increased albumin excretion in diabetes

J Clin Pathol 1990;43:3-8

OCCASIONAL ARTICLE: REVIEW

Increased albumin excretion in diabetes

J C Townsend

In 1964 Keen and Chlouverakis used aradioimmunoassay to measure the concentra-tion of albumin in the urine of diabetic subjec-ts.' Although the urine samples were negativefor protein by standard clinical tests, the rate ofalbumin excretion in the study group wasgreater than in normal controls. Many studiesand review articles have subsequently addres-sed the clinical implications and patho-physiological mechanisms of these findings.There has been little attempt, however, toassess the consequences of these activities onpathology services.The aim of this review is to give an historical

perspective to the main clinical developmentsof what is often called "microalbuminuria",outline the various analytical methods availa-ble, discuss some of the technical difficulties,and consider how pathologists may bestprovide a worthwhile service to clinicians.

Department ofClinical Chemistry,Royal HallamshireHospital, SheffieldJ C TownsendCorrespondence to:Dr J C Townsend,Department of ClinicalChemistry, RoyalHallamshire Hospital,Glossop Road, Sheffield S1O2JF, England.Accepted for publication31 August 1989

Terminology"Microalbuminuria" has been used to refer toalbuminuria that is greater than normal(2 5-26 mg/24 hour9) but undetectable by con-

ventional Albustix (less than 250 mg/l). It hasbeen variously defined as an albumin concen-

tration of between 30 and 140 ,ug/ml3; an

albuminexcretion rate (AER) of between 15and 150 pg/minute4; 20 and 200 ug/minute5; 30and 300 mg/24 hour6; and between 26 and250 mg/24 hour 2.

Other terms in common use are "incipientdiabetic nephropathy" (between 20 and 200 ug/minute) and "clinical or overt diabeticnephropathy"7 (greater than 200 ug/minute).The proposed definitions of both terms rely onexcluding other non-diabetic causes ofprotein-uria and should show an increased albuminexcretion rate on two out of three occasionswithin a six month period. There is no univer-sal agreement on these rigid criteria, whichimply definite disease entities rather thanevolutionary stages of one disease process.8The terms "microalbuminuria" and

"macroalbuminuria" won't be used in thisarticle. Both of these expressions are mislead-ing because they suggest that albumin ispresent in different molecular weight forms.Although this is true in some circumstances, innephrotic syndrome910 where polymeric andfragmentary forms ofalbumin occur, for exam-ple, in diabetes the "micro" and "macro" infact refer to the amount of albumin excreted.The recently proposed neologism of "paucial-buminuria"11 also fails to provide additionalclarification to the simple use of "albumin-uria"'2 which is the term used in this article.

Important studies

Three prospective studies of insulin dependentdiabetics have shown the prognostic value of araised urinary albumin excretion rate (AER)with regard to subsequent nephropathy. All theauthors arrived at the same conclusion-thatthere is a threshold rate of albumin excretionwhich identifies those patients likely to developdiabetic nephropathy. The important dif-ference among these studies, however, is thateach identifies a different threshold rate: Math-iesen, 70 ug/minute"; Viberti, 30 ,ug/minute"4;and Mogensen, 15 ,ug/minute'. This dis-agreement may be explained by differences inthe length of study, methods of urine collectionand statistical analyses.

In non-insulin dependent patients increasedalbumin excretion is reported not only topredict subsequent renal disease as evidencedby clinical proteinuria but also early mortality.3Albumin excretion now seems to be associatedwith cardiovascular disease in this group ofpatients.5 16 An important recent observation isthe association of increased albumin excretionwith coronary and peripheral vascular diseasein non-diabetic subjects."7

Studies have shown that control ofglycaemia,"3 blood pressure,920 and diet2l canreduce albumin excretion. In a group of 36randomised patients strict glycaemic control,judged by glycated haemoglobin and achievedby continuous subcutaneous insulin infusion,resulted in a significant reduction in the rate ofprogression from an AER of between 30-300mg/24 hours to clinical diabetic nephropathy(AER > 300 mg/24 hours) over a period of twoyears. 8Three questions arise from the above

observations. Does a raised albumin excretionactually predict later nephropathy or simplydetect disease at an earlier stage? Do manipula-tions which reduce the rate of albumin excre-tion really change the progression of renaldisease? Should urine albumin be measuredroutinely? These points have been addressedby Mogensen, who has drawn attention to thesix or seven year delay before the results of twolarge studies (diabetes control and complica-tions trial (DCCT) and the British microal-buminuria collaborative study) are available.Even then the question of whether strictmetabolic control will prevent diabeticnephropathy may not be answered. In themeantime he recommends guidelines forpatient management based on the measure-ment of blood pressure, glycated haemoblobin,and urinary albumin.22

3

on February 3, 2022 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.43.1.3 on 1 January 1990. Dow

nloaded from

http://jcp.bmj.com/

Townsend

MechanismsThe pathophysiology of diabetic proteinuriahas received considerable attention.223 Thereare, however, certain conflicting reports wor-thy of discussion. There should be a soundunderstanding of the mechanisms of albumin-uria before attempting to interpret the resultsof measurements. Albumin in urine may comefrom renal and non-renal sources. Urinarytract infection has been checked as a source ofproteinuria in clinical studies, but other factorsincluding mechanical problems of the ureters,drug effects,24 and especially perineal con-tamination in women25 have received littleattention.The renal excretion of any substance

represents a balance between glomerular filtra-tion, tubular reabsorption, and tubular secre-tion. It is unlikely that any active secretion ofalbumin occurs in the kidney26 so that excretionis determined by the processes of filtration andreabsorption. Filtration will be influenced byglomerular filtration rate and the size andelectrical charge of the molecule. More than95% of filtered albumin is reabsorbed by anearly saturated tubular mechanism (directevidence from animal studies27).The case for increased excretion of albumin

in diabetic subjects resulting from increasedfiltration rather than decreased tubular reab-sorption' is based on the observation that theexcretion of fl-2-microglobulin (an indicator ofrenal tubular damage) is not increased in thepresence of increased albumin excretion. Theexcretion of y globulin, which differs in elec-trical charge and molecular size, is alsoincreased, and the relative amounts of albuminand y-globulin change according to the degreeof proteinuria. This suggests a progressivechange in glomerular selectivity.The excretion of fl-2-microglobulin, how-

ever, may not be the best marker of tubularfunction. Several studies have investigated theurinary excretion of other low molecularweight proteins (K light chains, retinol bindingprotein, and a-l-microglobulin). In 1979Lopes-Virrella described a predominant excre-tion of low molecular weight proteins, suggest-ing tubular dysfunction, in juvenile diabeticswith short disease duration.28 Later work ondiabetics without albuminuria showed an in-creased excretion of K light chains which wasindependent of the excretion of ,B-2-micro-globulin.29 Recently the excretion of retinolbinding protein, a low molecular weightmarker of tubular function, has been shown tobe increased in diabetics with normal albuminexcretion.30 A study of childhood diabetesconcluded that fJ-2-microglobulin was anunsatisfactory marker oftubular function com-pared with the excretion of a-1-microglobulinand K light chains which were significantlyincreased in diabetic children.31 All ofthis worksuggests that there are disturbances of tubularfunction which precede albuminuria, but itdoes not resolve the questions of whichmechanism is responsible for albuminuria orwhether measurement of proteins other thanalbumin would provide more useful informa-tion for predicting nephropathy.

Diabetic subjects have increased concentra-tions of glycated proteins. Non-enzymaticallyglycated albumin is more anionic (pl < 4 7)than the native protein, and this resultingchange of charge could result in altered renalhandling of glycated albumin and explain someof the changes in AER associated withimproved glycaemic control. Studies of theexcretion rate of glycated proteins have usedvarious techniques and have produced con-tradictory results-namely, a preferentialretention or excretion of glycated protein.Gragnoli et al reported an increased renalelimination of glycated protein (expressed asnmol 5-hydroxy methyl furfural/mg protein) indiabetics which was related to the degree ofproteinuria.32 Ghiggeri et al, using isoelectricfocusing of chromatographically purifiedalbumin, concluded that glycation was themain determinant of albumin excretion indiabetics with normal or slightly raisedalbumin excretion.33 In contrast, Kverneland etal reported a preferential retention of glycatedalbumin (expressed as ,umol furosine/,umolalbumin), which could be explained by a loss ofanionic charge on the glomerular basementmembrane.34There is an increased transcapillary escape

rate (TER) of albumin in diabetic subjects.6This was established by following the declinein plasma radioactivity after an intravenousinjection ofradiolabelled albumin. The TER ofalbumin is increased by poor metabolic controland hypertension35 and is increased in patientswith increased albuminuria. The loss ofalbumin from the vascular compartment,however, is not explained by the increased lossof albumin in, the urine. Long term diabeticswithout albuminuria do not have increases inTER, suggesting that a common mechanism,probably endothelial dysfunction,36is respon-sible for both renal and extra-renal capillaryleakage.Whichever mechanism is responsible for the

increased albumin excretion, the mainstays ofclinical management will be control of hyper-tension and glycaemia together with the possi-ble restriction of dietary protein. Recent workinvestigating renal biopsies emphasises theimportance of hypertension by finding thatestablished abnormalities of glomerular struc-ture are only accurately predicted by thepresence of increased AER together withhypertension, reduced creatinine clearance, orboth.37 A positive family history of hyperten-sion or an increased red blood cell Na+/Li+exchange may also identify patients at risk ofnephropathy before any increase in urinaryalbumin excretion.38

AnalysisTo fulfil the requirements of adequate sen-sitivity, specificity, and reproducibility forurinary albumin in the range of interest (2-200mg/l) some type of immunological method isnecessary. There is a wide choice of availablemethods but four general approaches pre-dominate. Radioimmunoassay (RIA); non-isotopic immunoassay; nephelometry and

4



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


immunoturbidimetry (IT); radial immunodif-fusion (RID). Other less common methodsinclude zone electrophoresis,39 liquidchromatography,40 and latex particle coun-ting.4"The idea that diabetic nephropathy can be

predicted by a urinary albumin concentrationexceeding a critical value has naturally led tothe development of semiquantitative testswhich can identify urines with more than thatcritical albumin concentration and are suitablefor use in the clinic or ward side-room.

METHODSRadioimmunoassayRIA has been established since 196342 and hasbeen used in many of the important clinicalstudies. Several manufacturers now producekits specifically for the determination of urin-ary albumin, typically with analytical ranges inthe order of 1-100 mg/l, which may be expan-ded by sample dilution. The latter requiressome prior knowledge of the likely urinealbumin concentration. As an increased AERwas first shown by RIA there has been atendency for it to be used as the standardagainst which proposed methods are com-pared, although its performance is not alwayssatisfactory.43

Non-isotopic immunoassayThe use of non-isotopic labels overcomes twoof the main objections of RIA-the use ofradioactive materials and the short shelf life ofthe labelled reagent. For example, fluoresceinhas successfully replaced '25iodine as the labelin an otherwise identical assay system.44Enzyme linked immunoassay (ELISA) tech-

niques are simple but often require longincubation times with multiple washing andreagent addition steps. Few laboratories haveautomated equipment capable of washing, dis-pensing, and plate reading and therefore mightconsider the method too labour intensive. Themethod described by Fielding et al (an exampleof the sandwich type) requires a 250-foldsample dilution, five reagent addition steps,five hours of incubations and extra time forplate washing.45 Various modifications of thisapproach have been used' but do not achievethe same large assay range (3-1000 ,ug/l) des-cribed by Fielding. Competitive ELISA meth-ods have been developed either with immobil-ised antigen4748 or enzyme labelled antigen.49The labelled antigen method is attractivebecause it is fast (one hour), needs only twowashing steps, does not need sample pre-dilution and has a range of 0-9-200 mg/l whichis suitable for the concentrations of interest.This approach produces data which can beprocessed by conventional saturation analysiscurve fitting procedures. "Edge effects" arerecognised as a problem with microtitre platemethods454750 and refer to the discordantresults that obtain in the outer wells of themicrotitre plate. In practice this means betterprecision is achieved if only the inner 60 wellsof a 96 well plate are used, a considerable

sacrifice. This problem may be explained byinherent plate variability. Considering theincreasing use of microtitre plate methodsthere have been surprisingly few rigorousinvestigations of these phenomena.5'

ImmunoturbidimetrySpencer and Price investigated kineticimmunoturbidimetry and found its perfor-mance comparable with nephelometry andtherefore suitable for measuring albumin inurine at low concentrations.52 The widespreaduse of centrifugal fast analysers has allowedmany laboratories to have a simple, rapid, andreasonably economical method capable of deal-ing with large batches. The only requirementfor sample pretreatment is a centrifugation stepto clear the urine sample of potentially interfer-ing particulate matter. There is no lack of eitherpublished"55 or commercially available meth-ods. Most methods cover at least the 5-80 mg/lrange and offer within and between batchprecision of less than 5 and 1000, respectively.Several aspects of the technique are critical,especially the effect of polymer enhancementand antibody avridity.5' Other errors may resultfrom the choice of instrument wavelength andthe associated urine blank absorbance values.56There is also a report of significantly differentresults being generated for some urines; depen-ding on the species source of anti-albuminantibody, rabbit anti-serum produced the mostreproducible results and the best correlationwith an established commercial RIA.57

Radial immunodiffusionRadial immunodiffusion (RID) methods havebeen used in clinical studies," and commercialsystems are available. RID methods requirelong incubation times (often overnight) whichwould not permit a same day or "in clinic"service, but it is unlikely that urine albuminassays would ever be considered as urgentinvestigations.Watts et al give a comprehensive comparison

of RIA, RID, IT and ELISA (sandwichtype).50 They tested urine samples fromdiabetic patients in the range 1-120 mg/l andfound no systematic difference between RIAand RID. They reported IT and ELISA asgiving consistently lower values than RIA withgreater random error, the errors increasing atgreater albumin concentrations. ELISA gavethe poorest between batch performance (CV 8-10%), and a degree of bias (a problem alsoencountered by others58) which was consideredto be clinically significant and likely to result inthe misclassification of patients. They foundIT to be the most expensive method for operat-ing cost but this was assessed against an "inhouse" RIA method, which is probably manytimes less expensive than a commercial kit.They concluded that despite the greater tech-nical skill required, given the suitable staff, theattractions of low capital cost and the flexibilityof batch size would make RID their method ofchoice.

5



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Townsend

Semiquantitative methodsTests suitable for the clinic or ward side roominclude latex agglutination of both the directand indirect types.59" In the direct test latexparticles coated with anti-albumin antibody aremixed with free antibody and test urine sample.This results in visible semiquantitative agglu-tination if the concentration is between 24 and166 mg/l. Higher concentrations will give noagglutination and must be distinguished fromvery low values by a conventional Albustixtest.6' The indirect test uses latex particlescoated with albumin added to a mixture of testsample and anti-albumin antibody resulting invisible agglutination at concentrations of lessthan 40 mg/l.Another approach is a bromophenol dye

binding tablet. Urine is dropped on to thesurface of the tablet which is then washed bydrops of distilled water. A positive reaction isgiven by a blue-green colour which is com-pared with a standard. This simple method hasthe major disadvantage ofnot being specific foralbumin but reports suggest that it producesfew false negative results, when used to detecturinary albumin concentrations greater than 40mg/l and is therefore a reasonable "screening"method.6263A laboratory assay using a colorimetric

protein method has been described which givesa similar performance to other semiquan-titative methods and has the advantages ofspeed and low cost.64

METHOD INDEPENDENT FACTORSPotential sources of pre-analytical errorinclude the storage conditions ofurine samples.Standards must be stable and accuratelyprepared. It is also important to perfoim theanalysis on a representative sample and sominimise biological variations.

Albumin adherenceIt is known that polypeptides in urine mayadsorb on to the plastic ofcontainers65 but thereis no consensus regarding the behaviour ofalbumin in urine. Fielding et al included I%(v/v) of rabbit serum in the urine sample beforestorage to prevent adsorption.45 Many reportsdo not make specific reference to this phen-omenon but Feldt-Rasmussen et al state thatcontainers were tested for albumin adherence.46Unfortunately, they do not specify which con-tainers, how the adherence was investigated, orwhat results were produced. Torffvit et al alsostate that albumin adsorbs to plastic tubes andadded 0-0010% (v/v) of Tween 20 (an ionicdetergent) to prevent the phenomenon.47 Urinestored with the addition ofTween 20 gave up to50% greater albumin concentration than ifstored without the addition of the detergent.This problem of adherence has received

more attention regarding the preparation ofstandards. Gosling reported an overestimationof test albumin concentration unless bovineserum albumin was added to the standardsolutions to prevent adsorption onto the poly-styrene containers." The interpretation of this

finding was complicated by later work whichshowed up to 30% cross reactivity between thatparticular antibody and bovine serumalbumin.67" The problem of "apparent"adherence may be overcome by adding TritonX-100 (a detergent) to the albumin standards.69Unfortunately, there does not seem to havebeen a definitive study investigating theadherence of albumin to different material.This is of particular importance because theforgoing observations could be explained bythe loss of immunoreactivity of the standards.Silver et al discuss the possible 20% reductionof immunoreactivity in lyophilised comparedwith monomeric albumin.44 This is becauseaggregates form during the freeze dryingprocess and there may be batch to batchvariation of the amount of aggregate produced.

Storage conditionsThere is no standard method either for storingurine or for sample preparation prior to albu-min assay. The commonest storage tempera-ture is - 20°C, and few studies actually statehow long the samples have been maintained atthe chosen storage temperature. Some workershave added TRIS buffer to alkalinise theurine70 while others have added detergents,"proteins,45 or sodium azide preservative.39Some centrifuge urine after thawing"477' butmany others do not record whether this wasdone. The relevance of these different ap-proaches is that albumin may precipitate fromurine centrifuged after storage at - 20'C.72This observation has been endorsed7375 andrefuted.76 There is evidence that suchprecipitation may occur with other proteins aswell as albumin and is prevented by adjustingthe urine pH to neutral.7778

Choice of collectionThe easiest sample to obtain is a random(untimed) collection when a patient attendsclinic. This sample is likely to show the greatestvariations of albumin concentration, influen-ced by factors like previous exercise,79metabolic state,80 and rate of urine produc-tion." Short collection periods (20 minutes)have been used to investigate specific effectssuch as water loading, but the supervisionrequired would not be practicable outsidemetabolic wards. Overnight collections haveshown less variation than companion diurnalcollections82 and are less inconvenient forpatients than 24 hour collections. Coefficientsof variation are about 40%, however,83 whichare similar to those found for 24 hour collec-tions.84 A logical choice would be an untimedfirst morning collection which combines theconvenience of a small sample with the smallervariation of an overnight collection.

Children represent a special case becausetimed urine collections are so difficult to collectand there are considerable differences of rela-tive age, body mass, and height among sub-jects. Large (up to 10-fold) within subjectvariations ofalbumin excretion are described innormal children, emphasising the need for

6



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/


repeated determinations before decidingwhether albumin excretion is increased.85Marshal et al found that AER in both diabeticand normal children was similar (median 2 5and 2-4 jug/ml, respectively, overnight col-lection) and less than adult values (median 4-3yg/ml).86 They concluded that cut off valuespredictive of nephropathy in adults might notbe appropriate in children. They also showed apositive correlation between AER and age(girls having higher values than boys).Mathiesen et al, however,87 did not find anyassociation with age or sex.

UnitsA variety of units are used to express albuminexcretion. The choice is influenced by theduration of the urine collection: thus foruntimed samples albumin excretion may beexpressed simply as a concentration or as a ratioof albumin to creatinine concentrations; 24hour collections are usually expressed as mg/24hour; timed collections of less than 24 hours areoften expressed as ug/minute. In the case of a''rate" two errors are introduced from volumeand time estimates and in the case of the "ratio"any error in the creatinine estimation isintroduced. The inaccuracies of 24 hour urinecollections are well documented and the attrac-tion of expressing albumin excretion as a ratiowith creatinine ignores the variation of creatin-ine excretion both within and among subjects.Both Howey88 and Hutchison89 found littleadvantage in the albumin: creatinine ratio andprefer to use the uncomplicated albumin con-centration of the first morning urine sample.

ConclusionsLaboratories are likely to receive an increaseddemand for urinary albumin analyses, and ifthis is in the form of a service commitmentrather than as part of limited research projectspathologists must decide how to provide asatisfactory service. A laboratory's choice ofmethod is bound to be influenced by existingequipment, local expertise, and cost. The factthat over 20 years of experience have notproduced one preferred method suggests thatany of the main four quantitative approaches issatisfactory, although each has certain dis-advantages. Perhaps the more important con-sideration is whether all patients should have aquantitative assessment if reliable semiquan-titative tests are available. This point has majorcost implications because of the need to makeseveral determinations to gain a realisticmeasure of any individual's albumin excretion.In effect a more sensitive "Albustix" approachcould identify those patients in whom urinealbumin quantitation would influence man-agement.Whatever analytical method is used, sensible

comparisons of the results from different cen-tres will continue to be difficult in the absenceof a standard approach. The easiest aspect tocontrol is probably the most important-namely, the choice of urine sample. A firstmorning, untimed specimen combines thelower overnight variation with the convenienceof a random sample. It avoids any need for

timing on behalf of patient or staff and is notinconvenient to repeat should several differentdeterminations be required. Having obtainedthe sample in a straightforward manner, itseems sensible to express the results as simplyas possible-that is, as a concentration. Theredo not seem to be any clear advantages usingthe creatinine ratio which requires an extraassay and therefore time and cost. Laboratoriesusually batch the more specialised tests and ifthere is a delay in analysis the storage condi-tions and sample pretreatment may influencethe final result, just as may the choice andpreparation of standards.The prevalence of diabetes mellitus and the

high incidence of renal failure in the diabeticpopulation ensure a continued interest in theprevention or reduction of diabetic nephro-pathy. At present the exact role of measuringurinary albumin in patient management is notclear but rests on the expectation that a reduc-tion of albumin excretion will be associatedwith a reduction of the progression of renaldisease.

1 Keen H, Chlouverakis C. Urinary albumin excretion anddiabetes mellitus. Lancet 1964;ii:1 155-6.

2 Viberti GC, Wiseman MJ. The kidney in diabetes: sig-nificance ofthe early abnormalities. Clin Endocrinol Metab1986;51 :753-82.

3 Mogensen CE. Microalbuminuria predicts clinical protein-uria and early mortality in maturity onset diabetes. N EnglJMed 1984;310:356-60.

4 Mogensen CE, Christensen CK. Predicting diabeticnephropathy in insulin-dependent patients. N EnglJ Med1984;311:89-93.

5 Mogensen CE, Chachati A, Christensen CK, et al. Micro-albuminuria: an early marker of renal involvement indiabetes. Uremia Investigation 1986;9:85-95.

6 Feldt-Rasmussen B. Increased transcapillary escape rate ofalbumin in type I(insulin dependent) diabetic patientswith microalbuminuria. Diabetologia 1986;29:282-6.

7 Mogensen CE. Microalbuminuria. Recent Adv Diabetes1986;2:139-50.

8 Mogensen CE. Microalbuminuria as a predictor of clinicaldiabetic nephropathy. Kidney Int 1987;31:673-89.

9 Wiggins RC, Kshrisagar B, Kelsch RC, Wilson BS. Frag-mentation and polymeric complexes of albumin in humanurine. Clin Chim Acta 1985;149:155-63.

10 Birtwistle RJ, Hardwicke J. The role of urea in albumindimerisation in nephrotic urines. Clin Chim Acta 1985;151:41-8.

11 Trivin F, Giraudet P. "Microalbuminuria" or "pauci-albuminuria"? Clin Chem 1988;34:209-10.

12 Devgun MS. What's in a name? Microalbuminuria/albuminuria. Diabetic Med 1988;5:93.

13 Mathiesen ER, Oxenboll B, Johansen K, Svendsen PA,Deckert T. Incipient nephropathy in type 1 (insulindependent) diabetes. Diabetologia 1984;26:406-10.

14 Viberti GC, Jarrett RJ, Mahmud U, Hill RD, ArgyropoulosA, Keen H. Microalbuminuria as a predictor of clinicalnephropathy in insulin-dependent diabetes mellitus. Lanc-et 1982;i:1430-2.

15 Schmitz A, Vaeth M. Microalbuminuria: a major risk factorin non-insulin dependent diabetes. A 10 year follow-upstudy of 503 patients. Diabetic Med 1988;5:126-34.

16 Mattock MB, Keen H, Viberti GC, et al. Coronary heartdisease and urinary albumin excretion rate in type 2 (non-insulin-dependent) diabetic patients. Diabetalogia 1988;31:82-7.

17 Yudkin JS, Forrest RD, Jackson CA. Microalbuminuria as a

predictor of vascular diseases in non-diabetic subjects.Lancet 1988;ii:530-3.

18 Feldt-Rasmussen B, Mathiesen ER, Deckert T. Effect oftwo years of strict metabolic control on progression ofincipient nephropathy in insulin-dependent diabetes.Lancet 1986;ii: 1300-4.

19 Christensen CK, Mogensen CE. Antihypertensive treat-ment: long term reversal of progression of albuminuria inincipient diabetic nephropathy. A longitudinal study ofrenal function. J Diabetic Comp 1987;1:45-52.

20 Hommel E, Mathiesen E, Edsberg B, Bahnsen M, ParvingH-H. Acute reduction of arterial blood pressure reducesurinary albumin excretion in type I(insulin dependent)diabetic patients with incipient nephropathy. Diabetalogia1986;29:21 1-15.

21 Cohen D, Dodds R, Viberti GC. Effect of protein restrictionin insulin dependent diabetics at risk of nephropathy. BrMed J 1987;294:795-8.

22 Mogensen CE. Management of diabetic renal involvementand disease. Lancet 1988;i:867-9.

23 Mogensen CE, Christensen CK, Vittinghus E. The stages indiabetic renal disease with emphasis on the stage ofincipient diabetic nephropathy. Diabetes 1983;32(suppl

7



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

Townsend

2):64-78.24 Waller KV, Ward KM, Mahan JD, Wismatt DK. Current

concepts in proteinuria. Clin Chem 1989;35:755-65.25 Stoica GH, Purice S, Cheta D, Serban M, Balea M, Michiu

V. Albumin to creatinine ratios in early morning samplesofurine from healthy persons and patients without clinicalproteinuria. Revue Roumaine de Medicine Interne 1984;22:19-28.

26 Pesce AJ. First MR. Proteinuria: an integrated review. NewYork: Marcel Dekker Inc, 1979.

27 Bourdeau JE, Carone FA. Protein handling by the renaltubule. Nephron 1974;13:22-34.

28 Lopes-Virella MF, Virella G, Rosebrock G, Sagel J,Gonzalez J, Colwell J. Early diagnosis of renal malfunc-tion in diabetics. Diabetalogia 1979;16:165-71.

29 Teppo AM, Groop L. Urinary excretion of plasma proteinsin diabetic subjects. Increased excretion of Kappa lightchains in diabetic patients with and without proliferativeretinopathy. Diabetes 1985;34:589-94.

30 Holm J, Hemmingsen L, Neilsen NV, Thomsen M.Increased urinary excretion of the retinol-binding proteinin insulin dependent diabetes mellitus in the absence ofmicroalbuminuria. Clin Chim Acta 1987;170:345-50.

31 Walton C, Bodansky HJ, Wales JK, Forbes MA, CooperEH. Tubular dysfunction and microalbuminuria ininsulin dependent diabetes. Arch Dis Child 1988;63:244-9.

32 Gragnoli G, Signorini AM, Tanganelli I. Non-enzymicglycosylation of urinary proteins in type I (insulin depen-dent) diabetes: correlation with metabolic control and thedegree of proteinuria. Diabetalogia 1984;26:411-4.

33 Ghiggeri GM, Candiano G, Delfino G, Queirolo C. Elec-trical charge of serum and urinary albumin in normal anddiabetic humans. Kidney Int 1985;28:168-77.

34 Kverneland A, Feldt-Rasmussen B, Vidal P, et al. Evidenceof changes in renal charge selectivity in patients with type1(insulin-dependent) diabetes mellitus. Diabetalogia1986;29:634-9.

35 O'Hare JA, Ferriss JB, Twomey B, O'Sullivan DJ. Poormetabolic control, hypertension and microangiopathyindependently increase the transcapillary escape rate ofalbumin in diabetes. Diabetalogia 1983;25:260-3.

36 Jensen T, Feldt-Rasmussen B, Bjerre-Knudsen J, DeckertT. Features of endothelial dysfunction in early diabeticnephropathy. Lancet 1989;i:461-3.

37 Viberti GC, Bending JJ. Preventive approach to diabetickidney disease. Contrib Nephrol 1988;61:91-100.

38 Chavers BM, Bilous RW, Ellis EN, Steffes MW, MauerSM. Glomerular lesions and urinary albumin excretion intype I diabetes without overt proteinuria. N Engl J Med1989;320:966-70.

39 Vesterberg 0. Quantification of albumin in urine by a newmethod: zone immuno-electrophoresis assay (ZIA). ClinChim Acta 1981:113:305-10.

40 Turpeinen U, Koivunen E, Stenman U. Liquidchromatographic determination of ,B2 microglobulin, alacid glycoprotein and albumin in urine. Clin Chem 1987;33:1756-60.

41 Cambiaso CL, Collet-Cassart D, Lievens M. Immunoassayof low concentrations of albumin in urine by latex particlecounting. Clin Chem 1988;34:416-18.

42 Keen H, Chlouverakis C. An immunoassay for urinaryalbumin at low concentrations. Lancet 1963;ii:913-4.

43 Watts GF, AlbanoJDM, Bennett JE, MorisRW, Shaw KM,Polak A. Assessment of new radioimmunoassay kit fordetermining urinary albumin at low concentrations: com-parison with radial immunodiffusion. J Clin Pathol 1986;39:1151-4.

44 Silver A, Dawnay A, Landon J, Cattell WR. Immunoassaysfor low concentrations of albumin in urine. Clin Chem1986;32:1303-6.

45 Fielding BA, Price DA, Houlton CA. Enzyme immunoassayfor urinary albumin. Clin Chem 1983;29:355-7.

46 Feldt-Rasmussen B, Dinesen B, Deckert M. Enzymeimmunoassay: an improved determination of urinaryalbumin in diabetics with incipient nephropathy. Scand JClin Lab Invest 1985;45:539-44.

47 Torffvit 0, Wieslander J. A simplified enzyme-linkedimmunosorbent assay for urinary albumin. Scand J ClinLab Invest 1986;46:545-8.

48 Townsend JC. A competitive immunoenzymometric assayfor albumin in urine. Clin Chem 1986;32:1372-4.

49 Chesham J, Anderton SW, Kingdon CFM. Rapid, com-petitive enzymeimmunoassay for albumin in urine. ClinChem 1986;32:669-71.

50 Watts GF, Bennett JE, Rowe DJ, et al. Assessment ofimmunochemical methods for determining low concen-trations of albumin in urine. Clin Chem 1986;32:1544-8.

51 Kricka LJ, Carter TJN, Burt SM, et al. Variability in theadsorption properties of microtitre plates used as solidsupports in enzyme immunoassay. Clin Chem 1980;26:741-4.

52 Spencer K, Price CP. Kinetic immunoturbidimetry: Theestimation of albumin. Clin Chim Acta 1979;95:263-76.

53 Ross IS, Stroud RE, Scheymans J. An immunotur-bidimetric assay for microalbuminuria in normal anddiabetic urine. Transplant Proc 1986;18:1654-5.

54 Harmoinen A, Ala-Houhala I, Vuorinen P. Rapid andsensitive immunoassay for albumin determination inurine. Clin Chim Acta 1985;149:269-74.

55 Kearney EM, Mount JN, Watts GF, Slavin BM, Kind PRN.Simple immunoturbidimetric method for determiningurinary albumin at low concentrations using Cobas-Biocentrifugal analyser. J Clin Pathol 1987;40:465-8.

56 Harmoinen A, Vuorinen P, Jokela H. Turbidimetricmeasurement of microalbuminuria. Clin Chim Acta 1987;166:85-9.

57 Shukla PS, Endeman HJ, Kortlandt W. Effect of fivedifferent antisera on the immunoturbidimetric determina-tion of urinary albumin. Clin Chem 1988;34:430.

58 Alinovi R, Mutti A, Bergamaschi E, Franchini I. Com-petitive enzyme-linked immunosorbent assay (CELISA)of urinary albumin. Clin Chem 1988;34:993-4.

59 Viberti GC, Vergani D. Detection of potentially reversiblediabetic albuminuria. A three drop agglutination test forurinary albumin at low concentration. Diabetes 1982;31:973-5.

60 Paoli C, Bardelli F, Tarli P, Gragnoli G, Tanganelli I,Signorini AM. A simple latex agglutination test forurinary albumin screening. Clin Chim Acta 1987;166:67-71.

61 Hodgson BW, Watts GF. A simple side-room test to screenfor "microalbuminuria" in diabetes mellitus. Ann ClinBiochem 1987;24:581-4.

62 Linton AS, Rowe DJF. Ames "Microbumintest" evaluated,and its correlation with total protein and albumin concen-tration in urine. Clin Chem 1988;34:1927-8.

63 Schmitz A. Microalbutest: a new screening method fordetecting microalbuminuria in diabetes mellitus. UremiaInvest 1986;9:79-84.

64 Phillipou G, James SK, Seaborn CJ, Phillips PJ. Screeningfor microalbuminuria by use of a rapid, low costcolorimetric assay. Clin Chem 1989;35:456-8.

65 Livesey JH, Donald RA. Prevention of adsorption lossesduring radioimmunoassay of polypeptide hormones:effectiveness of albumins, gelatin, caseins, Tween 20 andplasma. Clin Chim Acta 1982;123:193-8.

66 Gosling P, Lawson N, Taylor AJ. Limitations in theimmunoturbidimetry of urinary albumin. Clin Chem1986;32: 1240.

67 Smith G, Wilcox AH, Barron JL. More on determination ofurinary albumin by centrifugal analysis. Clin Chem 1987;33:1497.

68 Bakker AJ. More on immunoturbidimetry of urine albumin.Clin Chem 1987;33:198.

69 Bakker AJ. Immunoturbidimetry of urinary albumin:prevention of adsorption of albumin; influence of otherurinary constituents. Clin Chem 1988;34:82-6.

70 Marshall SM, Alberti KGMM. Screening for early diabeticnephropathy. Ann Clin Biochem 1986;23:195-7.

71 Miles DW, Mogensen CE, Gundersen HJG.Radioimmunoassay for urinary albumin using a singleantibody. Scand JLab Invest 1970;26:5-1 1.

72 Townsend JC. Specimen preparation for microalbuminuriaassay. Clin Chem 1986;32:559.

73 Erman A, Rabinov M, Rosenfeld J. Albumin determinationin frozen urines-underestimated results. Clin Chim Acta1988;174:255-62.

74 Hutchison AS, O'Reilly D St J. More on urinary albuminand creatinine. Clin Chem 1987;33:740.

75 Elving LD, Bakkeren AJM, Jansen MJH, AngelinoCM de Kat, de Nobel E, Munster PJJ. Screening formicroalbuminuria in patients with diabetes mellitus:frozen storage of urine samples decreases their albumincontent. Clin Chem 1989;35:308-10.

76 Silver AC, Dawnay A, Landon J. Specimen preparation forassay of albumin in urine. Clin Chem 1987;33:199-200.

77 Townsend JC, Blair PJ, Forrest ARW. Effect of storage pHon precipitation of albumin from urine from diabetics.Clin Chem 1988;34:1355-6.

78 Townsend JC, Sadler WA, Shanks G. The effect of storagepH on the precipitation of proteins in deep frozen urinesamples. Ann Clin Biochem 1987;24:111-2.

79 Viberti GC, Jarrett RJ, McCartney M, Keen H. Increasedglomerular permeability to albumin induced by exercise indiabetic subjects. Diabetalogia 1978;14:293-300.

80 Christensen CK, Schmitz 0, Pedersen EB, Alberti KGMM,Mogensen CE. Effect of 3-hydroxybutyrate infusion onurinary protein excretion in healthy man. ScandJClin LabInvest 1986;46:239-43.

81 Viberti GC, Mogensen CE, Keen H, Jacobsen FK, JarretRJ, Christensen CK. Urinary excretion of albumin innormal man: the effect of water loading. Scand J Clin LabInvest 1982;42:147-51.

82 Chachati A, von Frenckell R, Foidart-Willems J, Godon JP,Lefebvre PJ. Variability of albumin excretion in insulindependent diabetics. Diabetic Med 1987;4:441-5.

83 Cohen DL, Close CF, Viberti GC. The variabili-ty ofovemight urinary albumin excretion in insulin-dependentdiabetic and normal subjects. Diabetic Med 1987;4:437-40.

84 Feldt-Rasmussen B, Mathiesen ER. Variability of urinaryalbumin excretion in incipient diabetic nephropathy.Diabetic Nephropathy 1984;3:101-3.

85 Rowe DJF, Bagga H, Betts PB. Normal variations in rate ofalbumin excretion and albumin to creatinine ratios inovernight and daytime urine collections in non-diabeticchildren. Br MedJ 1985;291:693-4.

86 Marshal SM, Hackett A, Court S, Parkin M, AlbertiKGMM. Albumin excretion in children and adolescentswith insulin-dependent diabetes. Diabetes Res 1986;3:345-8.

87 Mathiesen ER, Saurbrey N, Hommel E, Parving H.Prevalence of microalbuminuria in children with type 1(insulin-dependent) diabetes mellitus. Diabetalogia1986;29:640-3.

88 Howey JEA, Browning MCK, Fraser CG. Selecting theoptimum specimen for assessing slight albuminuria, and astrategy for clinical investigation: Novel uses of data onbiological variation. Clin Chem 1987;33:2034-8.

89 Hutchison AS, O'Reilly D St J, MacCuish AC. Albuminexcretion rate, albumin concentration, and albumin/creatinine ratio compared for screening diabetics for slightalbuminuria. Clin Chem 1988;34:2019-21.

8



j.com/

J Clin P


nloaded from

http://jcp.bmj.com/

increased albumin excretion in diabetes

Documents