8070

Saturday 29 April 1978

SEQUELÆ OF COVERT BACTERIURIA INSCHOOLGIRLS

A Four-Year Follow-Up Study

CARDIFF-OXFORD BACTERIURIA STUDY GROUP*

Royal Infirmary, Cardiff and Radcliffe Infirmary, Oxford

Summary A trial involving 208 girls aged 5-12who had bacteriuria on screening was

carried out to determine the effects of covert bacteriuriaon renal growth and scarring. 98 were observed for 4years without treatment and 110 were treated. 77% ofthe treated girls, but only 26% of the untreated controls,were free of infection for at least half of the 4-year fol-low-up. However, treatment had no effect on the emer-gence of symptoms, clearance of vesico-ureteric reflux

(V.U.R.), kidney growth, or the progression of kidneyscars. New kidney scars did not develop in previouslyunscarred kidneys. 12 (6%) of the 208 girls had pro-gression of pre-existing kidney scars; this high-riskgroup was characterised by kidney scarring on the initialX-ray, a high prevalence of V.U.R., and persistent orrecurrent bacteriuria. Screening for covert bacteriuriacannot therefore be recommended in schoolgirls sincekidney damage associated with infection arises before

the age of 5.

Introduction

ALTHOUGH it is feasible to screen schoolgirls forcovert bacteriuria,l-6 it is essential to know how harmfulthe condition is and how it responds to treatmentbeforeroutine screening is advocated. The course of covert bac-teriuria in treated girls has been described in detail’ butless is known of its natural history in untreatedchildren.8 We report on renal growth and the develop-ment of kidney scarring in schoolgirls with asympto-matic bacteriuria who were followed-up, with or withouttreatment, for 4 years, and relate the results to theamount of bacteriuria experienced during this period.

Method

Population16 800 schoolgirls, aged 5-12, in Cardiff and Oxford were

screened during 1971-72 for bacteriuria by methods pre-

*The members of the group are: Prof. A. W. ASSCHER, Dr E. W. L.FLETCHER, Dr H. H. JOHNSTON, Dr J. G. G. LEDINGHAM, Prof.M. S. F, MCLACHLAN, Dr R. T. MAYON-WHITE, Dr S. T. MELLER, DrG. SLEIGHT, Dr E. H. SMITH, Mr J. C. SMITH, Prof.M. SUSSMAN, DrE. R. VERRIER-JONES, and Dr L. A. WILLIAMS.

viously described.4 294 girls (1.8%) had bacteriuria-i.e., twoor more specimens with more than 105 viable bacteria/ml ofurine; Excretion urograms and micturating cystograms weretaken in 255 girls. 7 girls were then excluded from the studyand the other 248 girls were allocated randomly to a treatedand a control group. Follow-up was completed on 110 of 127in the treated group and on 98 of 121 in the control group. 8of the girls in the control group had urinary-tract infectionwith symptoms and required treatment. The results obtainedin this late treatment group were analysed separately but sincethere were no differences between this subgroup and the restof the control group, the findings in these 8 girls have been in-cluded with those of the rest of the control group. 10 girls werefollowed up incompletely because they left the survey areasand 30 girls refused regular supervision and/or the final radio-logical examinations:

Follow-upThe study period ran from the date of the first X-ray to the

date of the second. During this time midstream urine (M.s.u.)specimens were collected monthly in Oxford and every 2months in Cardiff. Specimens were regarded as infected if theygrew more than 105 colonies/ml in pure culture. Heavy mixedgrowths were considered contaminated and repeat specimenswere collected as soon as possible. The occasional specimenwith 104-105 colonies/ml of a single organism or with morethan 105 colonies/ml of a possible pathogen mixed with a muchsmaller number of probable contaminants was counted as posi-tive only if the suspect organism was confirmed as a pathogenby the findings in a preceding or subsequent specimen. Other-wise such doubtful specimens were excluded from analysis.

Girls in the treated group had appropriate antibacterialtherapy for their bacteriuria-usually co-trimoxazole, but alsoampicillin, nitrofurantoin, nalidixic acid, and pivmecillinam.Initially 7 or 14 day courses were given but longer courses(3-12 months) of low-dosage maintenance therapy were givenat the discretion of the physicians to girls with recurrent bac-teriuria. The control group was given no antimicrobial treat-ment by us. Girls in the late treatment group were given anti-bacterial drugs for symptomatic episodes only; otherwise theywere followed up in the same manner as were the untreated

group.

Duration of Bacteriuria

The extent to which the girls were exposed to bacteriuriaduring follow-up was measured by counting the intervals (indays) between each positive specimen and the preceding clearspecimen and dividing this total by the total length of the fol-low-up period in days, and expressed as a percentage of timeinfected during the 4-year follow-up. This method of expressingexposure to bacteriuria assumes that remissions begin immedi-

890

ately after the last uninfected specimen was obtained and thatany deviations from this assumption cancel each other out.There was a close correlation (r=0.97) between this measureand the ratio of infected to all non-infected urine specimenscollected from each child indicating that the urine collectionswere regularly spaced during the 4-year follow-up.

Radiological MethodsThe same radiological techniques9 for excretion urography

and micturating cystography were used in Oxford and Cardiff.Renal lengths were measured by radiologists who did not knowto which group the girls belonged. The renal lengths on thefirst X-rays were measured again when the second X-rays wereread to ensure that measurements were made from the same

points on the renal poles. Pyelonephritic scars were identifiedby the criteria of Hodson and Wilson.1O The incidence of newand/or increased kidney scarring was assessed by one of theradiologists (M.S.F.McL.) who viewed unnamed and undatedcopies of all initial and final films in- random order. Increasedkidney scarring was defined as further loss of renal substancecompared with pre-existing scars. Left and right kidneys wereconsidered separately in the analysis of renal growth and theirannual growth-rate was calculated by dividing the differencein renal lengths (in mm) between the two examinations by theinterval between the X-rays (in years). Kidney growth wasconsidered impaired when the annual growth-rate was morethan two standard deviations below the normal values pub-lished by Hodson et al." v.u.R. was graded as follows: grade1-reflux into ureter but not extending to kidney; grade2-reflux extending to kidney; grade 3-reflux extending tokidney and associated with calyceal dilatation.

Other Observations

Blood-pressures were measured at clinic visits with standardsphygmomanometers and appropriate cuff sizes for the smallergirls. The blood-pressure at the first visit was compared withthat at the last visit. Serum-creatinine and blood-urea concen-trations were measured by Technicon ’Auto Analyzer’ methodsN-11B and N-1C respectively on venous blood obtained at thefirst and the second urographic examinations.

Results

Clinical Observations

At the start of the trial the mean (±s.D.) ages of thegirls in the treatment and control groups were 8.7

(±1.8) and 8.6 (±1.9) years respectively. The mean ofthe follow-up period was 4.0 (±0’3) years. During fol-low-up 8 (8%) of the 98 girls in the untreated group hadan infection accompanied by frequency and dysuria orloin pain and fever. These girls were given appropriateanti-bacterial therapy. The infecting organisms andradiological findings in these girls did not differ fromthose of the control group as a whole. In the treated

group, 9 (8%) of the 110 girls had symptoms during fol-low-up. Body height increased equally in both groups(mean 21 and 22 cm in the treated and control groupsrespectively). The systolic and diastolic blood-pressuresin the treated and control groups did not differ signifi-cantly and had not changed by the end of the follow-up.None of the girls who completed the study had hyperten-sion (diastolic blood-pressure >90 mm Hg). Blood-ureaand serum-creatinine concentrations showed no signifi-cant differences either between the groups or betweenthe initial and final measurements (table i).

Bacteriological FindingsAll except 6 (3%) of the 208 initial infections were due

TABLE I-BLOOD-UREA AND SERUM-CREATININE

CONCENTRATIONS IN TREATED AND CONTROL GROUPS

to Escherichia coli. At the end of the study, 17 (15%) ofthe 110 girls in the treated group had bacteriuria com-pared with 44 (45%) of the 98 girls in the control group(20; n<0001). There was an overlap between thetreated and control groups in the extent to which bacteri-uria was present during follow-up partly because ofrecurrent infections in the treated group and partlybecause of spontaneous remissions and the effects of in-cidental antibacterial therapy for other infections in thecontrol group. Nevertheless none of the girls in thetreated group had persistent bacteriuria whereas 29

(30%) of the girls in the control group did. When all thegirls were grouped according to the proportion of timethat bacteriuria was present, it appeared that there weretwo distinct groups-those with bacteriuria less than25% of the time and those with bacteriuria for 75% ormore of the time (see accompanying fig.). For the analy-sis the intermediate group was divided into two

groups--one which had bacteriuria for 25-49% of thefollow-up period, and the other for 50-74% of the time,this division produced four groups of similar size.

Radiological FindingsRenal growth.-It was not possible to assess the

growth of 3 of the kidneys because bowel shadows ob-scured their poles. The annual growth-rates (table n)were not affected by treatment but the growth of kid-neys that were scarred and drained by ureters which

Percentage of time with bacteriuria

Duration of bacteriuria in treated and control subjects duringthe 4-year follow-up.

891

TABLE II-MEANS (&plusmn;S.D.) OR RENAL GROWTH-RATES (mm YEAR)

allowed reflux was significantly (P< 0-001) less than thatof the unscarred kidneys with normal ureters. Renalgrowth was not affected by the duration of the bacteri-uria (table III). The effect of bacteriuria upon kidneygrowth was further examined by studying the growth ofthe unscarred kidneys in 31 girls with unilateral kidneyscarring since the unscarred kidneys would be expectedto undergo both natural and compensatory growth andmight therefore more readily reveal a small effect of in-fection upon kidney growth. The mean annual growth-rate of the contralateral unscarred kidneys was 4.2 mm(il’3) compared with 3.6 mm (&plusmn;1.5) for kidneys ingirls who had no renal scarring. Even the enhancedgrowth of these "opposite" kidneys was found to be inde-pendent of the duration of the bacteriuria and of thepresence of v.u.R. Impaired kidney growth was observedin 9 girls (6 controls and 3 treated) with unscarred kid-neys. In 4 of the 9 girls, the kidneys had grown by lessthan 0.3 3 mm per year even though body growth was un-impaired ; in 2 of the 4 the deficiency in kidney growthwas bilateral and in 2 it was unilateral. 3 girls had a dif-ference in renal length of more than 1 -3 cm over the

4-year follow-up and in 2 of the girls with impaired kid-ney growth an initial difference in kidney length of morethan 1.5 cm increased to 2.5 cm by the end of the 4years. V.U.R. was present in 3 of the 9 girls with impairedkidney growth and was therefore not a commoner

finding in girls with impaired kidney growth than ingirls whose kidneys showed no impairment of growth.Bacteriuria was present for more than 50% of the time

TABLE III-EFFECT OF DURATION OF BACTERIURIA ON RENAL

GROWTH (mrn/YEAR)

I L I L-

Numbers in brackets refer to no. of kidneys.

in 6 (66%) of the 9 girls as compared with 57 (37%) ofthe 155 girls with unscarred kidneys who showed no im-pairment of kidney growth. This difference was not sig-nificant.

Kidney Scarring and V. U.R.-No new scars were seenin girls in whom the kidneys were normal at the initialX-ray examination. At the start, 31 girls had unilateralkidney scars and 13 had bilateral scars. 28 girls withscarred kidneys were in the treated group and 16 in thecontrol group. New and/or deepening scars were foundin 12_(27%) of the 44 girls with initially scarred kidneys;6 of the girls were in the treated group and 6 in the con-trol group (table Iv) and 1 of the girls in each groupshowed progression of kidney scars on both sides. None

TABLE IV-DETAILS OF 12 GIRLS WITH PROGRESSIVE RENALSCARRING

+refers to further loss of kidney substance on a pre-existing scar.

TABLE V-EFFECT OF INITIAL SEVERITY OF V.U.R., TREATMENT,AND DURATION OF BACTERIURIA ON COURSE OF V.U.R.

of the girls with progressive scarring had a rising blood-pressure or evidence of deterioration of kidney function(as shown by a rise of blood-urea or serum-creatinine).12 of the 14 kidneys which showed progressive scarringwere drained by ureters which allowed reflux; all of the12 girls with progressive scarring had v.u.R. as com-

pared with the overall prevalence of v.u.R. in 70 (34%)of the 208 girls in the trial. This difference was signifi-cant (xz=22.1; P<0.001). Treatment reduced theduration of the bacteriuria considerably but failed toprevent the progression of kidney damage. However, itmaybe of importance to note that the 3 girls with newscars were all in the untreated group (table iv). Therewas a significant trend for girls with renal scarringand/or v.u.R. to have bacteriuria for a greater portion ofthe time (==4; P<0.05). 8 girls who had v.u.R. at the

892

first radiological examination did not have a repeat mic-turating cystogram. In the remaining 62 girls withv.u.R. the reflux disappeared completely from 33 of the81 ureters which allowed reflux and was reducedin severity in another 17 over the 4 year follow-up(table v). These remissions occurred with equal fre-quency in the treated and control groups and were unre-lated to the duration of the bacteriuria. Of the 12ureters with grade-3 V.U.R., reflux disappeared in only 2(17%) as compared with the clearance of lesser degreesof reflux in 43 (62%) out of 69 ureters; this differencewas significant (=6 - 9; p < 0 - 01).

Discussion

The aim of this study was to determine the sequelxof covert bacteriuria in girls of school age and whethertreatment could prevent any or all of the deleteriouseffects that might occur. For such a study to be entirelysatisfactory the following conditions must be satisfied:(i) all girls entered into-the trial should be infected; (ii)the number of girls in the treated and control groupsshould be large enough to detect even small differencesbetween the groups; (iii) treatment should eradicate in-fection throughout the follow-up; (iv) the follow-upshould be prolonged and complete.

In so far as the presence of more than 105 bacteria perml of urine in two consecutive urine specimens indicatesinfection, we satisfied the first of these conditions. Thenumber of girls included in our study had at the outsetbeen estimated to be sufficient to detect a 10% differencein renal growth between the treated and control groupsover a 4-year follow-up; the second condition was there-fore also met. The difficulties we encountered in satisfy-ing the third and fourth conditions reflect the practi-calities of screening and subsequently treating largenumbers of girls with symptomless infection.The most outstanding observation was that new kid-

ney scars did not develop in girls whose kidneys werenormal at the start of the study. This strongly suggeststhat the kidney scars which are associated with infectionoriginate before the age of 5 and that future attempts atprevention need to be initiated before that age. Thisfinding is in keeping with those of Hodson et al 12 and ofRolleston et a113 who showed that the development ofkidney scars was confined to children aged less than 4in whom severe v.u.R. was associated with intrarenal re-flux. Rolleston et al. also showed 13 that intrarenal refluxdid not occur after the age of 4; neither did any of thechildren in our study have intrarenal reflux.The contrast between our treated and control groups

was blurred by spontaneous and latiogenic remissions incontrols and by the high recurrence rate of bacteriuriain the treated group. Nevertheless, treatment consider-ably reduced the period of exposure to bacteriuria in thetreated group--77% of the treated girls had infectionfor less than 2 years whereas only 38% of the controlshad bacteriuria for less than that time.

As in adult women 14 treatment did not preventsymptomatic infection. Since no changes in blood-pres-sure, blood-urea, or serum-creatinine were observed inuntreated controls it was not possible to judge the effectof treatment upon these variables. The reduction in

exposure to bacteriuria that we produced by treatmentdid not prevent the progression of pre-existing scars butit may have prevented new scar-formation since this

only occurred in the untreated group. Our treatment didnot affect the disappearance of v.u.R. This observationagrees with the recent findings of Edwards et alls in girlswith symptomatic infection. They found that disappear-ance of v.u.R. was as likely to occur in children who hadhad recurrences of infection during long-term prophy-laxis with antibacterials, as in those in whom infectionremained at bay. It is clear from our study that disap-pearance of V.U.R. is time-dependent and that the mostsevere degrees of reflux tend to persist more than thelesser degrees.

Treatment did not influence renal growth. The 3-4mm annual growth-rates of the kidney in bacteriuricgirls with unscarred kidneys were very similar to thoserecorded by Hodson et alll in normal children and bySavage et al 16 in bacteriuric children. As in the Dundeestudy" we found that renal growth was not affected bygrade 1 or 2 reflux. Since there were only 3 girls withunscarred kidneys drained by ureters with grade-3 re-flux, we were unable to judge the effect of severe refluxon renal growth. The observation that kidneys oppositeto scarred kidneys grew well whether bacteriuria and re-flux were present or not shows that renal growth in girlsof school age is not diminished by bacteriuria or themilder grades of reflux alone.

In 9 of the girls with normal initial radiologicalappearances kidney growth was markedly impaired (lessthan 0.75 mm/year) and in 12 girls renal scarring pro-gressed. Thus renal deterioration of one sort or the otheroccurred in 21 (10%) of the 208 girls. Since about 2%of schoolgirls have bacteriuria, our data suggest that theincidence of renal deterioration is of the order of 1 in500 schoolgirls. This remarkable frequency is not re-flected in the returns of dialysis and transplant units.’8Some of these girls with renal deterioration maylater-present with recurrent symptomatic infections and/orrise of blood-pressure either in pregnancy or when start-ing oral contraception or in the absence of these precipi-tating factors. We have not noted such sequelae in the4-year follow-up presented here. We have a limitedamount of information concerning the 40 girls who werenot followed-up completely. We arranged supervision forthose who moved to other parts of Britain and learntthat one of the girls who was originally in our treatedgroup presented with hypertensive encephalopathy andhad her unilaterally scarred kidney removed with resolu-tion of the raised blood-pressure.The girls in this study are still being followed-up to

clarify whether and how often a rise of blood-pressureand symptomatic infections occur. Meanwhile it is clearfrom our observations that screening for covert bacteri-uria in schoolgirls cannot be recommended becausetreatment which markedly reduced the duration of bac-teriuria produced no significant beneficial effects.Further attempts to prevent the kidney scarring associ-ated with infection and its possible late sequelae shouldfocus on children aged less than 5 and should aim toidentify high-risk factors, the most important of whichappears to be v.u.R.

We thank the Health Departments of Cardiff and Oxford, the fol-lowmg nurses, technicians and secretaries: E. M. Beane, S. Chick, J.Critchley, G. Elliott, G. Guthrie, J. Kidd, E. Long, R. Mackenzie, BMoss, M. Smith, P. Wilcock, and S. Williams.The study was supported by grants from the Kidney Research Umt

for Wales Foundation, the Clark Memorial Fund, and the PublicHealth Laboratory Service.

893

Requests for reprints should be addressed to Prof. A. W. Asscher,Royal Infirmary, Cardiff CF2 1SZ.

REFERENCES

1. Kunin, C. M., Deutscher, R., Paquin, A. J. Medicine, 1969, 43, 91.2. Meadow, S. R., White, R. H. R., Johnston, N. M. Br. med. J. 1969, iii, 81.3. Savage, D. C. L., Wilson, M. I., Ross, E. M., Fee, W. M. ibid. p. 75.4. Asscher, A. W., McLachlan, M. S. F., Verrier Jones, E. R. et al. Lancet,

1973, i, 1.5. Brundtland, G. H., Horig, B. Am. J. Epidem. 1973, 97, 246.6. Newcastle Asymptomatic Bacteriuria Study Group. Archs Dis. Childh. 1975,

50, 90.7. Kunin, C. M. New Engl. J. Med. 1970, 282, 4443.8. Asscher, A. W. Lancet, 1974, ii, 1365.9. McLachlan, M. S. F., Meller, S. T., Verrier Jones, E. R. et al. Archs Dis.

Childh. 1975, 50, 253.10. Hodson, C. J., Wilson, S. Br. med. J. 1965, ii, 191.11. Hodson, C. J., Drewe, J. A., Karn, M. N., King, A. Archs Dis. Childh. 1962,

37, 616.12. Hodson, C. J., Maling, T. M. J., McManamon, P. J., Lewis, M. G. Kidney

Int. 1975, 8 (suppl. 4), S50.13. Rolleston, G. L., Shannon, F. T., Utley, W. L. F. ibid. S59.14. Asscher, A. W., Sussman, M. Waters, W. E., Evans, J. A. S., Campbell, H.,

Evans, K. T., Williams, J. E. Br. med. J. 1969, i, 804.15. Edwards, D., Normand, I. C. S., Prescod, N., Smellie, J. M. ibid. 1977, ii,

285.16. Savage, D. C. L., Wilson, M. I., McHardy, M., Dewar, D. A. E., Fee,

W. M. Archs Dis. Childh. 1973, 48, 8.17. Savage, D. C. L., Howie, G., Adler, K., Wilson, M. I. Lancet, 1975, i, 358.18. Parsons, F. M., Brunner, F. P., Gurland, H. J., Harlen, H. Proc. Eur. Dial.

Transpl. Ass. 1972, 9, 3.

REPEATED ADJUVANT CHEMOTHERAPYWITH PHENYLALANINE MUSTARD OR

5-FLUOROURACIL, CYCLOPHOSPHAMIDE, ANDPREDNISONE WITH OR WITHOUT

RADIATION, AFTER MASTECTOMY FORBREAST CANCER

D. L. AHMANNP. W. SCANLONH. F. BISEL

J. H. EDMONSONS. FRYTAK

W. S. PAYNE

J. R. O’FALLONR. G. HAHN

J. N. INGLEM. J. O’CONNELL

J. RUBINDivisions of Medical Oncology and Therapeutic Radiobiology,Department of Surgery and Cancer Center Statistical Unit,

Mayo Clinic, Rochester, Minnesota 55901, U.S.A.

Summary 172 patients who had had mastectomyfor breast cancer were treated by

repeated adjuvant chemotherapy, either with phenyl-alanine mustard (P.A.M.) or a combination of cyclophos-phamide, 5-fluorouracil, and prednisone (C.F.P.) withand without radiotherapy. Tumours recurred signifi-cantly more frequently and mortality tended to be

higher in P.A.M.-treated patients than in patients onother treatment. The interval between surgery and dis-ease recurrence was significantly shorter for P.A.M.-

treated premenopausal but not postmenopausal patientsthan for patients of equivalent menstrual status treated

with C.F.P. with or without radiation. The associationsin premenopausal patients between the mode of treat-ment and both survival and the disease-free intervalwere significant before and after adjustment for vari-ations between the treatment groups in the number ofinvolved lymph nodes and the size of the primarytumour.

Introduction

BREAST cancer is often locally far advanced when firstdiagnosed and despite radical mastectomy the patientmay have a poor prognosis. Local lymph-node involve-ment and the size of the primary tumour-the likeli-hood of local metastasis increasing as the tumour getsbigger-both adversely affect survival. 1-5

In premenopausal patients with prognostically poorbreast cancer the disease-free interval after mastectomywas increased by repeated adjuvant therapy, either withphenylalanine mustard (P.A.M.)6,7 or with a combinationof cyclophosphamide, methotrexate, and 5-fluoroura-cil.s,9 In this study we have compared the effects of fol-lowing mastectomy with repeated adjuvant therapy,consisting of either p.A.M. alone or a combination of cyc-lophosphamide, 5-fluorouracil, and prednisone (C.F.P.)with and without radiotherapy.

, Patients

Originally we planned to study four regimens-namely, sur-gery followed by no further treatment, C.F.P. alone, C.F.P. withradiation, or radiation only. However, when the early resultswith P.A.M. became known/ we abandoned the "no treatment"and "radiation only" arms after 33 patients had been enteredinto the trial. From January, 1975, P.A.M. has been the thirdregimen to be tested.Of 185 patients entered in the trial 46 patients had surgery

elsewhere and were referred to the Mayo Clinic for adjuvanttherapy. 152 of the 185 patients had a modified radical mastec-tomy and 33 patients had a radical mastectomy (8 at the MayoClinic, 25 elsewhere). Patients with radical mastectomy weredistributed equally in the three treatment groups.

Patients (all of whom were willing to take part in a ran-domised clinical trial) were considered for entry into the studyif after operation there was either axillary lymph-node involve-ment or skin infiltration (exclusive of Paget’s disease) withulceration or involvement of pectoral muscle or fascia. Surgerymust have included a complete removal of the breast and axil-lary contents with preservation of the pectoral muscles wherepossible. At the Mayo Clinic modified radical mastectomy ispreferred in primary operable breast cancer, although the radi-cal operation is still used to create adequate tumour-free mar-gins around the primary tumour or its axillary metastases. 5

Patients were not admitted to the trial if they had bilateralbreast cancer, had had previous cytotoxic chemotherapy, wereaged over 70, or had a white blood-cell count <4000/jjd or a

platelet count < 130 000/1. Patients with severe emotional

problems, or who were pregnant or lactating, were also -excluded.

Methods

Before being randomly assigned to treatment groups the pa-

TABLE I-RECURRENCE AND SURVIVAL

* Expressed as rate per 1000 weeks observation.