Annals of the Rheumatic Diseases 1994; 53: 743-748

Urine neopterin: a new parameter for serialmonitoring of disease activity in patients withsystemic lupus erythematosus

Kean L Lim, Ken Muir, Richard J Powell

AbstractObjective-To investigate the role of serialmeasurement of urine neopterin concen-tration in monitoring the progression ofsystemic lupus erythematosus (SLE)disease activity scored using the BritishIsles Lupus Assessment Group (BILAG)index.Methods-We foliowed prospectively 68unselected SLE patients for a total of 464patient months during which 233 separateassessments were carried out. At eachassessment, urine neopterin, determinedby high performance liquid chromato-graphy, together with erythrocyte sedi-mentation rate (ESR) and plasma C3, C4,and C3d were measured and the SLEdisease activity scored by a singleobserver. Serial data sets were analysedusing time series modelling techniques.Results-Single time point analysisshowed a significant increase in urineneopterin concentrations in 14 patientswho suffered flares of their disease duringthe study period (p = 002). Thirty patientswith active disease went into diseaseremission with significant decreases intheir urine neopterin values (p = 0.02). Inthe time series analysis, a statisticallysignificant association was found betweenserial concentrations of urine neopterinand BILAG score (r = 0-6, p < 0.05); noother study parameter (ESR and serumC3, C4, and C3d) mirrored SLE diseaseactivity as effectively.Conclusions-This study provides initialevidence that changes in urine concen-trations of neopterin are significantlycorrelated with fluctuations in diseaseactivity over time, scored using theBILAG index, amongst individual patientswith SLE. Consequently, serial urineneopterin measurements appear to beclinically useful for monitoring diseaseactivity and may contribute substantiallyto therapeutic decision making in thesepatients.

(Ann Rheum Dis 1994; 53: 743-748)

A challenge in the clinical management ofpatients with systemic lupus erythematosus(SLE) is to identify laboratory parameters thatwill distinguish those with active and thosewith inactive disease, and parameters that canbe used to monitor serially disease activity and

disease progression. Erythrocyte sedimen-tation rate (ESR), serum or plasmacomplement concentrations, and antibodies todouble stranded DNA (dsDNA) are currentlyused in clinical practice as markers of diseaseactivity. However, some patients can havepersistently abnormal results in these tests andyet show few clinical symptoms or functionaldeterioration of a major organ, whilst otherpatients are severely ill with only minimalchanges in these parameters. Moreover, someparameters are only useful in certain patternsof organ involvement in SLE; for example,changes in anti-dsDNA and complementconcentrations are reported predominantly toaccompany flares of lupus nephritis. l 2Urine neopterin, an indirect marker of the

upregulation of the cellular immune system,3has previously been demonstrated to be ofvalue as a parameter of disease activity in anunselected group of SLE patients.4 In thatstudy patients with SLE, irrespective of diseaseactivity, exhibited significantly increasedconcentrations of urine neopterin comparedwith healthy controls. Using logistic regressionanalysis, it was shown that urine neopterin wassuperior to serum soluble interleukin-2 (sIL-2)receptors and anti-dsDNA, ESR, and plasmacomplement C3, C4 and C3 degradationproducts (C3d) as a predictor of diseaseactivity in patients with SLE. Urine neopterinappeared to be a more reliable index of globalclinical activity in SLE patients, irrespective oforgan involvement or treatment with corti-costeroid or cytotoxic agents, or both; however,the temporal relationship between urineneopterin and SLE disease activity was notaddressed. This study investigated the role ofserial urine neopterin in relationship to SLEdisease activity scored using the British IslesLupus Assessment Group (BILAG) index.5

Patients and methodsThe study population consisted of 68unselected patients regularly attending theconnective tissue disease clinic who fulfilledfour or more of the 1982 revised AmericanRheumatism Association criteria for theclassification of SLE.6 The mean (SD) age andthe mean (SD) disease duration of the patientcohort at time of recruitment were 43 (12-8)years and 11 5 (9 5) years, respectively, with amale to female ratio of 3:65. These patientswere studied prospectively for a total of 464patient months (median 6, range 0-18), duringwhich 233 separate assessments (median 3,

Department ofImmunology,University Hospital,Queens MedicalCentre,Nottingham,United KingdomK LLimR J PowellDepartment ofPublicHealth Medicine andEpidemiologyK MuirCorrespondence to:Dr K L Iim,Centre for RheumaticDiseases,Royal Infirmary,Glasgow G3 1 2ER.

Accepted for publication27 July 1994

743

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from

Lim, Muir, Powell

range 1-10) were carried out. None of thepatients was considered to have a concomitantviral or bacterial infection or condition otherthan SLE at time of assessment. Subsequentfollow up supported this clinical opinion.Informed consent was obtained from eachpatient, and the protocol was approved by theNottingham University Hospital EthicsCommittee.SLE disease activity was scored by a single

observer (KLL) using the computerisedBILAG index. A score of A or B in any of theeight organ systems assessed was considered torepresent active disease. A flare in diseaseactivity was defined as a change in BILAGindex grade from inactive to active; diseaseremission was defined as a change from activeto inactive on two successive assessments. Anumerical score for each assessment (BILAGscore) was also calculated from the individualscores for each organ system: A scored 9, B 3,C 1, and D and E each scored 0.

LABORATORY MEASUREMENTS

In all patients, venous blood was obtained attime of recruitment into this study formeasurement of haemoglobin concentration,white cell count with differential count, plateletcount, ESR, serum concentrations of urea andelectrolytes, liver function tests, and concen-

trations of antibodies to nuclear antigens,antibodies to extractable nuclear antigens,anti-dsDNA, and plasma C3, C4 and C3d. Anearly morning urine sample was obtained at thesame time for measurement of neopterin. ESR,plasma C3, C4 and C3d, and urine neopterinwere available also for all subsequentassessments.

Urine neopterin was measured, withoutprior knowledge of the BILAG status of thepatients, using reverse phase high performanceliquid chromatography (HPLC) as describedpreviously4 7 and expressed in pLmolmolcreatinine. In brief, urine samples, protectedfrom light and previously stored frozen, were

centrifuged to remove debris, diluted 1 in 10with water containing dimethylpterine as an

internal standard and injected directly onto a

Techsphere 5 ODS column (HPLC Tech-nology Ltd). A binary gradient elution was

used with an initial mobile phase of 2%methanol in 15 mmol/l phosphate buffer, pH6-4, increasing to 25% methanol after 12minutes, and neopterin was detected by itsnatural fluorescence (Xex 353 nm; Xem 438 nm).Urine creatinine was determined separatelyusing a kinetic alkaline picrate Gaffe) method;other test parameters of interest were

quantified by standard methods: ESR bySeditainer ESR System, plasma C3 and C4 bynephelometry, and C3d by double deckerimmunodiffusion.

STATISTICAL ANALYSIS

Spearman's rank correlation coefficients (r,)were computed for various study parametersand the BILAG score. Differences betweenparameters within a group were evaluated with

the Wilcoxon Matched Pairs Signed Rank Sumtest. In order to avoid statistical bias, forpatients who had more than one assessmentduring the study period we included only datafrom the initial assessment in the correlationanalysis.Urine neopterin measurements were made

at varying intervals and over a period ofmonths. In order to study the temporalrelationship between urine neopterin concen-tration and SLE disease activity, the data setwas therefore modelled as a time series.8Initially, an analysis of covariance was carriedout to assess the relationship between BILAGscore and patient factors. In this model, timeand patient were included as factors, urineneopterin as a covariate and BILAG score asthe dependent variable. The initial assessmentfor each patient was numerically coded 0, withsubsequent consecutive assessments coded 1,2, 3, etc, and only patients with more than fiveassessments each were included into themodel. Where time was found not to have asignificant effect, the association between themean BILAG score and the mean urineneopterin value for each patient was thenexamined by regression analysis. Similaranalyses were carried out for ESR and plasmaC3, C4 and C3d. For these analyses, theSPSSx' software package was used;9 p < 0 05was considered significant.

ResultsOf the 68 patients, 35 had more than twoassessments (nine had three, seven had four,five had five, four had six, seven had seven, onehad eight and two had 10 assessments) duringthe study period. According to the BILAGindex, 116 of these assessments were definedas active and 117 were inactive. Twenty fivepatients had active disease involving more thanone organ system and 24 had active diseaseinvolving one organ system only. Active diseaseinvolved the general system in 10 patients, themucocutaneous system in 24, the centralnervous system in nine, the musculoskeletalsystem in 18 patients, the cardiorespiratorysystem in six, vasculitis in 12 patients, the renalsystem in three, and the haematological systemin 17 patients, either in isolation or as part ofmultiple organ system disease.At the time of initial evaluation, urine

neopterin concentration was found to correlatesignificantly with the BILAG score (rs = 0-32,p = 0 008). Significant correlations were alsofound for C3 (rs = -0 38, p = 0 00 1), and C4(rs=-0 45, p=00001), but not for ESR(rs = 0 09) or C3d (r, = 0 08).

URINE NEOPTERIN AND DISEASE FLARE

Fourteen patients had flares of their diseaseduring the study period. Figure 1 shows plotsof their respective urine neopterin values.Median urine neopterin value was found to beincreased during the flare (342 pLmol/molcreatinine) compared with the inactive diseasephase (243 pLmol/mol creatinine) (differencesignificant, p = 0 02). Second successive

744

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from

Urine neopterin for serial monitoring of disease activity in SLE

750-

.)

c

._

4-

E0)

E

c._

a)0.

0

C

650-

750

._i-

EE

Ca

,.-

a)o

a)

0._

0

a)

CL

650

550

450

350 -

250 -

150 -

Dv2nd

Inactive Active assessment

Disease activity

Figure 1 Comparison of direction ofchange in urineneopterin concentrations between successive assessments withdisease activity in 14 patients who sufferedflares of theirdisease. Second successive assessments were carried out, onaverage, 1 9 months (range 1-4 months) after the first. m]and m2 = Median neopterin values for inactive and activedisease phases, respectively. BD = Patient referred to in text.

assessments were available for eight of the 14patients; four were found to have persistentlyactive disease with greater urine concentrationsof neopterin compared with correspondingvalues during the inactive disease phase.Despite the disease remaining active, urineneopterin values in three of these four patientswere lower in the second successive assessmentcompared with corresponding values duringtheir disease flares, and these values corres-ponded with lower BIIAG scores, suggestingclinical improvement. One patient (BD in fig1) had persistent tiredness and lethargy withconsistently increased urine neopterin concen-trations in addition to increased C3d and ESRvalues, but the BILAG grade became activeonly when she developed acute pleurisy and amalar rash, which responded to subsequenttreatment. Urine neopterin values for twopatients showed a downward trend despite aflare in their disease that subsequently settled.

In contrast, no significant differences in ESRand plasma C3, C4 and C3d were found inthese patients during their disease flares.

URINE NEOPTERIN AND DISEASE REMISSIONThirty patients with active disease went intoremission during the study period (fig 2).Second successive assessments were availablefor 19 patients, 16 of whom remained indisease remission. Urine neopterin concen-trations were significantly lower (p=0 02)when the patients had inactive as opposed toactive disease with median values of 249 and320 p,mol/mol creatinine, respectively. Urineneopterin values for the 16 patients who had

5-2nd

Active Inactive assessment

Disease activityFigure 2 Comparison of direction ofchange in urineneopterin concentrations between successive assessments withdisease activity in 30 patients with active disease who wentinto disease remission. Second successive assessments werecarried out, on average, 2-9 months (range 1-6 months)after the first. mi, m2, and m3 = Median neopterin valuesfor active, inactive, and second successive inactive diseasephases, respectively. NS = not significant.a

a second inactive assessment were alsosignificantly lower than the correspondingvalues during the active disease phase(p <0.01). In contrast, no significant differ-ences were found between urine neopterinvalues for the two inactive assessments.

Significant differences (p <0-01) were alsofound in ESR, plasma C4 and C3d values, butnot plasma C3 values, in the active comparedwith the inactive phases for each of thepatients.

TIME SERIES ANALYSISFigure 3 shows the serial plots of urineneopterin and BIIAG score with time in 14patients who completed more than fivecomplete assessments each. In the analysis ofcovariance, with time and patient as factors,urine neopterin as a covariate and BIIAGscore as the dependent variable, time did nothave a significant effect and could therefore beignored in the multivariate model (table).Subsequent multiple regression analysis,ignoring the effect of time, showed that serialmeasurements of urine neopterin provided asignificant predictor of the BILAG score

Analysis ofcovariance with urine neopterin as a covariate

Degrees offreedom F value Probability >F

Urine neopterin 1 10-8 0 002Time 17 1-4 0-186Patient 13 2-3 0-015Error 67

r%n ,

745

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from

Lim, Muir, Powell

5 10 15

1S

C)c0C.C

4)0..04t)z

SOO

'400 CC)00Q-300 Cr-

0z0-200 z

4)0U)

CO

m

4)

o0

0U)

*10020

a0a00c

0.4-

04)z

4)00U)

0

m

1590 16

14 -

0

°812'C )

.r (3'. 5O"500084)

z a

c0

00c._

04)z

10"

__C

C0f

"200 0

C-U)

4)

Z CO 6_ f::

* 10 15

Time (months)

c

0

0

CL

0.

0

z

4)

oU.)(A

m5

20

C

0 15-

o 0

C

*, 10'

z 5s

0'

C

C.)

C

0

C

0.0

a)z

4)

0

U)

-J

im

2

7 350

300 cC0

250 °C.,-4)

200 Q.0O)z% 150

- I , .I I-I 1000 5 10 1S 20

Time (months)

0

C

0

4,*00C

4)

S00 C.0

4)

-200z

100

5 10 15 20

lime (months)

Figure 3 Serial plots of urine neopterin concentration (pumol/mol creatinine) (0) and the BILAG score (M) with time in the 14 patients included into thetime series model.

(allowing for patient factors) (r = 0-6,p<O 05).

In contrast, analyses of covariance of therelationship between the other test parameters(plasma C3, C4 and C3d, and ESR) andBILAG score, all revealed a significant effect oftime (p < 0-05), but it was not feasible to definefurther this effect oftime because ofthe relativepaucity of the data sets.

DiscussionThis study provides evidence that changes inurine concentrations of neopterin are

significantly correlated with fluctuations indisease activity, scored using the BIIAG index,among individual patients with SLE.

Single time point analysis showed that urineneopterin was significantly increased duringflares in SLE disease activity. Similarly, a

746

4.)0t°> 4-n

@0

m 3-

co

2"-

1 "

4)t-10 I

0U)

0

5-

0"

41)o0CO)(DS

10 15 2

goo

700 C

QS00 00

Go OC

"500Z0.04)

"400

LSoo

600

"500 (C,0O00

3400 Z4)

200

300 Z

20020

-700CQ

- S600 0

C.)4)a

-400 04)z

10"

7.5.4)0U)o 5(9 S'-

-Jm

2.5

0'-

5

4.

CD0 o 3

(3< 2'-

1 '

0'1

I

T2

-I

6 I.0

m A

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from

Urine neopterin for serial monitoring of disease activity in SLE

significant decrease in urine neopterin concen-trations was observed in patients with activeSLE who went into disease remission. In thosepatients whose disease remained active or inremission on second successive assessments,urine neopterin concentrations were found tobe correspondingly greater or smaller com-pared with initial values. In addition, asignificant correlation between urine neopterinconcentration and BILAG score underlinedthe association of urine neopterin with clinicalactivity in SLE.

In the clinical setting, however, serialmeasurements of laboratory parameters arenormally made to monitor patients. A numberof studies have tried to address the temporalrelationship between SLE disease activity andserum concentrations of complement, comple-ment split products, and anti-dsDNA.' 2 10 Inthose studies, only graphical presentations andsingle time point analyses were used to describethe data sets, though the study parameters weremeasured sequentially over a period of manymonths or years, leading to loss of sensitivityand the problem of dependence between teststatistics. When clinical monitoring is basedupon serial quantitative measurements, thedetection and interpretation of abrupt changesin pattern of the data series are of paramountimportance. Often, however, such series of dataare difficult to interpret, as biological variationsand errors arising in the collection, measure-ment and processing of the data and variationswith time itself need to be taken into account.In this prospective study, we adopted timeseries modelling to investigate if serialmeasurements of urine neopterin could predictthe BILAG score and hence SLE diseaseactivity in individual patients.The graphical plots of serial urine neopterin

concentrations and BILAG score in individualpatients, shown in figure 3, suggest closerelationships between the two measurementsand with time. However, analysis of covarianceindicated that the time variable added nothingto the prediction based on the urine neopterincovariate. Consequently, the time variable wasignored in the subsequent regression model.This showed a statistically significant associ-ation between serial concentrations of urineneopterin and BILAG score, with an r value of0-6 (p < 0.05). However, it should be notedthat the entry point of each patient into theanalysis was arbitrarily called "time zero", eventhough patient recruitment into the study tookplace over a period of months. Thisdesignation may have influenced the outcomeof the time series modelling, but is inevitablein the clinical setting; the difficulty in analysingsuch data sets has been highlighted in otherstudies." Replication of these findings with alarger series of patients, using more regularmeasurements and a longer time frame, isrequired to validate our conclusions further.None of the other study paramaters was

found to mirror SLE disease activity aseffectively as urine neopterin, using both singletime point and time series analyses. Plasma C3and C4 did not decrease significantly duringSLE disease flares in individual patients, even

though these two parameters correlated signifi-cantly with the BIIAG score. In contrast, bothESR and plasma C3d failed to correlatesignificantly with BIIAG score and their valuesin individual patients did not increasesignificantly during SLE disease flares. Incorresponding analyses of covariance withserial measurements of ESR, plasma C3, C4,and C3d as covariates, time was found to havea significant effect on the BIIAG score.SLE is a chronic disease with variable

activity characterised by numerous and variedabnormalities of the immune system, in par-ticular upregulation of T lymphocytes. 2Recent attention has been focused on serumsIL-2 receptors, derived from surface boundIL-2 receptors on activated T lymphocytes invivo, as a sensitive and more reliable index ofclinical activity in patients with SLE and otherrheumatic diseases.'3 However, in our experi-ence, there was considerable overlap betweenserum sIL-2 receptor values in SLE patientswith active and inactive disease,4 suggestingthat other factors, particularly renal function,must have an important influence on serumconcentrations of sIL-2 receptors.'4 15Similarly, anti-dsDNA was found to be a poordiscriminant of disease activity in our patientcohort. Thus neither sIL-2 receptor nor anti-dsDNA was included as parameters in thisstudy. Neopterin is specifically produced byhuman macrophages when stimulated byinterferon gamma released from activated Tlymphocytes,'6 17 and is therefore an indirectmarker of the upregulation of the cellularimmune system. Urine neopterin concen-tration has been shown to correlate signifi-cantly with serum sIL-2 receptor concen-trations.'8 It is, then, not surprising to find thaturine neopterin is a more sensitive index ofdisease activity in SLE patients compared withestablished tests.Our calculations were based on the BILAG

index as the established index of SLE diseaseactivity. The BIIAG index is one of threecurrently available activity scales that havebeen shown to be reliable both between andwithin raters,'9 the other two being SystemicLupus Activity Measure (SLAM)20 and SLEDisease Activity Index (SLEDAI).21 Highcorrelations between these three clinical activityscales have been demonstrated, suggesting thatthey are all measuring the same thing.22 Ideally,the reliability of our calculations should havebeen validated against another independentclinical activity scale. However, the inclusion ofone or more of the study parameters in theother clinical activity scales makes themunsuitable for use as instruments to validatethese as parameters of disease activity.

In conclusion, this prospective longitudinalstudy has indicated that serial urine concen-trations of neopterin are significantly asso-ciated with SLE disease activity scored usingthe BIIAG index. Measurement of urineneopterin is non-invasive, relatively simple andinexpensive and appears to be clinically usefulfor isolated assessments of disease activity inaddition to serial monitoring of disease activityin the management of patients with SLE.

747

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from

Lim, Muir, Powell

The authors wish to thank Jim Hughes and Sandy Brown, QMCfor performing the urine neopterin assays and Andrea Thomas,West Midlands Health Authority for computing assistance inthe time series analyses.

1 ter Borg E T, Horst G, Hummer E J, Limburg P C,Kallenberg C G M. Measurement of increases in anti-double-stranded DNA antibody levels as a predictor ofdisease exacerbations in systemic lupus erythematosus: along-term, prospective study. Arthritis Rheumn 1990; 33:634-43.

2 Swaak A J G, Aarden L A, Statius van Eps L, Feltkamp TE W. Anti-dsDNA and complement profiles as prognosticguides in systemic lupus erythematosus. Arthritis Rheume1979;22: 226-35.

3 Wachter H, Fuchs D, Hausen A, Reibnegger G, Werner ER. Neopterin as marker for activation of cellularimmunity: immunologic basis and clinical application.Adv Clin Chem 1989; 27: 81-141.

4 Lim K L, Jones A C, Brown N S, Powell R J. Urineneopterin as a parameter of disease activity in patientswith systemic lupus erythematosus: comparisons withserum sIL-2R and antibodies to dsDNA, erythrocytesedimentation rate, and plasma C3, C4, and C3degradation products. Ann Rheum Dis 1993; 52: 429-35.

5 Symmons D P M, Coppock J S, Bacon P A, et al.Development and assessment of a computerised index ofclinical disease activity in systemic lupus erythematosus.QJMed 1988; 258: 927-37.

6 Tan E M, Cohen A S, Fries J F, et al. The 1982 revisedcriteria for the classification of systemic lupus erythe-matosus. Arthritis Rheum 1982; 25: 1271-7.

7 Slazyk W E, Splerto F W. Liquid-chromatographicmeasurement of biopterin and neopterin in serum andurine. Clin Chem 1990; 36: 1364-8.

8 Crowder M, Hand D. Analysis of repeated measures. In:Cox D, Hinkley D, Rubin D, Silverman B, eds.Monographs on statistics and applied probability. Vol 41.London: Chapman & Hall, 1990.

9 Statistical package for the social sciences'. Chicago: SPSS,1992.

10 Buyon J, Tamerius J, Belmont H M, Abramson S.Assessment of disease activity and impending flare in

patients with systemic lupus erythematosus: comparisonof the use of complement split products and conventionalmeasurements of complement. Arthritis Rheum 1992; 35:1028-37.

11 Smith A F M, West M. Monitoring renal transplants: anapplication of the multiprocess Kalman filter. Biometrics1983; 39: 867-78.

12 Via C S, Handwerger B S. T-cell and B-cell function inlupus. Curr Opin Rheumatol 1992; 4: 630-4.

13 Rubin L A. The soluble interleukin-2 receptor in rheumaticdisease. Arthritis Rheum 1990; 33: 1145-8.

14 Nelson D L, Wagner D K, Marcon L, et al. An analysis ofsoluble IL-2 receptors in human neoplastic disorders. In:Albertini A, Lentant C, Poeletti R, eds. Biotechnology inclinical medicine. New York; Raven Press, 1987.

15 Manoussakis M N, Papadopoulos G K, Drosos A A,Moutsopoulos H M. Soluble interleukin 2 receptormolecules in the serum of patients with autoimmunediseases. Clini Inmunol Imnunopathol 1989; 50: 321 -32.

16 Huber C, Fuchs D, Hausen A, et al. Pteridines as a newmarker to detect human T cells activated by allogeneic ormodified self major histocompatibility complex (MHC)determinants. J Inmunol 1983; 130: 1047-50.

17 Huber C, Batchelor J R, Fuchs D, et al. Immune response-associated production of neopterin. Release frommacrophages primarily under control of interferon-gamma. J Exp Med 1984; 160: 310-6.

18 Lim K L, Atta M, Hughes J, Brown N S, Powell R J. Serumlevels of soluble interleukin 2 receptor (IL-2R) and urineneopterin with disease activity in patients with systemiclupus erythematosus (SLE). Arthritis Rheumn 1992; 35:S209.

19 Hay E, Gordan C, Emery P. Assessment of lupus; where arewe now?Ann Rheum Dis 1993; 52: 169-72.

20 Liang M, Socher S, Roberts W, Esdaile J. Measurement ofsystemic lupus erythematosus activity in clinical research.Arthritis Rheum 1988; 31: 817-25.

21 Bombardier C, Gladman D D, Urowitz M B, Caron D,Chang C H. Derivation of the SLEDAI. A disease activityindex for lupus patients. The Committee on PrognosisStudies in SLE. Arthritis Rheum 1992; 35: 630-40.

22 Gladman D D, Goldsmith C H, Urowitz M B, et al.Crosscultural validation and reliability of 3 disease activityindices in systemic lupus erythematosus. 7 Rheumatol1992; 19: 608-11.

748

on June 26, 2020 by guest. Protected by copyright.

http://ard.bmj.com

/A

nn Rheum

Dis: first published as 10.1136/ard.53.11.743 on 1 N

ovember 1994. D

ownloaded from