might impair mineralisation in more normalbone.An alternative and simpler method of assessing

mineralisation is to measure the maximum numberof osteoid lamellae visible under polarised light."In hyperparathyroidism the "hyperosteoidosis" ismainly due to an increase in osteoid surfaces (andsometimes an increase in lamellar thickness), ratherthan to an increase in the number of osteoid lamellae.6Another simple method is to measure the thicknessof the osteoid seams themselves or to derive thisindirectly from the measurement of osteoid area andthe bone surface covered by osteoid-the osteoidindex (osteoid area divided by surface osteoid x100).6The prevalence of "osteomalacia" according to

some of these histological criteria is shown in Fig. 1.It is important to recognise that all these indices ofosteomalacia have their limitations, particularly inthe presence of hyperparathyroidism. However,without a clear understanding of what is meant byosteomalacia and the problems of its definition,judgements about its pathophysiology and treatmentcan be misleading.

CLINICAL, BIOCHEMICAL AND RADIOGRAPHICFEATURES OF OSTEOMALACIAThe diagnosis of osteomalacia in chronic renalfailure depends mainly on the histological examin-ation of bone, since most patients have no character-istic clinical, biochemical, or radiographic features.The frequency with which osteomalacia causessymptoms is very variable and in the Oxford RenalUnit, symptoms are confined to 10-20% of affectedpatients whereas in the Newcastle Renal Unit, wherethe incidence of osteomalacia is greater12 (see Fig. 2),symptoms appear also to be more frequent.Symptoms of osteomalacia include bone pain andtenderness, and proximal muscle weakness. Pains inthe lower limbs, pelvis and back are particularlycommon and may be worse on exercise. It is ourexperience'3 that bone pain occurs with equalfrequency in osteomalacia and osteitis fibrosa and istherefore not a diagnostic feature. Fractures,*Osteoid is normally laid down in tight lamellar bundles andwoven osteoid requires a degree of structural disorganisation.

Osteomalacia and chronic renal failure

Osteoid area Calcification front

100 /o sectonal area) (1/. trcbecularsurface

0O o°O 0

0t1 0 r 0 a&&

10-

1-1

a.

000

1000

IW

0 0*0

0

0

cPo

0

OPo

0401

Osteoid index( mean osteoid seain

width )

OS

*0 O

00

0

0

00Wo00

80 r

601-a(.2

40~ 40

doCL

0 20e

00

o OMo OF OM* OFo 'N-orOP

Fig. 1 Histological indices of osteomalacia (OM) in 51patients with end-stage chronic renal failure. Patientswere subdivided according to the presence (right handcolumns) or absence (left hand columns) of osteomalaciaas judged by the maximum number of osteoid lamellaevisible on sections under polarised light. Note the lack ofdiscrimination of the calcification front, the osteoidvolume, and the osteoid index in distinguishing patientswith and without osteomalacia. Each of the indicesdiscriminates patients with normal bone histology(N or OP) from those with osteomalacia (OF + OM),but discrimination is lost when patients with osteitisfibrosa (OF) are included.

particularly of the ribs, spine, pelvis and femoralneck, occur with a variable frequency and are morecommon in those dialysis centres with a highprevalence of osteomalacia.

Radiographic features are usually absent and anegative radiographic survey is therefore not helpfulin excluding osteomalacia. Looser's zones (Fig. 3)are characteristic of osteomalacia and are mostfrequently seen in the pelvis, but in dialysis-treatedpatients are a relatively infrequent finding. Manypatients with osteomalacia have radiographicchanges which are associated with hyperparathy-roidism. These changes include subperiosteal erosionsand intracortical porosity. Coarse trabecular mark-ings, periostealnewbone formationand osteosclerosisappear to be more consistently associated withosteomalacia than with osteitis fibrosa.

Children are particularly prone to renal bonedisease and may show radiographic features whichresemble rickets. A rachitic appearance on x-rayexamination does not always mean that osteomalaciais present. Thus, in uraemia there is often no

widening of the metaphyseal zone and the width ofthe growth plate is not as thick as in vitamin D

0

Oxford

0 2 4 6 >6D u ration of hoemodialysis (yecrs)

Fig. 2 The prevalence of osteomalacia in patientsestablished on long-term intermittent haemodialysis, asassessed by repeated bone biopsy. Note the largedifference in prevalence between Newcastle (a highaluminium area) and Oxford (low aluminium in thedialysate fluid) but a similar prevalence of osteomalaciaat the time of starting dialysis. Patients from the OxfordRenal Unit do not invariably develop osteomalacia evenafter many years of haemodialysis.

deficiency (though it may appear so radiographicallybecause of metaphyseal resorption below the growthplate: Fig. 3). These changes are more likely toreflect secondary hyperparathyroidism than vitaminD deficiency.14Few biochemical findings are characteristic of

osteomalacia in chronic renal failure. Plasmaconcentrations of calcium are lower in chronic renalfailure than in health, particularly in children, andmay be lower in patients with osteomalacia thanthose without.'5 These differences are still present inpatients on intermittent haemodialysis though thedifferences are less marked (Fig. 4).A proportion of patients with histologically

proven osteomalacia have normal or high con-centrations of plasma calcium but without evidenceof hyperparathyroidism. In such cases, the boneoften shows the absence of active-looking bone cells,which is a characteristic of aluminium-induced bonedisease (discussed later).Plasma phosphate concentrations tend to increase

as the glomerular filtration rate falls below 30ml/min. The level of plasma phosphate is alsodetermined by the diet, the use of oral phosphate-binding agents, and the dialysis regimen in patientswith end-stage renal failure. Patients with severechronic renal failure also malabsorb phosphate and

1297

C 1

0

0

00

Kaniis

Fig. 3 Radiographic appearance.s of renal bone disease. Left: Characteriitic radiographic features of osteomalacia inia child. Note the Looser's zonte of the tibial shaft, the metaphyseal splaying of the tibia and the increased width of thegrowth plate. Centre: Appearances resembling rickets at the wrist of a patient with hyperparathyroid bone disease.Note metaphyseal bone resorption. Right: Markedfeatures of hyperparathyroid bone disease of the wrist.

this may be the explanation for the osteomalacia andmarked hypophosphataemia seen in some patientsnot taking phosphate-binding agents. Plasmaconcentrations of phosphate tend to be lower inpatients with osteomalacia but osteomalacia cannotbe diagnosed from this alone, since the presence ofnormal or high plasma concentrations do not excludeosteomalacia (Fig. 4).Plasma alkaline phosphatase and plasma hydroxy-

proline are commonly raised in chronic renalfailure. An increase in alkaline phosphatase activitymay not always be due to an increase in bone-derived phosphatase but may be due to increasedactivities ofthe gut and liver isoenzymes. In populationstudies, plasma activities of alkaline phosphatasecorrelate well with histological indices of bone cellactivity-for example, osteoblast counts. Hence,alkaline phosphatase activity is usually increased inpatients with osteomalacia combined with secondaryhyperparathyroidism, but is normal or low in manypatients with osteomalacia alone (Fig. 4).

Since osteomalacia can rarely be diagnosed with-out histological assessment of bone, studies of itspathogenesis and the assessment of treatmentregimens must include the study of biopsy material,and require a critical awareness of the meaning ofthe histological variables measured.

PATHOGENESIS OF OSTEOMALACIA IN

CHRONIC RENAL FAILUREMany factors are thought to be important in thepathophysiology of osteomalacia (Table 3). Theevidence incriminating these factors is frequentlycircumstantial but their consideration is valuable indevising strategies for treatment.

Metabolism of vitamini DDeficiency of vitamin D in man retards skeletalgrowth and results in defective mineralisation ofmatrix produced both by chondrocytes and byosteoblasts. Reversal of these abnormalities byvitamin D provides convincing evidence for theimportance of vitamin D in skeletal homeostasis.However, it is not yet clear whether formation andmineralisation of bone and cartilage are directactions of vitamin D metabolites or whether theseprocesses are secondary to the actions of vitamin Dat non-skeletal sites of calcium and phosphatetransport such as the gut and kidney. In addition,vitamin D metabolites may regulate the secretion ofother hormones such as parathyroid hormone andcalcitonin, which themselves modify calcium andphosphate concentrations and bone metabolism.

In the past decade the major thrust in vitamin Dresearch has been directed to studies of its metabolism

1298

3 0 -

z 5 -

2-0-

3-0 -

Plasma 2 0-phosphate(mml/11)

1 0 -

700 -

400 -

PLasma20-

alkaline 200

phosphatase(lIU/) 100-

50 -

(n=32) (n=27) (n=28) (n=30)0

0

0ccoo

o _

o

0

0

080 9

_ __ _ _ _ _ _ _

-8 c~%'X

-aTo-_, q-

F f t ~~~~~t9

0

0

0

_0_0

0

0

Normal OM OM OF OF

Fig. 4 Biochemical findings in plasma (predialysisvalues) in patients on dialysis treatment. Patients havebeen grouped according to the type ofbone disease.OM = osteomalacia, OF = osteitis fibrosa. The brokenlines denote the limits of the normal range.

to active products.'6-'8 The kidney is involvedcritically in this process since it is the major site ofproduction of 1,25-dihydroxy vitaminD3 (1 ,25(OH)2-D3) and 24,25(OH)2-D3.

It has been recognised for a long time that largedoses of vitamin D may increase the intestinalabsorption of calcium and heal osteomalacia andosteitis fibrosa in patients with renal failure.19 The

1299

Table 3 Some factors thought to be important in thepathogenesis of osteomalacia in chronic renalfailure

Disturbed vitamin D metabolismInadequacy of intake of vitamin D or exposure to UV lightDefective production of 1,25(OH),-D,Excessive urinary losses of 25-OHD (nephrotic syndrome)? Defective production of 24,25(OH),-D,Accelerated metabolism of 25-OHD (anticonvulsants etc)

Toxic effects on boneAluminium intoxication-dialysate and ? aluminium-containing

phosphate-binding agents? FluorideAnticonvulsants (target organ resistance to vitamin D

metabolites)

Availability of calcium and phosphateLow plasma phosphate (malabsorption, hyperparathyroidism,

steroids, phosphate-binding agents, lengthy dialysis schedules)Very low calcium diets

UncertainTotal parathyroidectomyUraemic inhibitors (? pyrophosphate)? Acidosis

bone disease in renal failure is "vitamin D-resistant"in the sense that the doses of vitamin D required toproduce a biological response are greater than thoserequired to satisfy physiological needs in normalindividuals. There is now considerable evidence thatthis resistance in renal failure is due to a defect in themetabolism ofvitamin D.17The first step in the metabolism of vitamin D3 is its

conversion to 25-hydroxy vitamin D3 (25-OHD3)which occurs mainly in the liver (Fig. 5). Is renal bonedisease due to defective production, acceleratedmetabolism, or impaired action of this metabolite?Certainly plasma concentrations of25-OHD (D2 andD3) may be low in patients with the nephroticsyndrome and be associated with bone disease.20Apart from this exception, there is little evidence thatlow concentrations of 25-OHD are due to defects invitamin D metabolism, despite suggestions to thecontrary. Certain drugs, such as anticonvulsants andbarbiturates, induce hepatic microsomal enzymes,and might, therefore, increase the metabolism of25-OHD to inert products, but there is little directevidence for this. A more important effect ofanticonvulsants may be to block the action ofvitamin D metabolites on gut and bone.2' Lowconcentrations, when present, may therefore be dueeither to inadequate diet or reduced exposure tosunlight and might be expected to contribute toosteomalacia particularly when the degree of renalfailure is modest. The usual experience is that, whenanticonvulsants are avoided and patients allowed toeat normal diets (for example on dialysis treatment),plasma 25-OHD concentrations are normal.5

It is commonly thought that most of the actions ofvitamin D3 are mediated by metabolism of 25-OHD3

Osteomalacia and chronic renal failure

Plasmacalcium(mmolI1)

Kanis

D

I/f--O OH

Liver ?

CH2 CH2

HO3 25-OH -D3

dney

OHOH

OCH2

25,26 -(OH)2 D3

Kidney

H

OH

CH2

HO ^'(24R)-24, 25-(OH)2 D3

Fig. 5 Some of the major steps in the metabolism of vitamin D.

to 1,25(OH)2-D3. Since the kidney is probably thesole site of synthesis of 1,25(OH)2-D3 (apart from theplacenta in pregnancy), the development of osteo-malacia and its resistance to vitamin D and to25-OHD may result from impaired production of1,25(OH)2-D3 due to loss of renal tissue and perhapsalso to the inhibitory effects of hyperphosphataemiaon the renal 1-alpha hydroxylase.'6-18The evidence for a causal relation between vitamin

D resistance, defective 1-alpha hydroxylation, andosteomalacia, is based on several observations.Firstly, plasma concentrations of 1,25(OH)2-D3 andits rate of formation decrease when the glomerularfiltration rate is less than 40 ml/min, and 1,25(OH)2-D3 is usually undetectable in end-stage chronicrenal failure.'7 Secondly, x-ray and histologicalappearances of renal bone disease have featureswhich resemble nutritional vitamin D deficiency.Thirdly, administration of 1,25(OH)2-D3 or itssynthetic analogue 1-alpha hydroxy vitamin D3,reverses many of the biochemical and radiographicfeatures of osteomalacia. Moreover, the doses of1 ,25(OH)2-D3 required to maintain remission(0 25-0 5 ,ug daily) are close to its estimated daily

endogenous production rate in health, suggestingthat target organs are sensitive to physiologicalamounts of 1 ,25(OH)2-D3in contrast to 25-OHD3.1322The view that lack of 1,25(OH)2-D3 is the major

cause of osteomalacia in renal failure may be anoversimplification. Thus not all patients with severechronic renal failure have osteomalacia, and itsincidence does not invariably increase with time ondialysis despite defects in the metabolism of vitaminD.23 The incidence of osteomalacia varies widelybetween dialysis units (Fig. 2) suggesting the im-portance of dialysis-related factors in its aetiology.Moreover, the prevalence of osteomalacia is notincreased markedly in anephric patients, and in somesuch patients the rates of mineralisstion and boneformation are normal. Treatment with 1-alphahydroxylated metabolites may improve radiographicand biochemical indices of osteomalacia, but this isnot invariable, particularly when histological indicesof response are used.'3 22 23 Radiographic criteria ofhealing may be inadequate, particularly in childrenwhere the so-called rachitic appearances of theepiphyses may be due to hyperparathyroidism ratherthan to osteomalacia. These data suggest that factors

1300

1.

Osteomalacia and chronic renal failure

other than defective production of 1,25(OH)2-D3must contribute significantly to the osteomalacia ofchronic renal failure.The production of 24,25(OH)2-D3 may also be

deficient in renal failure since the kidney has anenzyme system capable of converting 25-OHD3 tothis metabolite. In chronic renal failure, plasmaconcentrations of 24,25(OH)2-D3 are, in somepatients, lower than would be predicted from thecirculating concentration of 25-OHD3.24 25 Pre-liminary observations suggest that patients withrenal failure and osteomalacia may have the lowerconcentrations of 24,25(OH)2-D3.25 The administra-tion of 24,25(OH)2-D3 alone in doses sufficient torestore plasma concentrations to normal hasmetabolic effects but does not heal osteomalacia(unpublished observations). There is increasingevidence that both 1,25(OH)2-D3 and 24,25(OH)2-D3are required for the actions of vitamin D on bone tobe complete.26 27 However, the relation betweendecreased synthesis of 24,25(OH)2-D3 and osteo-malacia in chronic renal failure will be clarified onlywhen we know whether or not the kidney is the solesite of production of this metabolite (a controversialissue) since nephrectomydoes not appear to aggravaterenal bone disease.23

Calcium and phosphate metabolismIt is possible that the skeletal effects of vitamin D arenot dependent upon direct actions of the metaboliteson bone itself but are due to consequent changes inthe plasma concentration of calcium, phosphate andparathyroid hormone. One of the reasons why thisquestion remains open is the difficulty in studyingmineralisation in vitro, while the effects of vitamin Dmetabolites in vivo are complex.

In many disorders, including renal failure, thelevel of plasma phosphate appears to be an importantdeterminant of osteomalacia. The concentration ofplasma phosphate in dialysis-treated patients hasbeen shown to correlate inversely with the degree ofosteomalacia, such that those patients with normalamounts of osteoid have the higher plasma phosphateconcentrations (Fig. 6).23 It may be relevant thatphosphate concentrations in such patients areconsiderably higher than the upper limit of normal inhealth, and hyperphosphataemia may thereforeprotect the patient from osteomalacia, despitedefective vitamin D metabolism. If such a protectivemechanism exists, hyperphosphataemia may maskthe true importance of disturbed vitamin D metab-olism in renal bone disease.The case for the phosphate effect on mineralisation

is greater than for calcium. Calcium deficiency insome experimental models leads to osteoporosisrather than to osteomalacia. However, it has been

201

0-aEE

a 1-5E

0. .Ia. 10.

1 3 5 70or moreMaximum number of lamel

omeFig. 6 Relation between plasma phosphate(Pi; mean of values taken immediately beforedialysis treatment ± SEM) and the number ofosteoid lamellae seen in histological sections.Osteomalacia (five or more lamellae) wasuncommon in patients with high concentrationsofplasma phosphate. Evidence of osteomalaciawas commonly noted in patients whose plasmaphosphate concentration lay below the upperlimit of the normal range (indicated by brokenlines).

shown recently that the delayed rate of osteoidmaturation in D-deficient rats is correctable whenhypocalcaemia is reversed by dietary calciumsupplements.28 There is controversy as to whethercalcium deficiency alone in the presence of adequatevitamin D nutrition can cause osteomalacia in man.The most convincing demonstration of calciumdeficiency rickets is in Bantus who have normalvitamin D status.29 In chronic renal failure it has beenreported that severe calcium deficiency may renderthe patient unresponsive to 1,25(OH)2-D3, whereashealing of osteomalacia occurs when the diet isadequately supplemented with calcium.30

Trace elements and water contaminantsA number of trace elements accumulate in chronicrenal failure due either to impaired excretion or toabsorption from the dialysate. These include arsenic,strontium, molybdenum, magnesium, manganese,copper, aluminium and fluoride, but, with thepossible exception of fluoride and aluminium, theirrole in the evolution of dialysis bone disease isunclear. Patients with osteomalacia do not responduniformly to treatment with 1,25(OH)2-D3 or itssynthetic analogue, 1-alpha-OHD313 22 despite

1301(n=2) (ii=21)(n=22)(n=lg)(n=19)tn.-7)(n=12)

T

1302

adequate control of plasma phosphate concentrationsand the administration of calcium supplements.Indeed the failure to respond to 1,25(OH)2-D3 maybe one method of separating patients in whomosteomalacia is due to other causes. It is also theimpression of many groups that osteomalacia morecommonly fails to respond to vitamin D treatment inpatients on intermittent haemodialysis that inpatients managed conservatively, and this has ledto renewed interest in the possibility that disturbancesinduced by haemodialysis itself may give rise toosteomalacia.The possible role of fluoride has been the most

extensively studied. Fluoride accumulates in chronicrenal failure during haemodialysis and is depositedin bone. It is also known that high doses in maninduce the formation of excessive amounts ofosteoid. Several groups have attempted to find acorrelation between fluoride and bone disease where-as others have looked at the effects of reducing thefluoride content of the dialysate. However, there isno consensus view to be obtained from the manystudies performed3' and it is now recognised that thebone disease attributed to fluoride may have beendue to other factors not recognised at the time.

There is an increasing body of evidence thataluminium retention may be an important factor inthe pathogenesis of osteomalacia in dialysis-treated patients. A form of osteomalacia associatedwith a high incidence of bone pain and fractures iscommon in certain geographical locations with ahigh aluminium content in the water (Fig. 2), and itsincidence appears to decrease with deionisation of thewater. There is also a good correlation (in the UK)between osteomalacia associated with fractures andthe aluminium content of the water used fordialysis.3233Flendriget al.34 showed that the incidenceof fractures (and of dialysis dementia) in their unitwas associated with a source of aluminium in thedialysate lines whereas other patients using the sametap water did not develop this complication. Thissuggests that either aluminium or some factorassociated with aluminium is responsible for thehigh prevalence of osteomalacia seen in some centres.The question arises whether or not phosphate-

binding agents containing aluminium salts may giverise to aluminium toxicity, since there is goodevidence that oral aluminium is absorbed.35 One ofthe difficulties of investigating this problem is thatskeletal retention of aluminium may remain formany years after the source has been withdrawn andaluminium toxicity may take many years to develop.Thus, it is difficult to assess the relative importanceof oral and water-borne aluminium. The evidence todate suggests that water-borne, rather than oralaluminium, is most commonly the major source of

Kanis

skeletal aluminium.

AcidosisThe role of acidosis in contributing to renal bonedisease has been advocated for many years. It hasbeen suggested that bone acts as a buffer by releasingalkaline bone salts and this is supported by thefinding of a decrease in the bicarbonate content ofbone from uraemic patients.36 The acute administra-tion of acid loads to normal man results in a neg-ative calcium balance37 and it has been noted that therate of mineralisation increases acutely whenalkalis are administered to acidotic patients withosteomalacia.38 The long-term effects of metabolicacidosis on bone are, however, less clear and it hasnot been regarded as a major factor, since thecorrection of acidosis appears to influence renal bonedisease in only a minority of patients.39 Chronicacidosis in the absence of uraemia or hypophos-phataemia-for example, chronic respiratory disease,diabetes mellitus, Gaucher's disease, is not character-ised by osteomalacia.

ParathyroidectomyTotal parathyroidectomy may be associated with thedevelopment of osteomalacia when osteitis fibrosahas healed.' 10 It is interesting that, apart frompatients with aluminium toxicity, a high proportionof patients with recognised osteomalacia and renaldisease who fail to respond to treatment withvitamin D metabolites have had previous parathy-roidectomies.'3 22 In some instances this lack ofresponse may be related to persistent hypophos-phataemia, but this is not invariably the case. Therelative ease with which woven osteoid, present inhyperparathyroidism, calcifies in the absence ofvitamin D (compared with lamellar osteoid) mayexplain the appearance of osteomalacia only afterparathyroidectomy when lamellar bone is laid down.A corollary is that osteitis fibrosa may protectagainst osteomalacia.

OtherfactorsA number of years ago Yendt et al40 noted thaturaemic plasma contained a factor which inhibitedthe calcification of rat cartilage. The nature of thisuraemic factor has not been elucidated. One candi-date might be pyrophosphate which is an inhibitor ofcalcium phosphate nucleation in vitro. Increasedconcentrations are found in patients with hypo-phosphatasia and in patients with chronic renalfailure, and it is conceivable that these might causefailure of mineralisation in both conditions. Inpopulation studies there does not appear to be aclear relation between histological indices ofosteomalacia and plasma pyrophosphate4' but this

Osteomalacia and chronic renal failure

does not necessarily exclude a regulatory role forpyrophosphate or other inhibitors of skeletalmetabolism.

TREATMENT OF OSTEOMALACIA IN CHRONICRENAL FAILUREThe strategy of treatment should be based not onlyon the presence of osteomalacia or associatedsymptoms but also on a careful assessment of theother skeletal abnormalities present, the otherconsequences of disturbed mineral metabolism(Table 1), and the mechanisms responsible for thedisorder. The proposed management of the chronicrenal disease itself should also be considered since,for example, therapeutic approaches may depend onthe probability of subsequent transplantation.There are a number of preventative measures

which should be considered in all patients withadvanced renal impairment. It is probable that thesevere restriction of dietary protein, as sometimespractised, is a greater factor in inducing morbiditythan it is in achieving beneficial effects. Though lowprotein diets restrict the amount of phosphate, thisis better achieved with the use of phosphate-bindingagents. Protein-deficient diets also tend to restrict theintake of vitamin C and pyridoxine which both act asessential cofactors in the formation and maturationof collagen. The relation between deficiency of thesefactors and osteomalacia is unclear and they may bemore important in the pathogenesis of osteoporosis.Both vitamin C and vitamin B6 should be given asdietary supplements.

Treatment ofdisturbedphosphate metabolismDespite compensatory decreases in tubular re-absorption of phosphate, plasma phosphate con-centrations are nearly always markedly increased inpatients with severe renal impairment (Fig. 4). Thevalue of lowering plasma phosphate concentrationsin preventing hyperparathyroid bone disease in manis uncertain but the control of plasma phosphate isimportant in the management and prevention ofextraskeletal calcification,42 and is usually achievedby the use of phosphate-binding agents, such asaluminium hydroxide. Calcium carbonate also bindsphosphate in the gut, and has potential advantages inthat it corrects acidosis, increases the dietary calciumload, and enables the ingestion of aluminium to beavoided; but in practice the amounts of calciumcarbonate required are large.

In order to avoid extraskeletal calcification,predialysis concentrations of plasma phosphateshould be less than 2-4 mmol/1.42 Factors whichinfluence the dose required of the phosphate-binding agent include the dietary intake of phosphate,concurrent treatment with vitamin D and its ana-

logues or metabolites, and the haemodialysistreatment schedule prescribed. Profound hypo-phosphataemia should also be avoided since it isassociated with osteomalacia.43 If the relationbetween predialysis concentrations of phosphate andhistological indices of osteomalacia is causal (Fig. 6),it is important to notethat theconcentrationsofplasmaphosphate below which osteomalacia may beinducedare considerably higher than those associatedwith impaired mineralisation in patients with normalrenal function. Thus the plasma phosphate con-centrations which best balance the risks of metastaticcalcification and osteomalacia probably lie between1-4 and 2-4 mmol/l in dialysis-treated patients.Phosphate therapy has been advocated as a

method of treating osteomalacia in chronic renalfailure. This is not without risk and should only beattempted, and then cautiously, in patients withtubular disorders. The infusion of phosphate duringthe dialysis treatment does not appear to improveosteomalacia and indeed may aggravate osteitisfibrosa.44

Treatment ofdisturbed cakcium metabolismUnlike the net intestinal absorption of phosphatewhich is largely dependent on the dietary load, thenet absorption ofcalcium is more critically dependenton the presence of the vitamin D metabolites,particularly 1,25(OH)2-D3. Nevertheless, net in-testinal transport of calcium can be augmented bylarge amounts of calcium carbonate (5-20 g daily)and this may improve osteomalacia.45 It is oftenmore practicable to give vitamin D or one of itsmetabolites, but net intestinal absorption of calciumcannot be greatly augmented if the diet is severelydeficient in calcium. Moreover, marked calciumdeficiency appears to impair the response to1,25(OH)2-D3.30 It is important, therefore, to ensurea normal dietary intake of calcium with the use ofcalcium supplements if necessary.The dialysis membrane provides a site for the loss

of calcium or its incorporation into the body, butthere is no evidence that the dialysate calciumconcentration influences the natural history ofosteomalacia.

Use ofvitamin D and related compoundsA variety of vitamin D compounds is available foruse in chronic renal failure. These include vitaminD2, D3, 25-OHD3, dihydrotachysterol (DHT),1,25(OH)2-D3 and 1-alpha-OHD3. A great deal ofclinical interest has focused on 1,25(OH)2-D3 andits synthetic analogue 1-alpha-OHD3 since theybypass the metabolic block caused by uraemia, butDHT is also biologically active without the necessityfor l-alpha-hydroxylation by the kidney. DHT and

1303

Kan is

1-alpha-OHD3 undergo hepatic hydroxylation andthe 25-OHDHT or 1,25(OH)2-D3 so formed are themajor circulating forms of these agents. It has beensuggested that anticonvulsants interfere with thehepatic production of these metabolites and thesedrugs should therefore be avoided for this reason.The clinical evidence for this is not clear though itshould be remembered that anticonvulsants probablyalso interfere with the target organ actions of all thevitamin D compounds.

All the vitamin D-like compounds available areeffective in relieving symptoms of bone pain andmuscle weakness, in increasing plasma calciumconcentrations, and in the majority of patients theysuppress raised plasma activities of alkaline phos-phatase and correct radiographic abnormalities(Figs. 7 and 8).13 19 22 Skeletal deformity in the youngcan probably be prevented and growth partlyrestored.46 47

The histological response to treatment is oftendisappointing, particularly in patients maintained onintermittent haemodialysis, where factors other thandisturbed vitamin D metabolism presumably play adominant role in the pathophysiology. Osteomalacia

1 oc- HCC(g/day) l

29 _

Plasma Ca2- 5

2-1

Plasma alkalinephosphatase(IU / l)

Plasma iPTH(,Ug / )

appears to respond more readily when associatedwith osteitis fibrosa. Once again this may be relatedto the different pathogenic mechanisms.Although most patients with end-stage chronic

renal failure have histological evidence of bonedisease, they are often symptomless. In patients ondialysis treatment, considerable differences exist inthe incidence and natural history of osteomalaciabetween renal units. Whether to treat asymptomaticpatients with vitamin D metabolites will thereforedepend on several factors. Our own policy at presentis not to treat those patients with abnormal bonehistology unless they have symptoms, markedhypocalcaemia or radiographic or biochemicalevidence of bone disease. Several trials are currentlyunderway to evaluate the efficacy of I-alpha-OHD3or 1,25(OH)2-D3 in preventing the development ofovert bone disease and this view may thereforerequire modification.Doses of the various agents required to maintain

the plasma calcium concentration within the normalrange and to reverse bone disease are indicated inTable 4. In general, the maximal dose which avoidshypercalcaemia decreases with time (Fig. 7). The

MT Renal bone disease

+ OF and OM+Normalx-ray4 Normal bone biopsy

200

100 , (o, {~~~~~~~~~~~~~~~~~~~~~~)iol

6-0r

3-0

6 12Duration of treatment ( months)

Fig. 7 Long-term treatment of osteomalacia (OM) with I-alpha-hydroxy-vitamin D3(1-alpha-HCC) in a dialysis-treated patient. Healing of osteomalacia occurred within 15months. Episodes ofhypercalcaemia occurred suddenly and the dose of I-alpha-HCCtolerated decreased progressively once plasma alkaline phosphatase had fallen to normalactivities. Remission from bone disease was maintained using a dose of J-alpha-HCC ofI ,ug thrice weekly. OF = osteitis fibrosa, i PTH = immunoreactive parathormone.

0 18 24s a a I --i

1304

Osteomalacia and chronic renal failure

:'3.1.78 o i

Fig. 8 Radiographic responses to treatment of osteomalacia with large doses of vitamin D.

Table 4 Usual dose requirements of vitamin D3 (or vitamin D2), dihydrotachysterol (DHT), I a-OHD3, and 1,25(OH)2-D3 in dietary deficiency rickets and in vitamin D resistance due to chronic renal failure. Note that, though largeramounts ofDHT than vitamin D are required to treat simple rickets, for the treatment of renal bone disease onlyslightly higher doses of DHT (or Ia-OHD3 or 1,25(OH)2-D3 up to x 4) are required in contrast to the much largerdoses of vitamin D3 (up to x 400) that are required

D, DHT la-OHD3 1,25(OH),-D,

Approximate daily dose required to treat or prevent rickets (,ug) 2 5-25 up to 200 up to 1 0 up to 0 5Potency relative to vitamin D, 100 10 250 500Approximate daily dose required to treat renal bone disease (,ug) 750-10 000 200-1000 0 5-2-0 0 25-2-0Potency relative to vitamin D, 100 1000 500 000 500 000

Note I jig Ds or 1,25(OH)2-D3 is equivalent to approximately 40 IU.

greatest risks of hypercalcaemia occur at the start oftreatment, particularly in patients who respondpoorly to treatment (aplastic osteomalacia), and laterwhen biochemical responses are nearing completion.Plasma calcium concentrations should be monitoredfrequently during these risk periods. It is also impor-tant to note that these agents increase the absorptionof phosphate and the requirements for phosphate-binding agents may be increased.

Despite advocates to the contrary, there is no

clinical evidence that 1,25(OH)2-D3, I-alpha-OHD3or DHT have any particular therapeutic actionsnot also possessed by other agents such as 25-OHD3or vitamin D.1348 The advantages of the 1-alpha-hydroxylated metabolites of vitamin D lie in the ease

with which doses are titrated according to require-ments and the speed with which toxic effects are

reversed on withdrawal of treatment.49Treatment with vitamin D or its metabolites is not

without risk. Prolonged raised plasma calcium andphosphate concentrations give rise to extraskeletalcalcification. Patients with pre-existing hyper-calcaemia and osteomalacia should be treated

cautiously, if at all, since such patients often respondpoorly to treatment. Prolonged hypercalcaemia mayalso impair renal function, sometimes irreversibly,and there has been recent interest in the suggestionthat vitamin D compounds may themselves benephrotoxic.50 It is difficult, however, to be sure howmuch any deterioration of renal function reflects thenatural history of the disorder or the effects of thehypercalcaemia or hyperphosphataemia,51 and othershave not noted such adverse effects when plasmacalcium and phosphate are well controlled. Theseconsiderations nevertheless re-emphasise the needfor close control of plasma calcium and phosphate,and the serial measurement of plasma creatinine inpatients not yet established on dialysis treatment.

Aluminium toxicityThis severely disabling condition does not respond totreatment with vitamin D or its metabolites. It mayrespond slowly to transplantation or adequateremoval of aluminium from the dialysis fluid. Thisslow response probably reflects the prolongedskeletal retention of aluminium and the difficulty

1305

Kanis

with which it is removed by haemodialysis, andindicates the importance of prophylaxis. Thus, wherealuminium concentrations in the dialysate exceed 20Mtg/l, the water should be treated by reverse osmosisor possibly by deionisation.33

ParathyroidectomySurgical removal of parathyroid glands is the mosteffective and rapid method of treating hyper-parathyroid bone disease. The place for parathy-roidectomy is beyond the scope of this review, but theoccurrence (or unmasking) of osteomalacia followingtotal parathyroidectomy is one of the reasons whypartial parathyroidectomy may be preferred.

Renal transplantationTheoretically renal transplantation would be thetreatment of choice for patients with osteomalacia.This rapidly restores the capacity to form 1,25(OH)2-D3 and of course reverses uraemia. Osteomalacia isoften slow to reverse, particularly when associatedwith aluminium retention. Bone disease, includingosteomalacia, may arise de novo in the transplantedpopulation. There is a high incidence of hypophos-phataemia in the transplant population52 and thismay be partly related to phosphate depletion by theuse of antacids, to the persistence of hyperpara-thyroidism which is slow to regress after transplan-tation, and to the effects of corticosteroids in de-creasing renal tubular reabsorption of phosphate.

I am grateful to the National Kidney ResearchFund, the Wellcome Trust, the Medical ResearchCouncil and the Special Trustees of the SheffieldArea Health Authority for their generous support ofthis work.

References

lDavid DS, ed. Calcium metabolismi in renal failure andnephrolithiasis. New York: John Wiley & Sons, 1977.

2 Avioli LV. Renal osteodystrophy in metabolic bone disease.Vol 2. Avioli LV, Krane SM, eds. New York: AcademicPress, 1978.

3 Peacock M. The clinical uses of 1 tY-hydroxyvitamin D3.Clin Fndocrinol (Oxf) 1977 ;7, suppi.

4 Coburn JW, Massry SG, eds. Uses and actions of 1,25-dihydroxyvitamin D3 in uremia. Contrib Nephrol 1980;18.

Kanis JA. Features and pathophysiology of renal bonedisease. In: Davies AM, ed. Haemodiali'sis review.Tunbridge Wells: Pitman, 1978:182-204.

6 Meunier P, Edouard C, Richard D, Laurent, J. Histo-morphometry of osteoid tissue: the hyperosteoidoses.In: Meunier PJ, ed. Bone histomorphometryl. Toulouse:Societe de la Nouvelle Imprimerie Fournie, 1976 :249-62.

Frost HM. The bone dynamics in osteoporosis andosteomalacia. Springfield: Charles C Thomas, 1966.

Woods CG. Why count osteoblasts? In: Meunier PJ ed.,HistomorphometrY. Toulouse: Societe de la Nouvelle

Imprimerie Fournie, 1976.9 Frost HM. Relation between bone tissue and cell popula-

tion dynamics, histology and tetracycline labelling. ClinOrthop 1966;49:65-75.

10 Ritz E, Krempien V, Mehls 0. Skeletal abnormalities inchronic renal insufficiency before and during haemo-dialysis. Kidney Int 1973;4:116-27.

'1 Woods CG, Morgan DB, Patterson CR, Grossman HH.Measurement of osteoid in bone biopsy. J PatholBacteriol 1968 ;95 :441-7.

12 Ellis HA, Pierides AM, Feest TG, Ward MK, Kerr DNS.Histopathology of renal osteodystrophy with particularreference to the effects of 1 o-hydroxyvitamin D:i inpatients treated by long-term haemodialysis. ClinEndocrinol (Oxf ) 1977 ;7 :31-8.

13 Kanis JA, Cundy T, Earnshaw M, et al. Treatment ofrenal bone disease with la-hydroxylated derivatives ofvitamin D3. Clinical, biochemical, radiographic anidhistological responses. Q J Med 1979;48:289-322.

14 Krempein B, Mehls 0, Ritz E. Orthological studies on thepathogenesis of epiphyseal slipping in uraemic children.Virchows Arch 1974;362:129-34.

15 Stanbury SW, Lumb GA, Mawer EG. Osteodystrophydeveloping spontaneously in the course of chronic renalfailure. Arch Intern Med 1969;124:274-81.

16 De Luca HF. Recent advances in our understanding ofthe vitamin D endocrine system. J Lab Clin Med 1976;87 :7-26.

17 Haussler MR, McCain TA. Basic and clinical conceptsrelated to vitamin D metabolism and action. N Engl JMed 1977 ;297:974-83, 1041-50.

18 Lawson DEM, ed. Vitami1in D. London: Acadeniic Press.1978.

19 Stanbury SW, Lumb GA. Metabolic studies in renalosteodystrophy. Medicine (Baltiniore) 1962 ;41 :1-31.

20 Malluche HH, Goldstein DA, Massry SG. Osteomalaciaand hyperparathyroid bone disease in patients withnephrotic syndrome. J Clin Invest 1979;63:494-500.

21 Jenkins MV, Harris M, Wills MR. The effect of phenytoinion parathyroid extract and 25-hydroxycholecalciferolinduced bone resorption: Adenosine 3'-5'. cyclicmonophosphate production. Calcified Tissue Research1974;16:163-7.

22 Coburn JW, Brickman AS, Sherrard DJ, et al. Clinicalefficacy of 1,25-dihydroxy-vitamin D: in renal osteo-

dystrophy In: Norman AW, Schaefer K, Coburn JW,et al, eds. Vitamin D: biochemical, chemical and clinicalaspects related to calcium metabolisnm. Berlin: WalterDeGruyter, 1977:657.

23 Kanis JA, Adams ND, Earnshaw M, et al. Vitamin Dosteomalacia on chronic renal failure. In: Norman AW.Schaefer K. Coburn JW, et ao, eds. Vitamin D: bio-chemical, chemical and clinical aspects related to calciummetabolism. Berlin: Walter DeGruyter, 1977:671.

21 Weisman Y, Eisenberg Z, Leib L, et al. Serum concentra-tions of 24,25-dihydroxy-vitamin D in different degreesof chronic renal failure. Br MedJ 1980;281:712-3.

25 Kanis JA, Russell RGG, Taylor CM, Cundy T, AndradeA, Heynen G. The relationships between disturbedmetabolism of vitamin D and bone disease in chronicrenal failure. Proc Eur Dial Transplant As.soc 1979;16:630-6.

26 Malluche HH, Henry H, Meyer-Sabellek W, et al. Effectsof interactions with 24R, 25-OHD3 and 1,25(OH)2-D:on bone. Am JPhysiol 1980;283:494-8.

27 Bordier P, Rasmussen H, Marie P, et al. Vitamin Dmetabolites and bone mineralisation in man. J ClinEndocrinol Metab 1978 ;46:284-94.

2 8 Howard GA, Baylink DJ. Matrix formation and osteoid

1306

Osteomalacia and chronic renal failurematuration in vitamin D deficient rats made normo-calcaemic by dietary means. Min Elect Metab 1980;3:44-50.

29 Pettifor JM, Ross FP, Moodley G, De Luca HF, TraversR, Glosseux FH. Calcium deficiency rickets associatedwith elevated 1,25-dihydroxyvitamin D concentrationsin a rural black population. In: Norman AW, SchaeferK, Herrath DV, et al, eds. Vitamin D: Basic research andits clinical application. Berlin: Walter DeGruyter, 1979:1125-7.

30 Cundy T, Kanis JA, Heynen G, et al. Lack of direct effectof 1,25-dihydroxy vitamin D3 on mineralisation of boneand secretion of parathyroid hormone. Clin Sci 1980;59:15P.

31 Rao STK, Freidman EA. Fluoride and bone disease inuraemia. Kidney Int 1975 ;7:125-9.

32 Platts MM, Goode GC, Hislop JS. Composition of thedomestic water supply and the incidence of fractures andencephalopathy in patients on home dialysis. Br Med J1977 ;ii :657-60.

33 Parkinson IS, Ward MK, Feest TG, Fawcett RPW, KerrDNS. Fracturing dialysis osteodystrophy and dialysisencephalopathy. An epidemiological survey. Lancet1979;i:406-9.

34 Flendrig JA, Kruis H, Das H. Aluminium and dialysisdementia. Lancet 1976;i:1235.

35 Clarkson EM, Luck VA, Hynson WV, et al. The effect ofaluminium hydroxide on calcium, phosphorus andaluminium balances, the serum parathyroid hormoneconcentration and the aluminium content of bone inpatients with chronic renal failure. Clin Sci 1972;53:519-31.

36 Pellegrino ED, Bilz RM. The composition of bone inuraemia. Observations on the reservoir functions of boneand demonstration of a labile fraction on carbonate.Medicine (Baltimore) 1965 ;44:397-418.

3 Leman J Jr, Litzow JR, Lennon EJ. The effect of acidlevels in normal man: turther evidence for the partici-pation of bone mineral in the defence against chronicmetabolic acidosis. J Clin Invest 1966:45:1608-14.

38 Cochran M, Hopkinson R. Effect of correction of metabolicacidosis on bone mineralisation rates in patients withrenal osteomalacia. Nephron 1975;15:98-1 10.

39 Fletcher RF, Jones JH, Morgan DB. Bone disease andchronic renal failure. Q J Med 1963;32:321-39.

40 Yendt ER, Connell TB, Howard JE. In vitro calcificationof rachitic rat cartilage in normal and pathologicalhuman sera with some observations on the pathogenesis

1307of renal rickets. Bulletin of the Johns Hopkins Hospital1955;96:1-19.

4] Bishop MC. Bone disease in the Oxford Renal Unit.University of London: MD Thesis, 1976.

42 Parafitt AM. Soft-tissue calcification in uraemia. ArchIntern Med 1969 ;124:544-56.

43 Ahmed KY, Varghese Z, Wills MR, et al. Persistenthypophosphataemia and osteomalacia in dialysispatients not on oral phosphate-binders: response to DHTtherapy. Lancet 1976 ;ii :439.

44 Feest TG, Ward MK, Ellis HA, Aljama P, Kerr DNS.Osteomalacic dialysis osteodystrophy: A trial ofphosphate enriched dialysis fluid. Br MedJ 1978 :i : 18-20.

41 Meyrier A, Marsac J, Richet G. The influence of a highcalcium carbonate intake on bone disease in patientsundergoing haemodialysis. Kidney Int 1973;4:146-53.

46 Kanis JA, Henderson RJ, Hevnen G, et al. Renal osteo-dystrophy in non-dialysed adolescents: long-term treat-ment with l-alpha-hydroxycholecalciferol. Arch DisChild 1977;52:473-81.

4 Chesney RW, Moorthy AV, Eisman JA, et al. Increasedgrowth after long-term oral 1a, 25-vitamin D in child-hood renal osteodystrophy. NFnglJ Med 1978;298:238.

48 Prior JC, Cameron EC, Ballon HS, Lirenman LS, MoriartyMV, Price JDE. Experience with 1,25-dihydroxychole-calciferol therapy in undergoing haemodialysis patientswith progressive vitamin D2-treated osteodystrophy.AmJMed 1979 ;67:583-9.

49 Kanis JA, Russell RGG. Rate of reversal of hypercalcaemiaand hypercalciuria induced by vitamin D and its1-alpha-hydroxylated derivatives. Br MedJ 1977;i:78-81.

10 Christiansen C, Rodbro P, Christensen MS, Hartnack B,Transbol I. Deterioration of renal function duringtreatment of chronic renal failure with 1,25-dihydroxy-cholecalciferol. Lancet 1978 ;ii :700-3.

51 Naik RB, Cundy T, Robinson BHB, Russell RGG, KanisJAK. Effects of vitamin D metabolism and analogues onrenal function. Nephron 1981:28:17-25.

52 Moorhead JF, Wills MR, Ahmed KY, Baillod RA,Varghese Z, Tatler GLV. Hypophosphataemic osteo-malacia after cadaveric renal transplantation. Lancet1974;i :694-7.

Requests for reprints to: Dr JA Kanis, Department ofHuman Metabolism and Clinical Biochemistry, Universityof Sheffield Medical School, Beech Hill Road, SheffieldS1O 2RX, England.