Annals of the Rheumatic Diseases, 1980, 39, 173-179

Renal handling of uric acid in normal and goutysubjects: evidence for a 4-component systemDENNIS J. LEVINSON AND LEIF B. SORENSEN

From the Department of Medicine (Division of Rheumatology), Michael Reese Hospital and Medical Center,and Department of Medicine, University of Chicago, Chicago, Illinois, USA

SUMMARY Bidirectional renal urate transport was studied in both control and gouty subjects.99-3 % of filtered urate undergoes reabsorption as assessed by pyrazinamide suppression of uratesecretion. The maximum uricosuric response to benzbromarone, equated with the minimumsecretory rate, amounted to 50% of the filtered load in normal persons and was lower in goutynormoproducers. Since benzbromarone selectively inhibits reabsorption of secreted urate, thedifference between secreted and excreted uric acid becomes a valid measure of urate reabsorptiondistal to the secretory site and amounts to 80% ofthe secretedload inboth groups. These dataconformto a 4-component model of renal urate handling in man.

In normal man about two-thirds of uric acid isexcreted in the urine. Until quite recently the '3-component mechanism' proposed by Gutman andYu (1961) has been widely accepted as a model forcharacterising the renal handling of urate. Accord-ing to that formulation plasma urate is essentiallycompletely filterable, at the glomerular membrane;virtually all the filtered urate is reabsorbed in theproximal tubule, and uric acid appearing in thefinal urine occurs almost entirely by a tubularsecretory process.

Recently evidence has accumulated suggestingthat the renal handing of uric acid is more complexthan had been previously proposed. It has beenpostulated that a second reabsorptive site existsdistal to the secretory site. Evidence for this comesin part from clinical studies of patients with Wilson'sdisease or Hodgkin's disease, who have increasedrenal clearance of uric acid. When these patientswere given pyrazinamide (PZA), a drug known tosuppress tubular secretion of uric acid, the urinebecame almost free of uric acid, indicating that theirinappropriate renal clearance of urate did not resultfrom a defect in proximal tubular reabsorption offiltered urate (Bennett et al., 1972; Wilson and Gold-stein, 1973). Furthermore, in pharmacologicalstudies it has been observed that the response touricosuric drugs is greatly diminished by pretreat-ment with PZA (Steele and Boner, 1973; Diamond

Accepted for publication 30 April 1979Correspondence to Professor L. B. Sorensen, Departmentof Medicine, Pritzker School of Medicine, 950 East 59thStreet, Chicago, Illinois 60637, USA.

and Paolina, 1973). These data could be interpretedto mean that uricosuria arising from either diseasestates or pharmacological manipulation is due tostimulation of tubular urate secretion or, alter-natively and more likely, results from diminishedreabsorption of secreted urate.

In this report we present data which conformto a 4-component system for renal handling ofurate, namely: (1) plasma urate is completelyfiltered at the glomerulus; (2) filtered urate issubsequently reabsorbed, presumably in the proxi-mal tubule; (3) tubular secretion of urate occursfurther distally in quantities approximately 50%of the filtered load; and (4) reabsorption of aboutfour-fifths of secreted urate occurs at a postsec-retory site. In addition, by using pharmacologicalmeasures we are able to assess the site of inappro-priate renal handling of urate in patients withprimary gout who have abnormally low urinaryuric acid secretion for their plasma urate values.

Materials and methods

Binding of urate to plasma proteins was studiedat different temperatures by equilibrium dialysisusing dual-compartment Plexiglass cells with acapacity of 1 ml (Chemical Rubber Company,Cleveland, Ohio). Fresh heparinised plasma wasseparated from dialysate made up of Krebs-Ringerphosphate solution, pH 7 4, by a regenerated cel-lulose membrane (average pour diameter: 4 8).3 ,ug of 14C-uric acid (specific activity 0 15 mc/mmol)was added to either sample or diffusate side. In

173

on June 2, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.39.2.173 on 1 A

pril 1980. Dow

nloaded from

http://ard.bmj.com/

174 Levinson, Sorensen

some binding studies cold uric acid was added tothe dialysate with the final uric acid concentrationvarying from 2 4 mg/dl (0.14 mmol/l) to 24 mg/dl(1 42 mmol/l). In studies using the highest con-centration of uric acid Krebs-Ringer phosphatesolution was replaced by 0-2 M Tris buffer, pH7 4, in which urate remained soluble. After in-cubation for 12 hours aliquots were withdrawnfrom the sample and diffusate sides for assay of 14Cand enzymatic determination of urate.The reabsorption of filtered urate was determined

in 10 healthy subjects who had been given PZAto inhibit tubular secretion of uric acid. Urinaryuric acid was determined in the 12 hour periodbetween 3 and 15 hours after oral intake of 4 g ofPZA. Reabsorbed uric acid was calculated as thedifference between filtered and excreted uric acid.Tubular secretory rate of urate was assessed in

8 healthy volunteers and in 12 patients with primarygout, all of whom had normal glomerular filtrationrates. In the gouty group 3 patients were overpro-ducers of uric acid, while 9 patients had normalproduction of urate but low urinary uric acidexcretion relative to their plasma urate levels.Adequate diuresis was accomplished with tap watergiven at '-hour intervals. Urine specimens wereobtained by voiding at hourly intervals.

After baseline values for plasma and urinary uricacid had been obtained 80 mg of micronised benz-bromarone, a potent uricosuric drug, was given ina single oral dose (Sorensen and Levinson, 1976).Uric acid excretion was determined in each hourlyfraction and expressed as Lg/minute (,mol/min)divided by the creatinine clearance to correct fordifferences in nephron mass. Maximum uricosuriawas usually obtained during the fourth hour afterbenzbromarone administration. Urate excretionduring maximum uricosuria related to the meanplasma urate concentration for that collectionperiod becomes a measure of minimum secretionat a particular plasma urate concentration. Uricacid determinations were performed by the enzy-matic spectrophotometric method of Praetorius(1949), while creatinine in serum and urine wasassayed on an SMA660 (Terrytown, NY) (Chasson,et al., 1961). Medications were discontinued 14 daysprior to all studies, which were conducted in themorning following an overnight fast. Informedconsent was obtained from both normal volunteersand the gouty patients.

Results

FILTERED URATE LOADPlasma from 15 normal subjects and 20 patientswith primary gout was evaluated for possible

binding of urate to plasma proteins. The distri-bution of 14C between plasma and dialysate isshown in Fig. 1. At 40C about 15 % of urate is boundto proteins, but as the temperature increases there isa marked reduction in binding of urate, such that at370C the mean concentration of 14C in 22 plasmasamples was 98 2% of the concentration in dialy-sate. Plasma from normal and gouty subjects hadsimilar binding characteristics. A plot of the ratiosof urate concentrations in plasma versus dialysateclosely resembled the data for 14C. All of the uratepresent in the 2 compartments of the cell at theonset of the dialysis study could be accounted forat the end of the incubation. Varying the concen-tration of uric acid in the dialysate from 2-4 mg/dl(0 14 mmol/l) to 24 mg/dl (1 42 mmol/l) did notalter the ratios of 14C or urate between plasma anddialysate. The use of serum instead of plasma gaveidentical results.The results of in-vitro protein binding studies

at 370C suggest that in-vivo binding of urate islikely to be nonexistent or of very low order and ofno physiological consequence. To the extent thatdata obtained with semipermeable membranes invitro can be extrapolated to the in-vivo situationin man, binding of urate to plasma proteins doesnot appear to be a factor affecting filtration of urateby the glomerulus. The filtered urate load cantherefore be assessed from the plasma urate con-centration and a measure of glomerular filtrationrate such as the creatinine clearance. The meanfiltered load for the 10 healthy subjects listed inTable 1 was 6*2 mg/minute (36*88 Vmol/min)in the control period (not shown).

1 20 r

E o0 ?la .2

u u?

- I1

108

104

100

0-96

S

4 20 37Temperature (*C)

Fig. 1 Effect of temperature on urate binding toplasma proteins. The concentration of uric acid in thedialysate rangedfrom 0 14 mmolll to 1 48 mmol/l.Open circles, normal men; closed circles, gouty subjects;X indicates absence of binding of 14C-labelled allantoin,the oxidation product of uric acid. No correctionwas made for Gibbs-Donnan effects.

XX I46

.

1 16-

1-12-

on June 2, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.39.2.173 on 1 A

pril 1980. Dow

nloaded from

http://ard.bmj.com/

Renal handling of uric acid in normal and gouty subjects: Evidence for a 4-component system 175

Table 1 Tubular reabsorption of uric acid in 10 normal subjects following oral intake of4 g pyrazinamideSex Plasma urate at 9 h Ccr (mIl/min) Uric acidfiltered Uric acid excreted Reabsorption of

after PZA (mg/dl) (JLg/min) (pmol/min) (jg/min) (ymol/min) filtered uric acid(mmol/l) (per cent)

M 6.97 (0-41) 121 8434 (50-16) 42 (0-25) 99.5M 6-50 (0-38) 132 8580 (51-03) 76 (0.45) 99.1M 6.28 (0-37) 116 7285 (43-33) 26 (0-15) 99-6M 6.50 (0-38) 114 7410 (44-08) 48 (0-29) 99-4M 5.48 (0-32) 80 4384 (26-08) 38 (0-23) 99-1M 6.49 (0-38) 160 10 384 (61-77) 120 (0.71) 98-8M 5-32 (0-31) 117 6224 (37-02) 67 (0-40) 98-9M 6-50 (0-38) 129 8385 (49-88) 47 (0-28) 99.4M 7-66 (0-45) 90 6894 (41-01) 25 (0-15) 99.6F 3-90 (0-23) 100 3900 (23-20) 9 (0-05) 99-8Mean 6-16 (0-36) 116 7188 (42-76) 50 (0.30) 99.3

TUBULAR REABSORPTION OFFILTERED URATEThe pronounced inhibition of tubular secretioncaused by PZA is evident in Fig. 2. A single dose of4 g is followed within a few hours by a strikingreduction in excretion of uric acid, which lasts foralmost 24 hours. Concomitantly plasma urateconcentration rises. The difference between thequantities of filtered and excreted uric acid duringmaximum PZA suppression is a measure of theproportion of filtered urate that has been reab-sorbed. Results of PZA studies in 10 healthy sub-jects are presented in Table 1. The mean excretionof uric acid in the 12-hour period from 3 to 15hours after intake of 4 g of PZA was 50 Vg/minute(0 30 ,umol/min), representing less than 1 % of thefiltered load. Assuming complete suppression oftubular secretion of urate, a mean of 99 - 3% (range98 8-99 8%) of the filtered load was reabsorbed.The tubular reabsorption may be even greater thandetermined from these studies, since tubularsecretion of urate may not be completely inhibitedby PZA.

Pyrazinamide. 4g orally

3

-

.D

-4 0 4 8 12 16 20 24 28 32 36Hours

Fig. 2 Effect of a single oral dose ofpyrazinamideon renal urate excretion in a normal subject. SIconversion: plasma urate mmolll = mg/dl x 0-0595.

The capacity for reabsorption of filtered urate isevident from a study in which the filtered load wasincreased 4-fold by the administration of 16 g ofRNA for several days (not shown). During thecontrol period the filtered load of urate was 20 mg/minute (118 96 ,mol/min). The fractional excretionof urate after PZA intake decreased to 0 * 5 %; thus,a minimum of 99 5% of filtered urate had beenreabsorbed.The marked suppression of urinary uric acid after

PZA administration implies that excreted uric acidis derived almost entirely, if not exclusively, fromtubular secretion in normal man.

TUBULAR SECRETION OF URIC ACIDReference has been made to a number of clinicaland pharmacological observations that tubularsecretion of urate greatly exceeds the quantity ofuric acid that appears in the final urine, and further-more that the difference between the two representssecreted urate that has been reabsorbed at a post-secretory site (Bennett et al., 1972; Wilson andGoldstein, 1973). To clarify this relationship wefirst examined the effects of benzbromarone on uricacid excretion in individuals who had been pre-treated with PZA to suppress tubular secretion ofuric acid.A representative study is shown in Fig. 3A. PZA

causes a predictable fall in urinary urate excretionfrom a baseline value of 445 ,ug/minute (2 * 65 ,umole/min) to a nadir of 31 ,g/minute (0-18 Vmol/min),equivalent to 0 5% of the filtered load. The typicaluricosuric response which normally reaches amaximum 3 to 4 hours after oral intake of a singledose of 80 mg of benzbromarone (Fig. 3B) iscompletely abolished by pretreatment with PZA.At a time when the maximum uricosuric effectnormally ensues, urinary uric acid was only 10%of the control excretion and plasma urate had in-creased paradoxically by 0 -65 mg/dl (0 -04 mmol/l),3-4 hours after ingestion of benzbromarone. Theresults of 4 such studies in normal subjects are

on June 2, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.39.2.173 on 1 A

pril 1980. Dow

nloaded from

http://ard.bmj.com/

176 Levinson, Sorensen

Pyrazinamide

I Benzbromaronej 80mg

Benzbromarone80mg

A

B

-.

-4 0 4 8 12 16 20 24 28 32 36Hours

Fig. 3 A. Inhibition of uricosuric response tobenzbromarone by pretreatment with pyrazinamide.B. A representative study of the uricosuric responseto benzbromarone. SI conversion: plasma uratemmol/l = mg/dl x 00595

Table 2 Eradication of uricosuric response tobenzbromarone by pretreatment with pyrazinamide*Time Purate UVurate(min) (mg/dl) (mmol/l) (pg/min) (pmol/min)

0-60 5*36+±079 (0-32+0.05) 514±50 (3 06+0-30)Pyrazinamide 4 g given orally

240-300 5*67±0*79 (0*33+0*05) 4410 (0*26±0*06)Benzbromarone 80 mg given orally

480-540 6-25±0-84 (0-37±0-05) 78±24 (0-46+0-14)1560 7-91±0-52 (0-47+0-03)

*Purate and Uvurate are expressed as the mean±SD in 4 normalsubjects.

summarised in Table 2. A uricosuric response tobenzbromarone is lacking in all cases. A slightincrease in urinary uric acid is noticeable afterbenzbromarone administration. Presumably thedrug inhibits reabsorption of the minute quantitiesof urate that arrives at a distal reabsorptive site,either because trivial amounts of urate escapeproximal reabsorption, or, more likely, because PZAfails to completely inhibit tubular secretion ofurate.

These results suggest that uricosuric drugs act byinhibiting reabsorption of secreted urate at a post-secretory site. This means in turn that tubularsecretion of urate can be determined by blockingdistal reabsorption completely. To the extent thatbenzbromarone selectively inhibits postsecretoryreabsorption of urate it is possible to assess tubularsecretion of urate by measuring maximum uric acidexcretion following administration of a suitabledose. In other words, the maximum uricosuricresponse can be equated to the minimum secretoryrate.The results of studies on tubular secretion in 8

healthy volunteers and 3 gouty subjects with over-production of uric acid are shown in Table 3 andFig. 4. Baseline urate excretion values during thecontrol period and maximum uricosuric responsesare plotted against plasma urate values over a rangeof 2 05 mg/dl (0-12 mmol/l) to 10 72 mg/dl (0 63mmol/l). In this group of subjects a linear relation-ship is seen between tubular urate secretion andplasma urate concentrations (r=0 * 956; P

Renal handling of uric acid in normal and gouty subjects: Evidence for a 4-component system 177

40

35-

m 30

c 25E, 20

15-

D 10-

D 5

00

Pl

Discussion

2 3 4 5 6lasma urate (mg /dl )

Fig. 4 Regression lines and 95 ,describing urate excretion as a fburate in 8 normal subjects and 3,Open squares depict control studjP

178 Levinson, Sorensen

uricosuria, presumably because of inhibition of bothtubular secretion, as well as reabsorption of urate.The finding that the fractional urate excretionamounts to only 0 7% after PZA administrationindicates that urate undergoes virtually completereabsorption at a site proximal to the tubular siteof urate secretion. The capacity for reabsorptionof filtered urate is extraordinary, as evidenced instudies in which the filtered urate load was increasedby administration of yeast RNA for several days.

Rieselbach (1976) reported studies in 6 normalsubjects whose mean filtered urate had been raisedto 17*4 mg/minute (103 *50 ,tmol/min). Theirfractional urate excretion remained as low as 0 - 51 %,indicating a mean reabsorption of 17 3 mg/min(102.90 l.mol/min). In a single study we found thatthe fractional urate excretion was 0 * 5 % after filteredurate had been increased to 20 mg/minute (118 96,umol/min).The pyrazinamide suppression test was originally

advocated as a test for estimating urate secretorytransport in man (Steele and Rieselbach, 1967). Thevalidity of the test for this purpose was based onthe premise that secretion of tubular urate occurs insegments of the nephron distal to all reabsorptivesites. With the finding that substantial quantitiesof secreted uric acid undergoes reabsorption furtherdistally in the nephron, it is clear that the testgrossly underestimates the magnitude of tubularsecretion. However, it remains a valid test forquantifying reabsorption of filtered urate.From the data obtained in studies with PZA it

may be concluded that uric acid is derived almostentirely from tubular secretion in normal man.Several pieces of evidence make it clear that onlya portion of secreted uric acid is excreted in theurine. The finding that the uricosuric response tobenzbromarone is completely nullified in subjectspretreated with PZA can best be interpreted tomean that the uricosuric action of this compoundis related to a direct inhibitory effect upon post-secretory reabsorption. Similarly, the uricosuricresponse to agents such as probenecid or intra-venous chlorothiazide has been found to be greatlydiminished in subjects pretreated with PZA (Steeland Boner, 1973; Diamond and Paolino, 1973).To the extent that it is possibletoinhibitpostsecretoryreabsorption of urate, it is possible to study the secre-tory transport mechanism more directly. Benzbroma-rone is particularly suitable for this purpose, since it isnot excreted by a renal tubular organic acid transportmechanism and therefore does not interfere withtubular secretion of uric acid (Podevin et al., 1967).The paradoxical effect of urate retention seen whenconventional uricosuric drugs are given in lowdosage is not observed in the case of benzbromarone.

The maximum urate excretion after benzbromaroneadministration becomes a measure of the tubularsecretory rate for a given plasma urate level.Increased availability of urate to the secretorysite results in a steep increase in the secretoryrate. Obviously, the values observed in the presentstudies represent minimum secretory rates, sincepostsecretory reabsorption may have been onlypartially inhibited by benzbromarone in the dosagegiven.

Reabsorption of uric acid at the postsecretorysite increases progressively with increasing avail-ability of urate. Thus within the plasma urate rangestudied the magnitude of postsecretory reabsorptionappears to be set at a constant percentage of thesecreted load.At least three-fourths of all patients with primary

gout have normal urate production and requirean elevated plasma urate concentration to obtaina normal level of uric acid excretion (Seegmilleret al., 1961). The technique of measuring the urico-suric response to benzbromarone has allowedidentification of a defect in the renal handling ofurate in these patients which can be localised to thesecretory site. Although gouty normoproducers canbe characterised as having a subnormal response tobenzbromarone, this drug, as is the case with otheruricosuric agents, decrease postsecretory fractionalreabsorption of uric acid, resulting in a loweringof plasma urate levels.

Recent reports of defects in urate reabsorptionin man support the proposed scheme for renalhandling of uric acid. One defect appears to involvediminished reabsorption of filtered urate in thaturate excretion was not significantly reduced afterPZA treatment (Greene et al., 1972). Another defectexists in patients with urate clearances higher thanthe glomerular filtration rate, which would appearto involve markedly impaired reabsorption of bothfiltered and secreted urate transport throughout thenephron (Simkin et al., 1973). We have reported athird defect confined solely to the postsecretoryreabsorptive site (Sorensen and Levinson, 1980).The exact mechanism and sites of tubular urate

transport remain to be clarified. Recent micro-puncture studies in the rat indicate substantial uratereabsorption in the loop of Henle (Greger et al.,1974). Variation in urine flow exerts a moderateeffect on urate excretion in man, suggesting areabsorptive site also in the distal tubule or collectingducts (Diamond et al., 1972).

This work was supported in part by a grant 683-F9022 fromthe Illinois Chapter of the Arthritis Foundation.

The authors are grateful to Ms Orneary Rogers for pre-paring the manuscript.

on June 2, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.39.2.173 on 1 A

pril 1980. Dow

nloaded from

http://ard.bmj.com/

Renal handling of uric acid in normal and gouty subjects: Evidence for a 4-component system 179

References

Bennett, J. S., Bond, J., Singer, I., and Gottlieb, A. J. (1972).Hypouricemia in Hodgkin's disease. Annals of InternalMedicine, 76, 751-756.

Chasson, A. L., Grady, H. T., and Stanley, M. A. (1961).Determination of creatinine by means of automatedchemical analysis. American Journal of Clinical Pathology,35, 83-88.

Diamond, H. S., and Paolina, J. S. (1973). Evidence for apostsecretory reabsorptive site for uric acid in man.Journal of Clinical Investigation, 52, 1491-1499.

Diamond, H. S., Lazarus, R., Kaplan, D., and Halberstam,D. (1972). Effect of urine flow rate on uric acid excretionin man. Arthritis and Rheumatism, 15, 338-346.

Fanelli, G. M., Jr., and Weiner, I. M. (1973). Pyrazinoateexcretion in the chimpanzee: relation to urate dispositionand the actions of uricosuric drugs. Journal of ClinicalInvestigation, 52, 1946-1957.

Greene, M. L., Marcus, R., Aurbach, G. D., Kazam, E. S.,and Seegmiller, J. E. (1972). Hypouricemia due to isolatedrenal tubular defect: Dalmatian dog mutation in man.American Journal of Medicine, 53, 361-367.

Greger, R., Lang, F., and Deetjan, P. (1974). Urate handlingby the rat kidney: IV. Reabsorption in the loops ofHenle. European Journal of Physiology, 352, 115-120.

Gutman, A. B., and Yu, T. F. (1961). A three componentsystem for regulation of renal excretion of uric acid in man.Transactions of the Association of American Physicians,74, 353-365.

Kaufman, J. M., Greene, M. L., and Seegmiller, J. E. (1968).Urine uric acid to creatinine ratio: A screening test forinherited disorders of purine metabolism. Journal ofPediatrics, 73, 583-592.

Klinenberg, J. R., and Kippen, I. (1970). The binding ofurate to plasma proteins determined by means of equilib-rium dialysis. Journal ofLaboratory and Clinical Medicine,75, 503-510.

Palmer, D., Levinson, D. J., and Sorensen, L. B. (1974).Urate-losing nephropathy as an isolated defect. Journalof Rheumatology, 1, (suppl. 1) 29A.

Podevin, R., Paillard, F., and Amiel, C. (1967). Action de labenzbromarone sur 1'excr6tion r6nale de l'acide urique.Revue Fran!Vaise d'Etudes Cliniques et Biologiques, 12,361-367.

Praetorius, E. (1949). An enzymatic method for the deter-mination of uric acid by ultraviolet spectrophotometry.Scandinavian Journal of Clinical and Laboratory Investi-gation, 1, 222-230.

Rieselbach, R. E. (1976). Renal handling of uric acid. PurineMetabolism in Man. Advances in Experimental Medicineand Biology, Vol. 76B, pp. 1-22. Edited by M. M. Muller,E. Kaiser, and J. E. Seegmiller. Plenum Press: New York.

Seegmiller, J. E., Grayzel, A. I., Laster, L., and Liddle, L.(1961). Uric acid production in gout. Journal of ClinicalInvestigation, 40, 1304-1314.

Simkin, P. A., Skeith, M. D., and Healey, L. A. (1973).Suppression of uric acid secretion in a patient with renalhypouricemia. Israel Journal of Medical Sciences, 9,11 13A.

Sorensen, L. B., and Levinson, D. J. (1976). Clinical evalua-tion of benzbromarone: A new uricosuric drug. Arthritisand Rheumatism, 19, 183-190.

Sorensen, L. B., and Levinson. D. J. (1980). Isolated defectin postsecretory reabsorption of uric acid. Annals of theRheumatic Diseases, 39, 180-183.

Steele, T. H., and Boner, G. (1973). Origins of the urico-suric response. Journal of Clinical Investigation, 52, 1368-1375.

Steele, T. H., and Rieselbach, R. E. (1967). The renalmechanism for urate homeostasis in normal man. AmericanJournal of Medicine, 43, 868-875.

Wilson, D. M., and Goldstein, N. P. (1973). Renal urateexcretion in patients with Wilson's disease. Kidney Inter-national, 4, 331-336.

on June 2, 2021 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.39.2.173 on 1 A

pril 1980. Dow

nloaded from

http://ard.bmj.com/