J Clin Pathol 1981 ;34:695-700

Review article

Biology and pathology of urinary tract infectionsLA HANSON,* A FASTH,4 U JODAL,4 B KAIJSER,t C SVANBORG EDtN*From the Departments of Clinical Immunology,* and Clinical Bacteriology,t Institute of MedicalMicrobiology and Department ofPediatrics,: University of Goteborg, Gdteborg, Sweden

The normal urinary tract is resistant to colonisationby bacteria.' Presence of significant bacteriuria(> 105 bacteria/ml) is, however, not necessarilyindicative of an infection of the urinary tract withserious consequences to the host. In fact, urinarytract infections (UTI) should be regarded as a groupof infectious diseases of varying clinical severity andprognosis and therefore they demand different levelsof priority-that is, for treatment, investigation,including radiology and follow-up. In patients withobstructive malformation UTI is always potentiallydangerous and surgical procedures or other radicalmeasures may be required. This review concentrateson the main group of patients with UTI-that is,those without obstructions of the urine flow, orother obvious underlying defects.

In school screening programmes the prevalence ofsignificant bacteriuria is around I % in girls.2 3 4About 3% of all girls will have had at least onesymptomatic infection before the age of 10 years.2The figures for adult women are similar. In malesUTI is less common, but a high frequency is foundduring the first years of life.2There are two main problems associated with

UTI: the first is quantitative. A very large proportionof bacterial infections in our society can be ascribedto UTI. Although the majority of infections areundramatic, a large population suffers from recurrentsymptomatic UTI and this makes considerabledemands on medical resources. The second problemconcerns complications, in that some of thesepatients suffer renal parenchymal reduction resultingin renal insufficiency and hypertension.The classical radiological sign of renal damage

caused by pyelonephritis is calyceal deformity withreduction of the corresponding parenchyma.5 In theepidemiological studies of Winberg et al.2 suchchanges were found in 5 % of 440 girls followed fromtheir first symptomatic UTI. In a group of 156 boys

Accepted for publication 10 December 1980

the frequency of renal scarring was 13%. Afterfollow-up for 8-15 years most of the patients hadnormal glomerular filtration owing to compensatoryhypertrophy of the contralateral kidney.6 About10-20% had bilateral scarring, however, and thesepatients in particular run the risk of a progressiverenal deterioration or the development of hyper-tension, or both.

Localisation of site of infection

UTI may be subdivided into three major clinicalgroups: acute pyelonephritis, acute cystitis, andasymptomatic bacteriuria. Bacteriuria is defined assignificant bacteriuria found on screening where thepatient was either asymptomatic or showed suchdiscrete symptoms that she or he has not soughtmedical advice.7 In contrast symptomatic UTI isdefined on the basis of the symptoms caused and oncomplementary laboratory tests.8 As indicators of aninfection involving the kidney (acute pyelonephritis)we have used temperature > 38 5°C, C-reactiveprotein > 20 mg/l, microsedimentation rate > 25mm/h, and low renal concentrating capacity duringthe acute stage of the infection. As diagnosticcriteria for acute cystitis we have used dysuria andfrequency, temperature not exceeding 38°C, absenceof back and loin pain and normal laboratory tests.

Tests for ascertaining the level of UTI have beenmuch studied especially in view of the limitations ofclinical diagnosis. In fact, we have no single fool-proof method although ureteric catheterisation maycome close to it. However, such a procedure isimpracticable. The bladder wash-out test wasintroduced as a practical simplification and has beenwidely accepted as a reference method, thoughunsupported by any studies of reliability. We used arange of tests in girls with symptomatic UTI.8Intermittent discharge of bacteria into the renalpelvis was a frequent finding and probably accountedfor an inadequate sensitivity of the bladder wash-out

695

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

696

test. In fact, the non-specific C-reactive protein(CRP) concentration was the best indicator of renalinvolvement when compared to the overall resultsfrom the range of tests. The usefulness of the CRPtest has been confirmed by others.9 In our routinework-up of the children with symptomatic UTI wetherefore include determination of CRP concentra-tions as well as of renal concentrating capacity.

Determination of serum antibodies to the 0-

antigen of the infecting Escherichia coli has enabledthe distinction to be drawn between acute pyelone-phritis and acute cystitis.101 The large number of0-antigens with a limited number of cross-reactionsmakes this method impracticable. With a pool ofantigens for routine titrations the sensitivity of themethod decreases.8 The detection of antibody-coated bacteria has been used as an indicator ofrenal infection.12 13 In children, however, thesensitivity was found to be low.8 1415 Recently, thereliability of the antibody-coated bacteria methodfor level diagnosis in adults has also been ques-tioned.16 Finally, two chemical methods should bementioned: it has been shown that the amounts oflactic dehydrogenase isoenzyme 517 and of 1B2-micro-globulin'8 are raised in the urine of patients withacute pyelonephritis.

VIRULENCE FACTORS IN GRAM-NEGATIVEBACTERIA CAUSING UTIDifferences between bacteria populations mayexplain the variable severity of UTI in the patient.Firstly, of the E coli bacteria causing acute pyelo-nephritis 80% belonged to only eight common 0-

antigen groups. In contrast, of the E coli strainscausing asymptomatic bacteriuria, 45% had an in-complete lipopolysaccharide (endotoxin) of the cellmembrane and were spontaneously agglutinating.The frequency of spontaneously agglutinatingbacteria in the urine of asymptomatic bacteriuriapatients was even higher than that in the stool ofhealthy age-matched children.'920

Secondly, the polysaccharide capsules of E coliare important virulence factors.21 Five differentK-antigens account for about 70% of E coli causingacute pyelonephritis in children. The Kl antigen wasfound on 39% of the urinary isolates.22

Thirdly, the bacteria causing acute pyelonephritiswere more resistant to the bactericidal activity ofnormal serum than were bacteria found in asympto-matic bacteriuria patients.23 The latter strains wereeven more sensitive than the ones found in normalstool.

Fourthly, E coli causing acute pyelonephritisattached efficiently to epithelial cells from theurinary tract in contrast to E coli from asympto-matic bacteriuria patients, the majority of which

Hanson, Fasth, Jodal, Kaijser, Svanborg Eden

attached poorly or not at all.24 25The speciesand tissuespecificity observed in the attachment of E coli inUTI patients indicates that adhesion is due not to a"general stickiness" of the bacterium, but to aspecific interaction of bacterial surface ligands-"adhesins"-with receptors in the epithelial cellsurface.26 The bacterial structures involved inattachment to epithelial cells have been classifiedaccording to the ability of bacteria to bind to othercells-for example, erythrocytes.27 Two main groupsof adhesins are found on E coli causing UTI.26 Onegroup induces agglutination of erythrocytes; reversedin the presence of D-mannose. These adhesins arefound on most enterobacteria without any obviousrelation to virulence. E coli strains carrying only theseadhesins attach poorly to human uroepithelial cellsbut bind to urinary slime-that is, Tamm-Horsfallprotein.28 29 The second group of adhesins induceagglutination of human erythrocytes, but this groupis not affected by D-mannose. E coli bacteria carryingthese adhesins attach to human uroepithelial cellsand are commonly found in patients with acutepyelonephritis but rarely in normal faecal isolates.26Both types of adhesins can be of a pilus or fimbriatenature,26 30 but the existence of non-fimbriateadhesins has also been suggested.27We recently reported that glycolipids isolated

from human urinary tract epithelial cells inhibit theattachment of E coli to uroepithelial cells from thesame donor.31 Globoseries glycolipids, especiallygloboside, fulfil the criteria of a receptor.25 Thebinding reaction, however, is complex and isolatedpili only partly satisfy the receptor. The density ornature of receptors may explain why E coli bindmore efficiently to uroepithelial cells from patientswith recurrent UTI than from individuals withoutUTI.33-35These various bacterial properties are commonly

found in E coli causing acute pyelonephritis, but lessoften in asymptomatic bacteriuria strains; they arelikely to be virulence factors, necessary for thebacteria to manifest the infection. In contrast,Proteus mirabilis rarely causes UTI in persons withnormal urinary tracts. P mirabilis differs from E coliin bacterial virulence factors. No polysaccharidecapsule has been described and P mirabilis alsodiffer from E coli in their capacity to attach tohuman uroepithelial cells.36 From patients withUTI, P mirabilis attached to the same extent asP mirabilis from other sources such as the stools ofhealthy carriers or blood of cases with septicaemia.The P mirabilis strains did not attach to transitionalepithelium from the bladder but to the squamousepithelium from the outer genital region.36 In addi-tion the attachment was more striking in the mid-menstrual period. The inability of P mirabilis to

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

Biology andpathology of urinary tract infections

attach to the bladder epithelium may explain itslower virulence compared to E coli.

HOST RESPONSE IN UTIAs in most infections the host defence is activatedin response to an attack of UTI. During acutepyelonephritis there is a systemic antibody response.Antibodies against the 0-antigen8 10 and occasionallythe K-antigen2' of the infecting strain have beenfound and recently antibodies to type 1 fimbriaewere described.37 A local antibody response con-sisting of IgG and secretory IgA antibodies can bedetected in the urine.38 Antibodies to fimbriae havenot so far been found in the urine. Antibodiesagainst several bacterial structures-for example,the 0-antigen or K-antigen, can protect againsthaematogenous or ascending pyelonephritis inexperimental animals.39 40 The antibodies may be ofvalue for the patient either to limit the damageincurred in the course of an infection or to preventcolonisation preceding recurrence. Antibodies incombination with the wash-out effect of the urinemay prevent colonisation of the urinary tract.

In patients with many recurrent infections it maybe possible to induce a protective immune responseby vaccination, as described in animals.39 40 Immu-nisation could be especially useful for the group whorisk renal damage. Such a vaccine should pre-sumably contain a broad selection of antigensincluding the fimbriae and polysaccharide capsuleantigens most commonly found among bacteriacausing UTI.

RISK GROUP OF UTI PATIENTSIn our prospective follow-up studies few patientshave developed renal parenchymal reduction,possibly because of close supervision and immediatetreatment of each symptomatic infection.2 Renaldamage that does occur is related to febrile UTIamong those patients4l but in the individual caseprediction is often difficult. The early attacks of UTIare more often symptomatic than the recurrenceswhich are "asymptomatic" in about 60 %.42 Noclinical or laboratory criteria have so far been usefulto predict those cases which are to develop renalscarring.Development of renal scarring is related to

vesicoureteric reflux shown by voiding cystourethro-graphy. The risk has been found especially high withdilatation of the refluxing renal unit.43 The matter iscomplex, however, and in our material scarring hasoccurred with low-grade reflux and even with nodemonstrable reflux. In addition, reflux has a con-siderable tendency to disappear during follow-up.44

In identifying patients at risk, some observationson antibodies against the Tamm-Horsfall urinary

glycoprotein are of considerable interest. Patientswith acute bacterial pyelonephritis, but not thosewith acute cystitis, had increasing IgG and IgAantibody concentrations against Tamm-Horsfallprotein during the course of the infection, reachinglevels well above healthy controls.45 The most markedincrease and the highest concentrations were inserum from patients with vesicoureteric reflux.Interestingly, the patients with renal damage andacute pyelonephritis had a poor response, withalmost no antibody increase in those with the mostsevere renal damage (as judged from serum creatin-ine and glomerular filtration rate).46 47 After theinfection, the IgG antibody concentrations of thepatients with scarred kidneys fell below the range ofthe controls and compared to pyelonephritis patientswith normal urography, there was no overlappingbetween the IgG antibody values. The measurementof IgG antibodies against Tamm-Horsfall proteincould become a risk group test for the early detec-tion of renal damage.The rise of autoantibodies against the Tamm-

Horsfall protein occurs in almost every case of acutepyelonephritis caused by various E coli and Proteusstrains. This observation together with the discoverythat all above the age of eight to nine months havesuch autoantibodies,45 suggest the possibility of across-reaction between the Tamm-Horsfall proteinand Gram-negative bacteria. In fact we found such across-reaction between the Tamm-Horsfall proteinand a protein from Gram-negative bacteria.48Antibody production might be induced by thecross-reacting antigens of colonising intestinalGram-negative bacteria. This may account for thepresence of autoantibodies against the Tamm-Horsfall protein found in all humans. The boost ofthis antibody response occurring during acutepyelonephritis may be due to the increased exposureof the Tamm-Horsfall protein to the immunesystem during the parenchymal infection, or to anincrease in cross-reacting bacterial antigen.

It is tempting to suggest that the development ofrenal damage associated with bacterial pyelone-phritis is related to autoimmunity to Tamm-Horsfall protein. In some cases of tubulointerstitialdisorders, deposits of Tamm-Horsfall protein havebeen found in the interstitial tissue.4951 Theselumps were occasionally surrounded by inflam-matory cells and in an experimental model the extentof these deposits was correlated to the concentrationof antibodies against Tamm-Horsfall protein.52 Thelow antibody concentration found in patients withrenal damage might be due to absorption of theantibodies in the interstitial infiltrates, or perhaps toan immunosuppressive effect of released tubularantigens as shown in experimental systems.53

697

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

Hanson, Fasth, Jodal, Kaijser, Svanborg Eden

The Tamm-Horsfall protein might also have otherinfluences on UTI as the protein can bind E coli inthe urinary tract.29 The Tamm-Horsfall protein orurinary slime may act as a non-specific defenderagainst Gram-negative bacterial infection of thebladder and kidney by preventing attachment to theepithelial surface. To overcome this barrier bacteriamay have developed the mannose-resistant tissue-binding type of fimbriae which are associated withacute pyelonephritis strains. On the other hand,binding to Tamm-Horsfall protein may facilitatecolonisation of the urinary tract of strains lackingthe tissue-binding capacity. This may explain whyE coli strains unable to attach to the urinary tractmucosa, persist for a long time in the urine of girlswith asymptomatic bacteriuria, without underlyingdefects like residual urine.

ASYMPTOMATIC BACTERIURIAThe bacteria found in patients with asymptomaticbacteriuria lack certain properties found in E colicausing symptomatic UTI. In some patients thebacteria changed to less virulent forms during long-term follow-up without treatment.54 Providing thekidneys were normal at the first radiological inves-tigation, non-treatment of girls with asymptomaticbacteriuria did not lead to deterioration ofthe kidneysduring a three-year follow-up. The only patient whodeveloped renal scarring had an attack of acutepyelonephritis due to a new strain, distinct from theasymptomatic bacteriuria strain.55Treatment of girls with asymptomatic bacteriuria

eradicated the bacteria rapidly in 98% of the cases.In 48% asymptomatic bacteriuria recurred at leastonce, and in 22% it recurred repeatedly. In somecases elimination of the low virulence bacteriaallowed a "recurrence" caused by a more virulentstrain originating from the stools leading to an attackof acute pyelonephritis. This was seen especially ingirls with renal damage. In the untreated group about10% lost the bacteria each year. After five years theincidence of bacteriuria was similar in the treatedand non-treated girls.On the basis of our follow-up studies in children it

seems that treatment of girls with a radiologicallynormal urinary tract is often unnecessary and some-times even harmful. Criteria of bacterial virulencesuch as the capacity to attach to uroepithelial cells orpresence of capsular substance especially E coli Klmay prove valuable in deciding whether to treat ornot. It must be remembered, however, that patientsat risk, such as those with established renal scarring,may be "compromised hosts" with an increasedsensitivity to bacteria carrying few virulence factors.In such patients all episodes of bacteriuria should betreated.

Conclusions

Combinations of several virulence factors are foundon most E coli bacteria causing acute pyelonephritis,but rarely on bacteria isolated from the normalintestinal flora. Examples of virulence factors are:capacity to attach to uroepithelial cells, presence ofendotoxin, and presence of polysaccharide capsules.The host defence system against these bacterialcomponents protects against pyelonephritis inexperimental models. Induction of protective immun-ity through vaccination may also be possible in thehuman and be of special value to the group at riskwho are susceptible to renal scarring after kidneyinfections.Symptomatic UTI in children may be separated

into acute pyelonephritis and acute cystitis. Evalu-ation of the clinical symptoms together with deter-mination of the non-specific C-reactive proteinconcentration and renal concentrating capacity are,in our opinion, the best indicators of renal infection.The risk of renal scarring is related to the number ofattacks of acute pyelonephritis. Analysis of anti-bodies against Tamm-Horsfall protein is a usefulindicator of the risk of renal scarring, since highantibody concentrations during infection are asso-ciated with vesicoureteric reflux and low antibodyconcentrations and a poor antibody response duringan attack of acute pyelonephritis are associated withrenal damage. In girls with significant bacteriuriawithout symptoms, diagnosis of severity is difficult.Asymptomatic bacteriuria may, however, be leftuntreated in patients with radiologically normalurinary tracts without an increased risk of renalscarring.

These studies were supported by grants from theSwedish Medical Research Council, Board forTechnical Development (project No. EKB-U-614),the Ellen, Walter and Lennart Hesselman Founda-tion for Scientific Research and Stiftung Volks-wagenwerk, W Germany. This review was writtenduring the tenure of a Fogarty Scholarship-in-Residence at the National Institute of Health, USAfor Lars Hanson between August 1979 and September1980.

References

Cox C, Hinman F. Experiments with induced bacteriuria,vesical emptying and bacterial growth on the mechanismof bladder defense to infection. J Urol 1969;86:739-48.

2 Winberg J, Anderson HJ, Bergstrom T, et al. Epidemiologyof symptomatic urinary tract infection in childhood.Acta Paediatr 1974;Suppl 252.

3 Kass EH, Zinner SH. Bacteriuria and renal disease. JInfect Dis 1969;120:27-46.

' Kunin CM, Deutscher R, Paquin A. Urinary tract infection

698

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

Biology andpathology of urinary tract infections

in school children: An epidemiologic, clinical, andlaboratory study. Medicine 1964;43:91-130.

'Hodson CJ, Wilson S. Natural history of chronic pyelone-phritic scarring. Br Med J 1965;ii:171-94.

6 Winberg J, Claesson I, Jacobsson B, Jodal U, Peterson H.Renal growth after acute pyelonephritis in childhood.An epidemiological approach. In: Hodson CJ,Kincaid-Smith P, eds. Reflux nephropathy. New York:Masson, 1979;309-22.

7 Lindberg U, Claesson 1, Hanson LA, Jodal U. Asympto-matic bacteriuria in schoolgirls. I. Clinical and labora-tory finding. Acta Paediatr Scand 1975 ;64:425-31.

8 Jodal U, Lindberg U, Lincoln K. Level diagnosis ofsymptomatic urinary tract infections in childhood. ActaPaediatr Scand 1975;64:201-8.

9 Wientzen RL, McCracken GH, Petruska ML, SwinsonSG, Kaijser B, Hanson LA. Localization and therapy ofurinary tract infections of childhood. Pediatrics 1979;63:467-74.

0 Winberg J, Anderson HJ, Hanson LA, Lincoln K. Studiesof urinary tract infections in infancy and childhood. I.

Antibody response in different types of urinary tractinfections caused by coliform bacteria. Br Med J 1963;ii :524-7.

Vosti KL, Monto AS, Rantz LA. Host-parasite interactionin patients with infections due to Escherichia coli. II.

Serologic response of the host. JLab Clin Med 1965;66:613-26.

2 Jones SR, Smith JW, Sanford JP. Localization of urinary-tract infections by detection of antibody-coated bacteriain urine sediment. N Engl J Med 1974 ;290:591-3.

12 Thomas V, Shelokov A, Forland M. Antibody-coatedbacteria in the urine and the site of urinary-tract infec-tion. N Engl J Med 1974;290:588-91.

14 Forsum U, Hjelm E, Jonsell G. Antibody-coated bacteriain the urine of children with urinary tract infections.Acta Paediatr Scand 1976;65:639-42.

s5 Hellerstein S, Kennedy E, Nussbaum L, Rice K. Locali-zation of the site of urinary tract infections by means ofantibody-coated bacteria in the urinary sediments. JPediatr 1978;92:188-93.

6 Mundt KA, Polk BF. Identification of site of urinary-tractinfections by antibody-coated bacteria assay. Lancet1979;ii:1 172-5.

17 Carvajal HF, Passey RB, Berger M, Travis LB, LorentzWB. Urinary lactic dehydrogenase isoenzyme 5 in thedifferential diagnosis of kidney and bladder infections.Kidney Int 1975;8:176-84.

18 Schardijn G, Statius van Eps LW, Swaak AJG. Urinarybeta2-microglobulin in upper and lower urinary-tractinfections. Lancet 1979;i:805-7.

1 Lindberg U, Jodal U, Hanson LA, Kaijser B. Asympto-matic bacteriuria in schoolgirls. IV. Difficulties of leveldiagnosis and the possible relation to the character ofinfecting bacteria. Acta Paediatr Scand 1975;64:574-80.

20 Lidin-Janson G, Lindberg U. Asymptomatic bacteriuria inschoolgirls. VI. The correlation between urinary andfaecal Escherichia coli. Relation to the duration of thebacteriuria and the sampling technique. Acta PaediatrScand 1977;66:349-54.

21 Kaijser B, Vahlne G. Escherichia coli K antigen. Zbl BaktHyg I Abt Orig A 1979;243:271-88.

Kaijser B, Hanson LA, Jodal U, Lidin-Janson G, RobbinsJB. Frequency of E coli K antigens in urinary tractinfections in children. Lancet 1977 ;i:663-4.

22 Lidin-Janson G, Hanson LA, Kaijser B, Lincoln K,Lindberg U, Olling S, Wedel H. Comparison of Escheri-chia coli from bacteriuric patients with those from fecesof healthy schoolchildren. J Infect Dis 1977;136:346-53.

24 Svanborg Ed6n C, Hanson LA, Jodal U, Lindberg U,

Sohl Akerlund A. Variable adhesion to normal humanurinary tract epithelial cells of Escherichia coli strainsassociated with various forms of urinary tract infections.Lancet 1977;ii:790-2.

215 Svanborg Eden C, Eriksson B, Hanson LA, Jodal U,Kaijser B, Lidin-Janson G, Lindberg U, Olling S.Adhesion to normal human uro-epithelial cells ofEscherichia coli from children with various forms ofurinary tract infections. J Pediatr 1978;93:398-403.

26 Svanborg Ed6n C, Hagberg L, Hanson LA, Korhonen T,Leffler H, Olling S. Adhesion of Escherichia coli inurinary tract infection. In: Adhesion and MicroorganismPathogenicity. (Ciba Found Symp 80) London: PitmanMedical (in press).

27 Duguid JP. The function of bacterial fimbriae. ArchImmunol Ther Exp 1968;16:173-88.

28 0rskov 1, 0rskov F, Birch-Andersen A. Comparison ofEscherichia coli fimbrial antigen F7 with type 1 fimbriae.Infect Immun 1980;27:657-66.

29 Fasth A, Hanson LA, Svanborg Eden C. Tamm-Horsfallglycoprotein that associates with Escherichia coli mayinhibit attachment to human urinary tract epithelial cellsin vitro. FEMS Microbiol Lett (submitted).

30 Korhonen T, Ed6n S, Svanborg Eden C. Binding of purifiedEscherichia coli pili to human urinary tract epithelialcells. FEMS Microbiol Lett 1980;7:237-41.

31 Svanborg Eden C, Leffler H. Glycosphingolipids as possiblereceptors for Escherichia coli attaching to human urinarysediment epithelial cells. Scand J Infect Dis 1980; Suppl24:144-7.

32 Leffler H, Svanborg Eden C. Chemical identification of aglycosphingolipid receptor for Escherichia coli attachingto human urinary tract epithelial cells and agglutinatinghuman erythrocytes. FEMS Microbiol Lett 1980;8:127-34.

33 Fowler JE, Stamey TA. Studies of introital colonization inwomen with recurrent urinary infections. VII. The roleof bacterial adherence. J Urol 1977;ii:472-6.

34 Kallenius G, Winberg J. Bacterial adherence to periureteralepithelial cells in girls prone to urinary tract infections.Lancet 1978;ii:510.

35 Svanborg Eden C, Jodal U. Attachment of Escherichia colito sediment epithelial cells from UTI prone and healthychildren. Infect Immun 1979;26:837-40.

36 Svanborg Ed6n C, Larsson P, Lomberg H. Attachment ofProteus mirabilis bacteria to human urinary tractepithelial cells is different from that of Escherichia coli.Infect Immun 1980;27:804-7.

37 Silverblatt F, Rene P. Serologic response to E coli pili inpyelonephritis. Abstract ICAAC Boston, 1979.

38 Sohl Akerlund A, Ahlstedt S, Hanson LA, Jodal U.Antibody responses in urine and serum against Escheri-chia coli 0 antigen in childhood urinary tract infection.Acta Pathol Microbiol Sect C 1979;87:29-36.

89 Kaijser B, Larsson P, Olling S. Protection against ascendingE coli pyelonephritis in rats and significance of localimmunity. Infect Immun 1978 ;20:78-81.

40 Kaijser B, Larsson P, Schneerson R. Protection againstacute, ascending pyelonephritis caused by E coli in ratsusing isolated capsular antigen conjugated to bovineserum albumin (in manuscript).

41 Jodal U, Bjure J, Claesson I, Hanson LA, Jacobsson B,Lindberg U, Peterson H, Winberg J. Hlufigkeit vonRezidiven und von Nierenparenchymnarben-Verleichvon Madchen mit symptomatischer Harnwegsinfektionund asymptomatischer Bacteriurie. In: Olbing H, ed.Rezidivierende nicht-obstructive Harnwegsinfektionen beiKindern. Berlin: Springer Verlag, 1980;1 1-20.

42 Bergstrom T. A prospective study of urinary tract infectionin infancy and childhood. Thesis, Goteborg, 1972.

699

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

70043 Rolleston GL, Shannon FT, Otley WLF. Follow-up of

vesico-ureteric reflux in the newborn. Kidney Int 1975;8:559-64.

44 Edwards D, Norman ICS, Prescod N, Smellie JM.Disappearance of vesicoureteric reflux during long-termprophylaxis of urinary tract infection in children. Br MedJ 1977;ii:285-8.

46 Fasth A, Hanson LA, Jodal U, Peterson H. Autoantibodiesto Tamm-Horsfall protein associated with urinary tractinfections in girls. J Pediatrics 1979;95:54-60.

46 Fasth A, Bengtson U, Kaijser B, Wieslander J. Antibodiesto Tamm-Horsfall protein associated with kidneyscarring and urinary tract infections in adults. Kidney Int1981; (in press).

47 Fasth A, Bjure J, Hellstrom M, Jacobsson B, Jodal U.Autoantibodies to Tamm-Horsfall glycoprotein inchildren with renal damage associated with urinary tractinfections. Acta Paediatr Scand 1980 ;69 :709-15.

48 Fasth A, Ahlstedt S, Hanson LA, Jann B, Jann K, KaijserB. Cross-reactions between the Tamm-Horsfall glyco-protein and Escherichia coli. Int Archs Allergy applImmunol 1980;63:303-1 1.

49 Resnick JS, Sisson S, Vernier RL. Tamm-Horsfall protein.Abnormal localization in renal disease. Lab Invest 1978;38:550-5.

Hanson, Fasth, Jodal, Kaijser, Svanborg Eden

50 Solez K, Heptinstall RH. Intrarenal urinary extravasationwith formation of venous polyps containing Tamm-Horsfall protein. J Urol 1975;119:180-3.

51 Zager RA, Cotran RS, Hoyer JR. Pathologic localizationof Tamm-Horsfall protein in interstitial deposits in renaldisease. Lab Invest 1978;38:52-7.

52 Hoyer JR. Tubulointerstitial complex nephritis in ratsimmunized with Tamm-Horsfall protein. Kidney Int1980;17:284-92.

53 Neilson EG, Phillips SM. Cell mediated immunity ininterstitial nephritis. J Immunol 1979;123 :2373-80.

54 Lindblad BS, Ekengren K. The long term prognosis cfnon-obstructive urinary tract infection in infancy andchildhood after the advent of sulphonamides. ActaPaediatr Scand 1969;58:25-32.

55 Lindberg U, Claesson I, Hanson LA, Jodal U. Asympto-matic bacteriuria in schoolgirls. VIII. Clinical courseduring a 3-year follow-up. J Pediatr 1978;92:194-9.

Requests for reprints to: Dr LA Hanson, Department ofClinical Immunology, University of Goteborg, Guid-hedsgatan 10, S413 46 Goteborg, Sweden.

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from

doi: 10.1136/jcp.34.7.695 1981 34: 695-700J Clin Pathol

 L A Hanson, A Fasth, U Jodal, et al. urinary tract infections.Biology and pathology of

http://jcp.bmj.com/content/34/7/695.citationUpdated information and services can be found at:

These include:

serviceEmail alerting

corner of the online article.this article. Sign up in the box at the top right Receive free email alerts when new articles cite

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on July 16, 2011 - Published by jcp.bmj.comDownloaded from