The Pathobiology of Subclinical Pyelonephritis—An Experimental Evaluation

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Kidney International, Vol. 41(1992), pp. 1356—1365

The pathobiology of subclinical pyelonephritis—An experimental evaluation

THOMAS E. MILLER, GLENNE FINDON, STEPHEN P. RAINER, and JOHN B. GAVIN

Departments of Medicine and Pathology, University of Auckland, Auckland,New Zealand

The pathobiology of subclinical pyelonephritis-'--An experimental eval- this information from clinical studies, the present investigation uation. Clinical studies have demonstrated a poor correlation between was undertaken to develop and characterize an animal model of localization tests, which are designed to determine the site of urinary subclinical pyelonephritis. tract infection, and symptoms of upper urinary tract infection. One explanation is that microorganisms may be present in the kidney but not initiate an inflammatory response with associated symptoms. An animal Methods model has been developed to obtain quantitative information on the comparative pathobiology of lesion-inducing and non-lesion-inducing Animals infections. In this model, non-manipulated kidneys had acquired a persistent microbial flora within 48 hours of the lower urinary tract Adult male and female Hooded Oxford (HO) rats, three to becoming infected. This bacterial invasion was not associated withfour months old and weighing between 200 and 250 grams were gross or histologic changes within the renal parenchyma, but minor foci used in these experiments. of inflammatory cells were seen beneath the epithelium lining the calyces. Ureteric urines from such kidneys contained many leukocytes Bacterial strain and high numbers of bacteria. These results showed that the kidneys were infected, rather than colonized, and the term 'subclinical" The strain of Escherichia coil 075 employed in these exper- infection was appropriate. Antimicrobial agents were variably success- iments was the same as that used in previous studies of ful at eradicating the infection. These experimental observations support the concept of subclinical pyelonephritis and may explain theexperimentally induced urinary tract infection [6, 7]. absence of symptoms in the clinically equivalent situation in humans. Bacteriologic analysis Acute pyelonephritis was induced in the left kidney of 70 Studies of patients with significant bacteriuria, who haveanesthetized Hooded Oxford (HO) rats, by the direct innocu- undergone ureteric catheterization or bladder washout testslation of l0 E. co/i 075 into the surgically exposed kidney using a glass microcapillary pipette. The inoculum was delivered in 5 have shown that in 30 to 60% of cases, bacteriuria originates in the upper urinary tract [1—5]. The same studies have also1.d amounts into two sites in the poles of the kidney [8]. Because of the absence of vesico-ureteric valves in the rat, reflux of demonstrated a poor correlation between upper urinary tractinfected urine occurs, and, in this strain, leads to bacterial localization and appropriate clinical symptoms. The suggestionlocalization in the contralateral kidney. The animals were has therefore been made that clinical manifestations of acutehoused under standardized conditions without any additional pyelonephritis are determined by factors other than bacterialmanipulations. Autopsies were carried out on seven groups of invasion of the kidney per se. For example, renal tissue may10 animals at regular intervals over the following 60 days. often be colonized rather than infected and invading microor-Microbial numbers were then quantified in the directly chal- ganisms may fail to stimulate the inflammatory response asso-lenged kidneys with classic pyelonephritis and in the non- ciated with symptomatic infections, Clinically it is important to manipulated (right) kidneys. The kidneys were removed asep- differentiate between individuals with asymptomatic bacterialtically, decapsulated and weighed, then homogenized in 9 ml of invasion of the upper urinary tract and those with cystitis. The normal saline using a rotating PTFE pestle in a heavy-walled clinical and laboratory findings are similar in these two groups, but in cystitis, the infectious process is limited to the lowerglass tube. Serial tenfold dilutions of the homogenate were then urinary tract. Although the term "subclinical pyelonephritis"made in normal saline and Columbia agar (Gibco Laboratories, adequately describes the symptomless bacterial invasion of theLife Technology Inc., Madison, Wisconsin, USA) pour plates renal parenchyma, there are insufficient details of its naturalwere prepared to determine the bacterial numbers per gram of history available for it to be accepted as a clinical entity.renal tissue, Because of the considerable difficulties involved in obtaining Comparative pathology Pyelonephritis was induced in the left kidney of 66 HO rats Received for publication August 12, 1991 which were divided into three groups and sacrificed 7, 30 or 60 and in revised form November 6, 1991 days later. At autopsy both directly challenged and non-manip- Accepted for publication November 8, 1991 ulated kidneys were examined for gross and histopathologic © 1992 by the International Society of Nephrology changes. The gross pathology of each kidney was quantitated

1356 Miller Ct a!: Experimental subclinical pyelonephritis 1357

Table1. Dosages and pharmacologic parameters of antimicrobial agents

MIC Peak concentration Antibiotic p.g/ml Dose Route Frequency Serum Urine Tissue Ampicillin 1.6 50 mg/kg i.m. 2 x daily 25 g/ml 400 pg/mI 40 pg/g (Ampicyn) Gentamicin 1.6 7.5 mg/kg i.m. 2x daily 8 pg/mla 40 pg/ml 45 pg/g (Garamicin) Cotrimoxazole 0.25 50 mg/kg p.o. 2 x daily Trim 3 pg/mI Trim 300 pg/mI Trim 22 pg/g (Bactrim) Sulpha 4 pg/mI" Suipha 20 pg/ml Sulpha 2.4 pg/g Ceftriaxone 0.1 200 mg/kg i.m. Daily 160 pg/mlC NA NA (Rocephin) Ciprofloxacin 0.02 20 mg/kg p.o. 2x daily 2.5 pg/mI" 345 pg/mI 4 pg/g (Ciproxin) Aztreonam 0.1 20mg/kg i.m. 2x daily 38 pg/mla 632 pg/mI 190 sg/g (Azactam) Trimethoprim NA 10 mg/kg p.o. 2x daily 3 pg/mlb 300 pg/mI 22 pgIg (Triprim) NA is not available. aAuthor'sunpublished data bReeves[lii Glauser and Bonard [12] d Campoli-Richardset al [13]

using a computer program, written in BASIC and installed on a National Panasonic RL-H7000W PC. The program produces two mirror image kidney profiles on the screen, representing both sides of one kidney. The dimensions of the profiles were scaled to those of the normal rat kidney and a grid was superimposed onto each kidney to aid orientation. Using a computer mouse, the scarred areas on the excised kidneys were shaded in on the screen. The software then calculated the size of the shaded (scarred) areas on both screen profiles and expressed this as a percentage of the total kidney surface area. Kidneys for light microscopic examination were placed in 0.1% cetylpyridinium chloride in 10% formaldehyde. After routine processing, 7 to 10 p. thick sections were prepared and stained with hematoxylin and eosin. The lesion size was determined using a paper profile of each kidney section on which was recorded the area of tissue showing histopathologic evidence of damage. The mean proportion of section occupied by lesions was calculated in eight representative sections by computer- graphic analysis as detailed above. Microbial distribution Twenty-four animals were challenged in the left kidney as above, and six animals each sacrificed 7, 14, 30 and 60 days later. Each kidney was removed and placed in a petri dish, on a thin bed of solidified wax (Paraplast, Sherwood Medical Fig. 1. Diagrammatic representation of the procedure used for kidney Industries, St. Louis, Missoun, USA). Molten wax was thendissection. Each kidney yielded approximately 40 pieces, each weigh- poured over the entire kidney and allowed to solidify. A singleing on average, 25 mg. edged razor blade was used to make five cuts along the length of the kidney to form six slices. Each section was then removed from the wax bed and laid on the flat surface of the wax. The slices of tissue were further dissected, each cut from top totube. The homogenate was then poured onto a Columbia agar bottom into portions weighing between 10 and 40 mg, so that in pour plate mixed with molten agar, and the total number of total, each kidney was divided into 30 to 40 pieces (Fig. 1).bacteria determined in each portion. This was expressed as a Segment numbers were then recorded on a paper profile and percentage of the total number of bacteria in the whole kidney individual portions weighed and homogenized in I ml of normal and marked on a paper profile. For graphic presentation, 2% of saline using a flat bottomed glass rod in a flat bottomed glassthe total bacteria was represented as a single dot. 1358 Miller et a!: Experimental subclinical pyelonephritis

A Classical pyelonephritis 8— S. S.. 7- S S..S • 6— ::: S.S S. S. . •5•S 5 S. S...... •.. •.S. 4 S. • S...

a) (a U) 2 >, a) C <1 ______I ____I I I I I I 9- 0 1 3 7 14 21 30 60 E Time,days after challenge 0) 0.a)

a) C-) BRenal colonization

a) -U 7 E C 6 S 0 -J .•. • 5 S. S. •SS . S.. • . • .. . .f...... S 4 •. S. •• •55• 5 . 3 S

2 Fig. 2. Bacteriologic status, over a 60 day period, of the left and right kidneys in <1 •• experimentally induced pyelonephritis.(A) I I I I I I IInfectionwas established in the left kidney 3 7 14 21 30 60 using a direct challenge with E. coli 075 (B) Right kidneys were not manipulated at any Time, days after challenge stage,

Histopathologic and electron microscopic evaluation of data was used to identify foci of bacterial sequestration and the bacterial foci in non-manipulated kidneys adjacent tissue samples were processed and examined for Using the dissection procedure described above for bacterialhistopathologic changes, using light and electron microscopy. distribution, focal areas of renal tissue from 16 non-manipulatedFifty-two tissue samples were examined by light microscopy at kidneys, which had been infected for 7, 14, 30 and 60 days, werefive different levels and compared with an equivalent number of examined histopathologically. Alternate samples of tissue weresamples from uninfected kidneys. For electron microscopic taken for bacteriologic analysis and the adjacent pieces placedexamination of bacterial foci in the non-manipulated kidneys, in fixative solution for histologic examination. The bacteriologicsix animals were challenged in the left kidney and left for 60 Miller et a!: Experimental subclinical pyelonephritis 1359

Bacteriology A Gross pathology 15-i La >6 •,... . a) C) 6 . Co La C 10 ii La £ . .. C La aa >- a . 5 Laa a Lia E . C 5 La La S. ••. Li 4 LJL B Histopathology 15 >a) <1 t!.!!j LJ L.J 10 aaa 7 30 60 a La a at a aaaiii Lii Cytology I ci ' La Lii >200 .• ••• Co a 200 0 L...... J S S Subclin. Classic Subclin. Classic Subclin. Classic infect. 150 . infect.pyelo. infect. pyelo. pyelo. C.) a) . 30 60 : 0 1oo • Time, days post-challenge Fig.4. Quantitative analysis of gross and histopathologic changes in 50 :: • directly challenged kidneys (left) with classic pyelonephritis and unmanipulated right kidneys with subclinical infection. 0 L.!.!..J 7 30 60 Time, days after challenge Fig.3. Bacterial and leukocyte numbers in ureteric urine from unmanipulated right kidneys colonized with E. coli. Samples were obtained 7, 30and 60 days after challenge. Antimicrobial therapy The minimum inhibitory concentrations of each antimicrobial to the strain of E. co/i 075 used to establish infection were determined before use. A standard bacterial inoculum was added to a series of tubes containing a decreasing amount of the days without any further manipulation. A bacterial focus wasagent in broth and the effect on the bacterial growth observed. identified as above in the contralateral kidneys, adjacent areasAntibiotic sensitivity, using the disk method of testing, was also carried out and the size of the zone of inhibition determined for fixed in 2.5% glutaraldehyde in 0.1 M buffer at pH 7.4 phosphate each antibiotic. A dose was then chosen for each antibiotic that and processed as previously described [91.The70 to 90 nanom-gave blood, urine and tissue levels of antibiotic in experimental eter thick sections were examined in a Philips 4IOLS transmis-animals comparable to those found clinically. The source of this sion electron microscope. information and the doses used experimentally are shown in Table 1 [11—131. The left kidneys of eight groups often animals were challenged as described above. Seven of the groups were Bacteriologic and cytologic content of ureteric urine treated for ten days with the appropriate antibiotics commenc- ing 30 days after challenge. Ureteric catheters were inserted in E. coli renal infection was established in the left kidney of 36 the right ureters, two days after the final dose of the antibiotic HO rats, which were then divided into three equal groups. No and samples collected over a period of several hours. Bacteri- further manipulations were carried out until cannulae wereologic examination of the urine was carried out as described placed in ureters draining the right kidney 7, 30 and 60 dayspreviously. The animals were then sacrificed and the unmanip- later. Ureteric catheters were also placed in 25 normal animals.ulated kidneys removed for bacteriologic evaluation as detailed Ureteric urine was collected from animals as previously de-above. Residual inhibition was occasionally noted in animals scribed [10]. Urine was collected (1 to 2 ml) into tubes packedtreated with gentamicin so homogenates of tissue from these in ice from anesthetized animals over a period of four to sixanimals were washed twice with saline before being subjected hours. Samples were examined promptly at the completion of to bacteriologic analysis. the experiment. Quantitative bacteriologic analysis was carried out using serial dilutions of the sample spread on MacConkey Statistical analysis agar plates (Difco Laboratories, Detroit, Michigan, USA) containing crystal violet and 1% lactose. Total leukocyte numbers Data were analyzed using Wilcoxon's sum of ranks method were determined in a wet preparation using a hemocytometer.for non-parametric data. 1360 Milleret al: Experimental subclinical pyelonephritis

AClassical pyelonephritis

Fig. 5. Bacterial distribution, over a 60 day period, in kidneys with (A)classicpyelonephritis and (U) subclinical pyelonephritis. The area of lesion formation is shown on kidneys that were directly challenged. Each dot represents 2% of the total bacterial isolated from the kidney (represented as log10 numbers in triangles).

Results constant level, but in a few kidneys, spontaneous clearance Bacteriologic status of directly challenged and non- occurred (Fig. 2). manipulated kidneys Bacteriologic and cytologic content of ureteric urine Bacterial numbers in the directly challenged (left) kidneys, reached a peak at 24 hours and thereafter declined slowly over Seven days after challenge of the left kidney, high bacterial the following week to reach a stable population of lO to lOcounts were found in the right ureteric urine from 8 of 12 bacteria per gram of tissue. Contralateral kidneys were some-animals examined (Fig. 3). Leukocytes were present in 11/12 of times colonized within 24 hours, but by three days, microbialthe ureteric samples. Ureteric urine draining right kidneys, invasion had occurred. Bacterial numbers were maintained at aexamined 30 and 60 days later, showed a similar incidence of Miller et a!: Experimental subclinical pyelonephritis 1361

B Subclinical pyelonephritis

infection to that found at seven days, although absolute bacte-filled with leukocytes and necrotic debris. After three to four rial numbers were reduced. In contrast there was a substantiallyweeks, the lesions showed an infiltrate of chronic inflammatory higher number of leukocytes in the ureteric urine after 30 and 60cells, tubular destruction and glomerular crowding. No surface days, than at seven days. Ureteric samples from all untreatedlesions were apparent on the contralateral unmanipulated kid- control animals were sterile and free of leukocytes. neys and the renal parenchyma was histologically normal in most cases (33 of 36 kidneys). In three kidneys, evidence of an Comparative pathology inflammatory response was seen as an intraluminal aggregation Kidneys which had been directly challenged (classic pyelo-of neutrophils in a single tubule. However, minor changes were nephritis) all showed characteristic cortical lesions on grosscommonly observed in the mucosa lining the calyx. Localized examination (Fig. 4). These were seen to be wedge-shapedinflammatory foci containing neutrophils and lymphocytes were areas which traversed the cortex, medulla and renal pelvis.present within the transitional cell epithelium and also in the Early lesions consisted of interstitial foci of neutrophils withsubepithelial tissues. These lesions were first identified after some macrophages and the tubules and collecting ducts wereseven days and showed a progression over 60 days towards 1362 Milleret a!: Experimental subclinicalpyelonephritis multiple small subepithelial collections of lymphocytes, approx-[1—51. The reasons why a "level" diagnosis based on the imately 1 mm in diameter. bacteriologic and cytologic analysis of ureteric samples might not coincide with the clinical presentation have never been Bacteriallocalizationinclassic and subclinical pyelonephritis clear, although some studies suggest that autonomic neuropa- Microorganisms were usually found in focal aggregates in thethy may prevent the expression of renal symptoms [221. Our directly challenged kidneys with classic pyelonephritis. Thebelief, based on the results of the present study, is that foci were mainly in the area of the gross lesion, up to 30 daysmicrobial invasion of the upper urinary tract sometimes occurs after challenge but thereafter they were more evenly distributedwithout the development of classic pyelonephritis. In this (Fig. 5A). In the unmanipulated kidneys, microorganisms weresituation, microorganisms establish a presence in the kidney but commonly spread evenly throughout the kidney, but focalfail to stimulate an inflammatory response of sufficient intensity aggregates were occasionally observed (Fig. 5B). to generate clinical symptoms. Once invaded, such kidneys maintain a substantial microbial population for many weeks and Histopathologic evaluation of bacterial foci in subclinical continue to excrete leukocytes in the ureteric urine. However pyelonephritis this invasion is not associated with the gross changes normally Areas of the renal parenchyma containing bacterial focifound in experimentally induced renal infection [8, 231, and appeared normal when examined by light microscopy. How-histologic evidence of bacterial invasion is limited to micro- ever discrete focal areas of the renal calyx showed evidence ofscopic foci of acute inflammation in the epithelial lining of the infiltration by neutrophils, lymphocytes and plasma cells be-calyx. In the absence of morphologic evidence of infection, neath the transitional epithelium. Electron microscopy of cor-these unmanipulated kidneys were initially considered to be tical and medullary regions within bacterial foci showed mild"colonized", rather than infected. However the consistent vascular congestion, occasional neutrophilic leukocytes and ademonstration of leukocytes in ureteric urine draining the small increase in mononuclear cells, predominantly lympho-non-manipulated kidneys suggested that microbial invasion of cytes. Occasionally there was evidence of parenchymal cellthe renal parenchyma was viewed by the host as an infectious injury in the form of cytoplasmic vacuolation, mitochondrialchallenge rather than bacterial colonization. The term subclin- inclusions and mitochondrial swelling. Fibrin and calcificationical pyelonephritis is therefore an appropriate description of an within some renal tubules was also present. Samples taken fromevent involving a cellular response by the host to microbial similar areas of the kidney in uninfected animals, showed noinvasion of the kidney that does not induce lesions typically evidence of these changes. associated with symptomatic infection. A number of investigators, dating back to Box in 1908 [24] Antimicrobial therapy and Thomsen-Walker in 1926 [25] have noted the discrepancies Bacterial numbers in the right kidneys of animals withbetween the symptoms of urinary tract infection and bacterial subclinical pyelonephritis treated with ampicillin, cotrimox-localization within the urinary tract. They concluded that azole, ciprofloxacin, ceftriaxone and gentamicin were signifi-factors other than the site of infection determine symptoms. cantly reduced compared to untreated animals. However onlyOur hypothesis was that symptoms of upper UTI are related to ceftriaxone and ampicillin were able to sterilize 50% or more ofpathologic changes resulting from the host response to micro- the kidneys. Trimethoprim and aztreonam had no effect. Ure-bial invasion of the kidney and not to upper tract localization teric urine was effectively sterilized by ampicillin and ceftriax-per Se. Experimental studies [26], and a clinical report which one. Ciprofloxacin, gentamicin and cotrimoxazole were suc-demonstrated a close association between significant bacteri- cessful in 60 to 70% of cases, but bacterial numbers were noturia, upper UT! symptoms and laboratory evidence of lesion affecte4 by trimethoprim and aztreonam (Fig. 6). No differencesformation, support this belief [27]. There is also an increasing in antimicrobial sensitivity were demonstrated when isolatesawareness that the full spectrum of the interaction between host obtained from animals, which failed to respond to antimicrobialand pathogen in renal infection has yet to be described. Com- therapy, were compared with the parent strain. ments by Komaroff [281 and Stamm et al [29], drawing attention to a group of patients who present with symptoms of acute Discussion cystitis, but in reality have early 'occult" renal infection, are Problems in correlating the bacteriologic status of the in-consistent with the view that bacterial invasion of the kidney fected kidney with functional and pathologic changes has madedoes not necessarily lead to clinical pyelonephritis. This con- the natural history of renal infection in man a difficult topic tocept is in keeping with the present finding that similar numbers study. Nonetheless, recent reports, which have taken intoof microorganisms in the directly challenged and unmanipulated account revised criteria for the bacteriologic, histopathologiccontralateral kidneys are associated with widely differing de- and anatomic diagnosis of urinary tract infection (UT!) [14, 15],grees of tissue damage. the contribution of reflux nephropathy to renal scar formation Recent experiments designed to provide a molecular expla- [16] and long term outcome for individuals with recurrent UT!nation for differences in the ability of microorganisms to invade [17—21], suggest that a pattern is emerging. One consistentrenal tissue have indicated that the DNA sequences encoding observation to come from the studies of patients with provenfor the gal-gal pili are critical determinants and that bacteria UTI has been that bacteriuria could be localized to the upperwhich express these genes persist longer in the kidney than urinary tract in over half the cases. However, there was a poorisogenic strains which lack the adhesins. For renal invasion to correlation between bacterial invasion of the upper urinary tractoccur, there is also a requirement for virulence factors [30]. The and the presence of corresponding symptoms of pyelonephritiscurrent study showed that, even in the same animal, a single Miller et a!: Experimental subclinical pyelonephritis 1363

A Renal tissue 6- I • I . In I .2 5—1 • I C . S. S • E 4 •. •S• S I . .. . . I S S •5 3- . I S •5S S o S S C 2— • . I S.. o I -J SS• S <1 -1 S... S. S 5555. 5•• I I I I I I I I I I _____ Non-Rx Amp Cotrim Trimeth Aztreo Cipro Ceftriaxone Gent

B Ureteric urine 7— S

Q) 6 S .. S . •5 5% S S S . •5 S. •5 .0 4 55 5 S o • 5 5. S. C •5 5 o0) 5 -J 3 5 . 55555 <2 %%f.% •...,. .%?...... •.•..• I I I J I I I I I .1 I _I Non-Rx Amp Cotrim Trimeth Aztreo Cipro Ceftriaxone Gent Fig. 6. Bacterial numbers in (A) renal tissue and (B) ureteric urine, from animals with subclinical pyelonephritis treated for 10 days with various chemotherapeuticagents. Treatment commenced 30 days after the establishment of infection.

microbial strain induced lesions characteristic of classic pyelo-was involved [29]. The importance of successful management of nephritis in one kidney but minimal changes, consistent withsubclinical pyelonephritis has been emphasized by Clark et at subclinical infection, in the other. These observations, together[321, who showed that infection-associated abnormalities of with those of Lomberg et at, who found similar characteristicsrenal function are reversible if the microorganisms are eradi- among strains of E. coli causing pyelonephritis in children withcated. Failure to recognize persistent subclinical infection could and without renal scarring [31], suggest that host factors as wellexplain why the management of cystitis has a 10 to 15% failure as microbial characteristics, determine the outcome of bacterialrate and a high incidence of recurrent infection [33, 34]. Thus invasion of the kidney. both clinical and experimental studies indicate that an effective Recognition that the kidney can be subclinically infected hasantimicrobial regimen for subclinical pyelonephritis has yet to important implications for antimicrobial therapy. A recentbe established and reinforces the need for aggressive therapy review of the management of acute bacterial cystitis, with andonce bacterial invasion of the kidney is suspected. without evidence of upper urinary tract involvement, indicated The aim of the present study was to provide experimental that both ampicillin and co-trimoxazole can provide effectiveevidence for the concept of subclinical pyelonephritis. Cur- therapy when infection is confined to the lower urinary tract. Inrently this diagnosis is made on the basis of clinical evaluation contrast, treatment is ineffective when the upper urinary tractand localization tests, but there is no practical way of reaching 1364 Miller et at: Experimental subclinical pyelonephritis such a diagnosis at the initial visit. If subclinical pyelonephritis by QUINN EL, KASS EH, Boston, Little, Brown and Co., 1960, is to be diagnosed promptly, non-invasive diagnostic proce- p 345 dures will need to be developed and shown to be reliable. The19. GOWER PE: A prospective study of patients with radiological pyelonephritis, papillary necrosis and obstructive atrophy. Quart J possibilities for this include quantification of urinary cytokines Med 45:315—349, 1976 [35—371,renalenzymes [38], urinary immunoglobulin levels20. PARKER J, KUNIN C: Pyelonephritis in young women. A 10- to [39], renal antigens [40, 411, and concentrating capacity [42]. 20-year follow-up. J Am Med Assoc 224:585—590,1973 The availability of the model of subclinical pyelonephritis 21.BULLENM, KINCAID-SMITH P: Asymptomatic pregnancy bacteri- characterized in the present study should prove useful in uria—A follow-up study 4—7 years after delivery, in Renal Infection developing more effective means for its diagnosis and treat- and Renal Scarring, edited by KINCAID-SMITH P, FAIRLEY KF, ment. Melbourne, Mercedes Publishing Service, 1971, p 33 22.MEYRIER A, CONDAMIN M-C, FERNET M, LABIINE-ROUSSEL A, SIMON P, CALLARD P, RAINFRAY M, SOILLEUX M, GROC A: Acknowledgments Frequency of development of early cortical scarring in acute The assistance of Linda Graham in the preparation and examination primary pyelonephritis. Kidney lot 35:696—703, 1989 of the tissue by electron microscopy is gratefully acknowledged. 23. LECAMWASAM JP, MILLER TE: Antimicrobial agents in the management of urinary tract infection: An experimental evaluation. J Reprint requests to Dr. T.E. Miller, Department of Medicine, Auck- Lab C/in Med 114:510—519,1989 land Hospital, Park Road, Auckland, New Zealand. 24. Box CR: A lecture on certain bacterial infections in the urinary tract. Lancet i:77—79,1908 References 25.THOMSEN-WALKER J: Chronic urinary infection with B. coli com- munis. Postgrad Med J 1:41—45, 1926 1. FAIRLEY KF, GROUNDS AD, CARSON NE, LAIRD EC, GUTCH RC, 26. ORMROD D, CAWLEY S, MILLER T: Neutrophil-mediated tissue MCCALLUM PHG, LEIGHTON P, SLEEMAN RL, O'KEEFE CM: Site destruction in experimental pyelonephritis, in Host-Parasite Inter- of infection in acute urinary-tract infection in general practice. actions in Urinary Tract Infections, edited by KASS EH, SVAN- Lancet ii:615—618, 1971 BoRG-EDEN C, Chicago, University of Chicago Press, 1986, p 365 2. EYKYN S, LLOYD-DAVIES RW, SHUTI-LEWORTII KED, VINNI- 27. SANDBERG T, LIDIN-JANSON G, SVANBORG-EDEN C: Host re- COMBE J: The localization of urinary tract infection by ureteric sponse in women with symptomatic urinary tract infection. Scand catheterization. Invest Urol 9:271—275, 1972 J Infect Dis 2 1:67—73, 1989 3. BOUTROS P, MOURTADA H, RONALD AR: Urinary infection local- 28. KOMAROFF AL: Acute dysuria in women. NEngIJ Med 310:368— ization. Am J Obstet Gynecol 112:379—381, 1972 375, 1984 4. HARDING GKM, MARRIE TJ, RONALD AR, HOBAN S, MUIR P: Urinary tract infection localization in women. J Am Med Assoc 29. STAMM WE, HOOTEN TM, JOHNSON JR, JOHNSON C, STAPLETON A, ROBERTS PL, MOSELEY SL, FIHN SD: Urinary tract infections: 240:1147—1 150, 1978 5. BUSCH R, HULAND H: Correlation of symptoms and results of From pathogenesis to treatment. J Infect Dis 159:400—406, 1989 direct bacterial localization in patients with urinary tract infections. 30. O'HANLEY P, LARK D, FALKOW S. SCHOOLNIK 0: Molecular basis J Urol 132:282—285, 1984 of Escherichia colt colonization of the upper urinary tract in 6. MILLER TE, FINDON G, CAWLEY S: Cellular basis of host defense BALB/c mice. Gal-gal pili immunization prevents Escherichia coli in pyelonephritis. I. Chronic infection. Brit J Exp Pathol 67: l3—23, pyelonephritis In the BALB/c mouse model of human pyelonephri- 1986 tis. J C/in Invest 75:347—360, 1985 7. MILLER TE, FINDON G, CAWLEY 5, CLARKE I: Cellular basis of 31. LOMEERO H, HELLSTROM M, JODAL U, KAIJSER B, SVANBORG- host defence in pyelonephritis. II. Acute infection. Brit J Exp EDEN C: Characteristics of Escherichia coli causing acute pyelo- Pathol 67:191—200, 1986 nephritis in children with and without renal scarring, in Host- 8. MILLER TE, ROBINSON KB: Experimental pyelonephritis: A new Parasite Interactions in Urinary Tract Infections, edited by KASS method for inducing pyelonephritis in the rat. J Infect Dis 127:307— EH, SVANBORG EDEN C, Chicago, University of Chicago Press, 310, 1973 1986, p 181 9. ARMIGER LC, HARLAND SMITH SJ: Applications of a large-speci- 32. CLARK H, RONALD AR, CUTLER RE, TURCK M: The con-elation men resin-embedding technique in experimental cardiac pathology. between site of infection and maximal concentrating ability in Exp Pathol 32:233—239, 1987 bacteriuria. J Infect Dis 120:47—53, 1969 10. MILLER TE, FINDON G, LECAMWASAM JP, YAP P: Ureteric 33. STAMEY TA, GOVAN DE, PALMER JM: The localization and catheterization in the diagnosis of pyelonephritis—An experimental treatment of urinary tract infections: The role of bactericidal urine evaluation. Kidney mt 38:835—842, 1990 levels as opposed to serum levels. Medicine 44:1—36, 1965 11. REEVES DS: Sulphamethoxazole/trimethoprim: The first two years. 34.FANG LST, TOLKOFF-RUBIN NE, RUBIN RH: Efficacy of single- J Clin Pathol 24:430—437, 1971 dose and conventional amoxicillin therapy in urinary-tract infection 12. GLAUSER MP, BONARD M: Treatment of experimental ascending Escherichia coli pyelonephritis with ceftriaxone alone and in com- localized by the antibody-coated bacteria technic. N EngI J Med bination with gentamicin. Chemotherapy 28:410—416, 1982 298:413—416, 1978 13. CAMPOLI-RICHARDS DM, MONK JP, PRICE A, BENFIELD P, TODD 35. Di MAN P. VAN KOOTEN C, AARDEN L, ENGBERG I, LINDER H, PA, WARD A: Ciprofloxacin. A review of its antibacterial activity, SVANBORG-EDEN C: Interleukin-6 induced at mucosal surfaces by gram-negative bacterial infection, Infect I,nrnun 57:3383—3388, 1989 pharmacokinetic properties and therapeutic use. Drugs 35:373—447, 1988 36.UNDERH, ENGBERG I, VAN KOOTEN C, DE MAN P, SVANBORG- 14, STAMM WE: Quantitative urine cultures revisited. Eur J Clin EDEN C: Effects of anti-inflammatory agents on mucosal inflamma- Microbiol 3:279—281, 1984 tion induced by infection with gram-negative bacteria. Infect Im- 15. KUNIN CM: Natural history of "lower" urinary tract infections. mun 58:2056—2060, 1990 Infection 18(Suppl 2): S44—S49, 1990 37. HEDGES S, ANDERSON P. LIDIN-JANSON G, DE MAN P, SVANBORG 16. BAILEY RR: Second C. J. Hoc/son Symposium on Reflux Nephrop- C:Interleukin-6 response to deliberate colonization of the human athy. Christchurch, New Zealand, Design Printing Services, 1991 urinary tract with gram-negative bacteria. InfectImmun 59:421— 17. ALWALL N: On controversial and open questions about the course 427,1991 and complications of non-obstructive urinary tract infection in adult 38. SANDBERGT,BEEGMARKJ, HULTBERG B,JAGENBUROR,TROLL- women. Acta Med Scand 203:369—377, 1978 FORS B: Diagnostic potential of urinary enzymes and f32-microglob- 18. FREEDMAN LR: Prolonged observations of a group of patients with ulin in acute urinary tract infection. Acics Med Scand 219:489—495, acute urinary tract infections, in Biology of Pyelonephritis, edited 1986 Miller et a!: Experimental subclinical pyelonephritis 1365

39.SVANBORG-EDEN C, KULHAVY R, MARILD S, PRINCE Si, tologic dissection of the human kidney using monoclonal antibod- MESTECKY J: Urinary immunoglobulins in healthy individuals and ies. J Urol 133:502—505, 1985 children with acute pyelonephritis. Scand J Immunol 21:305—313, 41. TOLKOFF-RUBIN NE: Monoclonal antibodies in the diagnosis of 1985 renal disease: A preliminary report. Kidney mt29:142—152,1986 40. BANDERNH,CORDON-CARDO C, FINSTEAD CL, WHITMORE WF, 42. BRUMFITT W, HAMILTON-MILLER iMT: Urinary infection in the VAUGHAN ED, OETTGEN HF, MELAMED M, OLD U: Immunohis- 1990's: The state of the art. Infection 18(Suppl 2):S34—S39, 1990