POSTGRAD. MED. J. (1965), 41, 485

Clinical Review

PYELONEPHRITIS IN CHILDRENAn Interim Review of Recent Literature

MALCOLM MACGREGOR, M.D., F.R.C.P.From the South Warwickshire Hospital Group

PYELONEPHRITIS is now a fast-changing subject,with which the general reader may keep abreastonly if frequent attempts are made to bringtogether published work from different sources.This survey aims to provide a balanced accountof recent concepts, but is in no sense exhaustive.

Incidence of the DiseaseAmong childhood infections those of the

urinary tract are second in frequency only torespiratory infections, and are the commonestbacterial infections under two years of age(Pryles, 1960; Deluca, Fisher and Swenson,1963). The incidence of overt urinary infectionsin the general population is estimated at 8 per1,000 per annum (Percival, Brumfitt andLouvois, 1964), and in American schoolgirlsat 1.4% of the school population per annum(Kunin, Deutscher and Paquin, 1964). BetweenI and 4%/, of hospital admissions of childrenare for this disease (Stansfeld, 1954); 3 to 4%/of recent admissions to an acute p,ediatric wardin London were for urinary infections (Burke,1961). On the other hand, in the Newcastlesurvey of 1,000 families, only 3 out of 847children were noted to have pyelonephritis infive years; all relapsed. The incidence of "signi-ficant bacteriuria", which is often symptomless,(for definition see below) is higher, especially ingirls; 1% of schoolgirls in a recent largeAmerican survey were found to have this(Kunin, Southall and Paquin, 1962; Kunin andothers, 1964). Forty per cent of young womenwith symptomless bacteriuria in early pregnancylater developed pyelitis of pregnancy (Kincaid-Smith, 1964; Percival, Brumfitt and Louvois,1964). The highest incidence of bacteriuriawas found in girls aged between 15 and 19(Kunin, Deutscher and Paquin, 1964). Mostclinicians find that the onset of infection ismost frequent under one year old (Stansfeld,1954: Smellie, Hodson, Edwards and Normand,1964), though some assert that the peak in-cidence is among girls aged three to five(Deluca, Fisher and Swenson; Gross, Randolphand Wise, 1963).

But in infancy the diagnosis is often missed:in one autopsy series it had been missed clinic-ally in 8.3% (Pryle and Neumann, 1962). T-heinitial urinary infection often occurs in thenewborn period; in one series 0.3%/, of hospitalbirths (Smellie and others, 1964) and in another1.5% (James, 1959) were considered to be in-fected. In fact, the incidence among the new-born may be higher than in other age groups;congenital defects in the kidney may pre-dispose (Porter and Giles, 1956). Postmortemstudies suggest that the prevalence of urinary in-fection is still underestimated (Kleeman, Hewittand Gaze, 1960); about 2%/, of routine autopsieson American children disclose evidence ofpyelonephritis (Pryles and Neumann, 1962;Spark, Travis, Dodge, Dalschmer and Hopps,1962; Macaulay, 1964), but the difficulties indiagnosis of this infection from postmortemmaterial must be borne in mind. An increasedincidence of pyelonephritis in siblings has beennoted (Kunin, Deutscher and Paquin, 1964).

The Course and Prognosis of ChildhoodUrinary InfectionsA tendency for the infection to relapse or

to become persistent is an alarming feature ofoyelonephritis. Some 80-95%/, of treated primaryinfections are "cured" (when defined as freefrom bacteriuria six weeks after cessation oftreatment), but about half of these have furtherperiods of infection (Macaulay and Sutton,1957; Lancet, 1963). Recurrences seem morecommon in older children than in infants(Burke, 1961) and especially so in girls of thesix to ten age group (Williams and Sturdy,1961). Persistence of infection is common,mainly in girls (Dunn, Hine and MacGregor,1964; Woodruff and Everett. 1954), and onethird to one half of continuing infections areasymptomatic (Steele, Leadbetter and Crawford,1963: Dunn and others, 1964). Admittedlythe significance of "significant bacteriuria" asthe only finding is still debatable, but thereseems to be little tendency for this to die outspontaneously (Kass, 1956; Kleeman and others,

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

1960; Spence, Murphy, McGovern, Herdionand Pryles, 1964). Persistent infection has beenfound with different frequency in different sur-veys, in 10% (Burke, 1961), in 25% (Dunnand others, 1964), in 39% (Woodruff andEverett, 1954). Of the chronic and recurrentgroups the outcome varies, but the prognosisis believed to be worst when the onset is undertwo years of age. Some clear up, spontaneously(Kunin, Deutsches and Paquin, 1964) or withtreatment (Turck, Browder, Lindmeyer, Brown,Anderson and Petersdorf, 1962; Macaulay,1964) or at adolescence (Turner-Warwick, 1962;Williams, 1964). Others progress to renal failure(Rosenheim, 1963). The follow-up in adultsafter 20 years can be appalling. In a surveyrelating to the pre-antibiotic era, 43%/0 of youngwomen who had been treated for acute pye-lonephritis had serious urological disease after16 years (Hanley, 1964). There is not yet muchinformation about the long-term prognosis forchildren. The incidence of toxemia of pregnancyand pyelitis of pregnancy is known to be un-usually high when there is a history of urinaryinfection in childhood (Steele and others, 1963),although a history of childhood infection wasfound in only five out of 75 pregnant womenwith pyelitis in one enquiry (Woodruff andEverett, 1954). There seems to be little correla-tion between the number of recurrences and theeffect after some years on the kidneys (Steeleand others, 1963). The equal sex incidenceof chronic pyelonephritis at autopsy in adultsdoes not seem consistent with the much higherincidence of symptomatic infections in girls,suggesting that the one is not the direct ante-cedent of the other (Macaulay, 1964). However,some girls with recurrent infections have beenobserved to develop progressive loss of kidneyparenchyma with a fatal issue (Williams, 1965).Although pvelonephritis is the second common-est cause of hypertension in adults, in two-thirdsof such cases no history of recurrent urinaryinfection could be obtained (Kincaid-Smith.McMichael and Murphy, 1958). Indeed, it issaid that a history of acute pyelonephritis israre in advanced chronic nvelonephritis (Kim-melstiel, Kim, Beres and Wellman. 1961). Onthe other hand, in a recent series of 200 childrenseen for urinary infection at a hospital, 13%/,had X-ray evidence of chronic pyelonephritis(Smellie and others, 1964).

Whatever its ultimate results, chronic urinaryinfection is difficult to eradicate. With chemo-therapy the prognosis is still depressing (Mac-aulay and Sutton, 1957). Uncomplicated primaryinfections treated continuously for up to two

years still show a 7% relapse rate afterwards(Campanacci, Bonomini and Zuchelli, 1963).In chronic or recurrent infections in adultsthere is only a 10% (Lancet, 1963) or 20%(Turck and others, 1962) cure rate, in spite ofreally prolonged treatment (up to two years)(Williams, 1963). In children, after conventionaltreatment of acute infections there may be a250/% persistence rate (Dunn and others, 1964).There is not much evidence that prolongeddrug treatment will produce better figures(Lancet, 1963), unless perhaps in infants (Stans-feld and Webb, 1954), though assertions aremade to the contrary (Smallpeice, 1958). Inadults, non-antibiotic urinary disinfectants suchas methenamine mandelate (Holland and West,1963) are capable of giving as good results asthe powerful modem antibiotics tetracycline,chloramphenicol and kanamycin (Turck andothers, 1962), especially in older age groups(Lindmeyer, Turck and Petersdorf, 1962). Be-cause so many infections are clinically silent.prolonged follow-up of treated cases is essential(Lancet, 1963). Such follow-up studies inchildren are in progress in a number of centresand should soon provide facts about the trueprognosis in childhood. Early treatment ofprimary infections is important (Deluca andothers, 1963), for delay in diagnosis increasesthe relapse rate (Stansfeld and Webb, 1954).Once pyelonephritic scarring of a kidney hasoccurred, infection is very difficult to eradicate(British Medical Journal, 1964). Delay in diag-nosis is common, and averaged 18 months inone recent survey (Deluca and others, 1963).In another (Spence and others, 1964), 47%of children sent to hospital had had symptomsfor one year or more, and in a third series,only 30%/O of children had been sent to hospitalwith a suggestion of the correct diagnosis(Smellie and others, 1964).

Clinical FeaturesIn infancy, anorexia, loss of weight, vomiting

and failure to thrive are the main symptoms;in older children, fever or abdominal pain(Burke, 1961: Smellie and others, 1964). Only25%/, have micturition symptoms (Burke, 1961).Hematuria occurs in 10% of acute attacks(Kleeman and others, 1960; Burke, 1961). Re-current febrile urinary infections are known tooccur in older girls (Williams and Sturdy, 1961).and the symptoms tend to be the same eachtime (Burke, 1961). In older children chronicinfection may be quite silent, but close question-ing will often uncover some disregarded urinarysymptom, such as enuresis, frequency or

486 August, 1965

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

MAcGREGOR: Pyelonephritis in Children

dysuria (Dunn and others, 1964). Ureteric re-flux occasionally gives rise to pain in the loinon micturition (Forsythe and Wallace, 1958).Episodes of partial or complete urinary reten-tion may be described (Gross and others, 1963).Impor,tant clues may be revealed if questionsabout the rate and force of the urine stream,and whether straining is present, are a part ofthe routine enquiry (New England J. Med.,1963; Gross and others, 1963). However, ana-lysis of symptoms does not allow a judgementto be made of which cases have urologicalabnormalities, and which have not. (Forsytheand Wallace, 1958; Kunin and others, 1964).Neither is the severity of symptoms proportionalto the extent of existing, or of future, renaldamage (Steele and others, 1963). The growthof patients with X-ray changes of pyelonephritisis sometimes a little reduced. There is occasion-ally hypertension, or another congenital mal-formation (Smellie and others, 1964). Polyuriaand nocturia are early signs of serious renalimpairment (Kleeman and others, 1960).

Diagnosis of Urinary InfectionThe danger of catheterisation as a means of

introducing infection is increasingly recognized,especially in obstetric practice (Brumfit't, Daviesand Rosser, 1961). The risk of infection inchildren is probably less (tho-ugh it was 70/oin one series (Vertanen, Oksanen and Pettonen,1962); catheterisation should not be withheldif needed for urgent diagnosis, retention, non-cooperation or uncertainty (Pryles, 1960). Butin children the bacteriological correlation bet-ween catheter and clean-voided specimens canbe 96% (Pryles, 1960) and at all ages mid-stream specimens without bacterial contamina-tion can be obtained from a high proportionof both sexes (Clarke, 1960; Pryles, 1960;Vertanen and others, 1962; Cattell and Lefford,1963). The most precise results are obtainedfrom suprapubic aspiration of the bladder(Monzon, Ory, Dobson, Carter and Yow, 1958);this method gives only a 93%/0 bacterial cor-relation with catheter urine, but is higher if thefirst few millilitres of catheter urine are dis-carded,(Pryles, 1960). Many different,techniquesof perineal cleansing are employed before takingclean specimens (Brumfitt and others, 1961;Cattell and Lefford, 1963; Houston, 1964);benzylkonium chloride 1 in 1000 is much used(Little, 1962), polybactrin spray has advocates(McLeod, Mason and Pilley, 1963), but soap-and-water seems satisfactory (Clarke, 1960).Different methods of cleansing do not seem toinfluence urine cell counts (Lincoln and Win-

berg, 1964). The plastic bags, much used forobtaining specimens from young children, areliable to introduce contamination into bacterialcounts, but have less influence on cell counts(Houston, 1964).

Proteinuria is of little aid in detecting urinaryinfection (Burke, 1961; Lancet, 1962a and b;Steele and others, 1963; Dunn and others, 1964;Smellie and others, 1964); albumin was foundas frequently in the urines of unaffected school-girls, as in those with significant bacteriuria(Kunin and others, 1964). The lack of a simplebut reliable screening test for infection is stilllamented (Lancet, 1962). Controversy hasfocussed upon the relative accuracy of methodswhich estimate bacteria, and of those whichcount cells, in the urine. The correlations bet-ween urinary cell concentration, timed cell ex-cretion rates, and bacterial colony counts arenot close, and misleading results can come fromrelying upon any one of them (Osborn andSmith, 1963). In America colony counts arepreferred (Brumfitt and Percival, 1964; Spenceand others, 1964). For children in hospital puscell counts have been found as useful asbacterial counts in identifying infections; witheither method there were 13-25% doubtfulresults; when the methods were used in con-junction uncertainty was reduced to 6%(Houston, 1964).

Direct examination of freshly voided urinemay show organisms under the microscope;bacteriuria of this magnitude is never due tocontamination, so this is a valuable screeningtechnique (Gardborg, 1959; Dunn and others,1964). Again, if the uncentrifuged sediment ofa urine shows organisms after Gram staining,the bacterial count is nearly always above100,000/ml. (Pryles, 1960; Kunin and others,1964). Laboratory techniques for demonstratingbacteriuria employ quantitative, semiquantita-tive, or chemical screening methods. Urine isan excellent culture medium, so within twohours at room temperature a large increase inbacterial numbers may occur (Pryles, 1960).This can be prevented by refrigeration of freshspecimens at 40C, which preserves the bacterialcontent unaltered for 2-7 days (Pryles, 1960;Houston, 1964). Bacterial colony counts onlarge numbers of fresh urines show a bimodaldistribution, with peaks at less than 10,000 andat more than 100,000 organisms/ml. There ismuch evidence that organisms derived fromthe urethra or vulva never exceed 104 per ml.,whereas bacteria rapidly multiply in the bladderto reach numbers in excess of 105 per ml.,the level of "significant bacteriuria" (Kass,

August, 1965 487

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

1956, 1960). Significant levels may have to beslightly higher in the case of young children(Houston, 1964). Quantitative bacterial countson urine are made by the surface viable, orpour plate techniques (Brumfitt and Percival,1964). These methods are too time-consumingfor most routine laboratories, so semi- quantita-tive modifications are often used. Using freshuncentrifuged urine and a standard loop, colonycounts of more than 200 equate witih morethan 105 organisms per ml., and counts of lessthan 90 with less than 104 organism per ml.This is sufficiently accurate for clinical practice(O'Sullivan, Fitzgerald, Meynell and Malins,1960). Another technique employs a blottingpaper strip dipped into urine and pressed on tothe culture medium; results have been stand-ardised against pour plate counts (Leigh andWilliams, 1964). Conventional methods of des-criptive reporting on the bacteriology of urinehave been compared with the results of bacterialcounting; it is claimed that if standards arerigorous, closely similar conclusions will bedrawn from each method, and that experimentalresults are not necessarily invalidated by failureto carry out bacterial counts (Clark, 1960;Guttman and Stokes, 1963). The T.T.C. testis a chemical screening method based uponnaked-eye observation of colour change in theurine after incubation of urine with test materialin a water bath (Simmons and Williams, 1962).It is claimed to give a positive reading withbacterial counts of gram negative organismsover 100,000/mI., and has proved successfulin detecting 86% of antenatal women withasymptomatic bacteriuria of pregnancy (Kin-caid-Smith, Bullen, Mills, Fussell, Huston,and Goon, 1964), and even better results arereported. The level of certain enzymes is alsoreported to be raised in infected urines (Bren-ner and Gilbert, 1963).

Pyuria may result from other causes thanurinary infection, from vaginitis, trauma, de-hydration and nephritis, as well as at the endof a urinary infection when organisms havegone (Pryles, 1960). Until recently the criteriaaccepted by many authors as "significantpyuria" were not uniform (Lancet, 1962). Acommonly employed criterion has been "morethan five cells -per high power field" (represent-ing, if centrifuged, approximately 25 cells/cu. mm., if uncentrifuged, approximately 250cells/cu. mm. (Stansfeld, 1962, 1963). On thisbasis, only 35% of children with significantbaoteriuria had also significant pyuria (Kuninand others, 1962), and, in another series, onlyhalf the infected children had pyuria (Steele

and others, 1963). But this standard is no moreacceptable than are the still much employeddescriptive reports ("moderate numbers" or"scanty pus cells") which for CSF, with a similarrange of cell numbers, are never countenanced(Stansfeld, 1962; Little, 1964). Of course,chronic pyelonephritis with a cell-free urine iswell recognized (Clarke, 1960; Deluca andothers, 1963; Leather, Wills and Gault, 1963),and both pyuria and bacteriuria may be inter-mittent (Stansfeld, 1954; Dunn and others,1964); but bacilluria without pyuria, using asensitive criterion, occurred only three times in250 specimens from children (Stansfeld, 1962).Timed white-cell excretion rates, which do notseem to be much influenced by the rate ofurine flow, are available for adults, but not yetfor children (Houston, 1964). An upper limitof excretion of 400,000 cells per hour is notexceeded in normal men or women (Kass,1956; Little, 1962; Gadeholt, 1964); this isprobably the best measurement of pyuria(Gadeholt, 1964). Centrifugation in any cellcounting technique introduces important errors,as it leads to considerable and unpredictableloss of cells (Gadeholt, 1964; Little, 1964). Cellcounts in uncentrifuged specimens have beenfound in adul,ts to give a good correlation withtimed cell excretion rates; more than 10 cells/cu. mm. means more than 400,000 excreted perhour, while less than three cells per cu. mm.means less than 400,000 per hour with faircertainty; there is no sex difference (Little,1964). Quantitative methods of leucocyte count-ing in a counting chamber are increasingly usedfor children (Stansfeld, 1954), and have lessvariation than methods based on high powerfield counts. It is however important to ensurea urinary pH of between six and eight (Stansfeld,1962).Semiquantitative wet film screening techniques

of cell counting have proved useful in somehands (Dunn and others, 1964), though criticizedby others in respect of variability (McGrackieand Kennedy, 1963). After infancy, 92% ofurines from normal children contain less than10 leucocytes/cu. mm. (Stansfeld, 1962). Oldergirls give more misleading cell counts (Houston,1964). One centre accepts less than 10/cu. mm.as normal, more than 100 as infected, and inter-mediate values as dubious (Houston, 1964). Alevel of 50/cu. mm. or more is regarded byothers as indicating infection (Stansfeld, 1954).In the newborn, counts greater than 25/cu. mm.in boys, or 50/cu. mm. in girls, are abnormal(James, 1959; Lincoln and Winberg, 1964).Neonates, especially males, may excrete epithe-

488 August, 1965

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

MAcGREGOR: Pyelonephritis in Children

lial cells in considerable number and causeconfusion (Lincoln and Winberg, 1964); differ-ential staining may then make counting easier(Sternheimer and Malbin, 1957; Prescott andBrodie, 1964; Houston, 1964). Taking a levelof 10 leucocytes/cu. mm. as "significantpyuria", some 14% of admissions to a children'shospital were found to have it, though countsbelow 50/cu. mm. seldom turned out to beabnormal (Stansfeld, 1954, 1962). In general,those who have studied the subject over along period now accept that the diagnosticvalue of leucocyte counts is comparable tothat of bacterial counts (Osborn and Smith,1963). Wlhen chronic pyelonephritis is suspected,attempts to stimulate leucocyte excretion canbe made, using intravenous bacterial pyrogen(Leather and others, 1963), or, probably withmore reliability ('Montgomerie and North,1963), with oral steroid drugs (Huth, Chalmers,Macdonald and de Wardener, 1961; Little andde Wardener, 1962). There are no reports ofthe use of these methods for children. There isneed for a test to distinguish reliably betweeninfection involving the kidney, and that confinedto the lower urinary tract (Brumfltt and Perci-val, 1964). The presence of leucocyte casts inurine means renal infection, and by using aspecial stain, morphological differences areapparent between leucocytes of renal origin andthose from more distal sources (Sternheimerand Malibin, 1957; Poirier and fackson, 1957),which can be used as a diagnostic aid. Whenthe serum from a patient with an acute urinaryinfection is used for direct agglutination againstorganisms obtained from the urine, a highand rising titre seems to correlate with involve-ment of renal tissue. A titre of more than 1: 160suggests such infection: it falls rapidly wheninfection is controlled (Percival and others,1964). However, in another large group ofbacteriuric girls no such relationship was found(Kunin and others, 1964).

Reasons for Persistence of InfectionThe insidious advance towards chronic

pyelonephritis, which is so resistant to treat-ment, has prompted much speculation aboutpredisposing factors. Major structural abnormal-ities of the renal tract are readily demonstrated,and do not contribute to the problemunder discussion. The megacystis-megauretersyndrome, for example, is found in 6.5%/, ofpersistent urinary infections (Kunin and others,1964). A retrospective survey of a group ofgirls with recurrent urinary infections, pre-viously investigated, failed to demonstrate any

child in whom serious obstruction had beenoverlooked (Williams and Sturdy, 1961). Butabout 10%/, of affected girls show recurrence orpersistence of infection, and a large proportion(variously estimated at 25% (Dunn and others,1964), 50% (Spence and others, 1964) or 75%(Macaulay, 1964) have structurally normalurinary tracts. It is plain that damage fromone attack confers in some way a susceptibilityto further attacks (Lancet, 1963), and in ex-perimental pyelonephritis intra-renal scarringpermits of infection more readily, as also doareas of renal dysplasia (Kleeman and others,1960). Non-bacterial agencies, such as an under-lying virus infection, or auto-immune reactions,or progressive renal vascular damage have beenconsidered, without good evidence to favourthem (Brumfitt and Percival, 1964). Bacteriaare in fact probably responsible; it has beensuggested that -they could persist within thekidney as spheroplasts or protoplasts (Brum-fitt, Percival and Williams, 1964), but thetendency for one organism to be replaced byanother after treatment argues to the contrary(Gard'borg, 1959; Turck and others, 1962).There may be "pyelopathic" strains of E.Coli(Brumfitt and Percival, 1964; Sweet and Wolin-sky, 1964). Again, anti-bacterial substances areknown to exist in normal urine, and may bereduced in diseased states (Kleeman and others,1960). But of late years increasing attention isbeing given to vesico-ureter reflux as a factorpromoting persistence of infection.

Vesico-ureteric RefluxRetrograde movement of urine from the

bladder up the ureters rarely if ever occurs innormal children (Hodson and Edwards, 1960;Williams, 1964; Spence and others, 1964),although it is a normal finding in rats, and canbe borne in dogs under experimental conditionsfor long periods without ill-effects except someprogressive depression in ureteric peristalsis(Scott, 1962). Reflux occurs in man in congenitalmegaureter, in bladder wall abnormalities, withneurogenic lesions, with lower urinary obstruc-tion, and with chronic pyelonephritis (Turner-Warwick, 1962; Gross and others, 1963; NewEngland J. Med, 1963). It is also observedfrequently in association with trigonitis andurethritis in young women as a transient pheno-menon, and in this way probably disposes toascending infection of the kidneys (Hanley,1964). Its occurrence in obstructive uropathyis generally accounted for by the protrusion ofa mucosal saccule at the ureteric orifice, whilein non-obstructive urinary infections reflux is

August, 1965 489

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

ascribed either to a congenital deficiency ofureteric muscle, to a shortened intramuralcourse from a contracted bladder, or to rigidityfrom cedema of the bladder wall (Williams,1964). In differen.t patients, reflux may occurat different levels of intra-vesical pressure, andof bladder filling (Melick, Brodeur and Karel-los, 1962b). Ureteral competence is achieved-by a longitudinal crumpling of intra-vesicalmucosa (a "whistle-valve" effect), and if thisfails reflux may occur without any rise of in-travesical pressure (Melick, 1962a; Stephens andLenaghan, 1962). Reflux may remain unilateral,as free reflux quickly reduces pressure withinthe bladder. It may disappear spontaneously(Turner-Warwick, 1962), and because i.t is notpresent at one examination, it is no proof thatit was never present (Johnston, 1963); it isoccasionally seen with a sterile urine and noevidence of previous infection. It may occurfor the first time during a urinary infection(Hodson and Edwards, 1960), and is recordedin 18% of all infections in girls (Kunin andothers, 1964). In recurrent urinary infections re-flux is commoner in children than in adults, andis reported in 15-30% of non-obstructivepyelonephritis (Palken and Kennelly, 1960;Deluca and others, 1963). Radiographic evid-ence of chronic pyelonephritis without refluxat any time has seldom been found (Williams,1965), and the triad of chronic pyuria, vesico-ureteric reflux and pyelonephritic scarring ofthe kidney is a familiar one (Johnston, 1963).Reflux has been reported in 70-80% of childrenwi.th chronic pyelonephritis (Allen and Burrows,1964; Hodson, 1964). There is frequent associa-tion with hydronephrosis, with renal scarringand with duplex kidneys (Smellie and others,1964). There is good evidence that the effectsof reflux back pressure on the kidneys maybe serious (New England J. Med., 1963), andthat pyelographic abnormalities may be caused,presumably via effects upon the calyceal bloodvessels (Williams, 1964). These effects arecalyceal blunting, general diminution in size ofthe kidney, and failure of the kidney to grow(Hodson, 1959; Hodson and Edwards, 1960;Hodson, Drewe, Karn and King, 1962). Inmost cases a combination of backpressure andinfection is present, but the only certain radio-graphic indication of infection (whose othereffects are indistinguishable from those of re-flux) is coarse localised scarring of the renalsubstance (Hodson, 1959), which takes two toseven years to develop (Hodson, 1964). Refluxis present so often when there is X-ray evidenceof renal damage in pyelonephritis, that it may

be a necessary condition for damage to occur(Allen and Burrows, 1964). The ill-effect of re-flux is well demonstrated in patients with con-genital doulble ureter, the upper calyx of whosekidneys is always drained by a ureter openinglow in the bladder, after an intramural courseof exceptional length. The other ureter has ashort intramural course, and therefore is apt tobecome incompetent, so allowing reflux to occur.In these patients it is always the lower calyceswhich become damaged and infected (Williams,1962). In adults, where reflux and chronicinfection are associated, the untreated prog-nosis before renal failure is two or three decadesat most (Hodson and Edwards, 1960). If thereis an element of obstruction, the renal failuremay be of a kind with a normal blood urealevel, but with excessive water and electrolytelosses (Lancet, 1962; Jones and Mills, 1964).There may be renal osteodystrophy (Hodsonand Edwards, 1964).

Lower Urinary ObstructionThere is no general agreement upon how

often persistent infection is associated withminor degrees of obstruction to the lowerurinary tract, though this is a well-recognizedassociation with ureteric reflux (Hodson, 1959;Hodson and Edwards, 1960; Williams andSturdy, 1961; Turner-Warwick, 1962; Delucaand others, 1963; New England J. Med., 1963;Gross and others, 1963). In a long follow-upaveraging sixteen years it was found that allchildren with advancing renal disease hadobstruction to urinary outflow (Steele and others,1963). Organic abnormalities of the urinarytract, including reflux, are found in 50-75%/ ofchildren with persistent infections (Allen andBurrows, 1964; Dunn and others, 1964). Obs-tructive uropathy has been reported in 16%(Deluca and others, 1963) and in 35%/, (Steeleand others, 1963) of large surveys of suchchildren. Obstruction is most often at thebladderneck in girls, and the lack of clear-cut objectivecriteria for this diagnosis, has made the entityan unsatisfactory one. Bladder neck obstruc-tion cannot be diagnosed endoscopically, anddiagnosis depends upon the observation of non-progressive bladder muscle hypertrophy, withovergrowth of tissue at the bladder outlet, in-the absence of more distal obstruction (Williamsand Sturdy, 1961). In practical terms this meansthe presence of residual urine (not due to nerv-ousness or to reflux), and, on cystoscopy, thepresence of trabeculation of the bladder wall.Suggestive additional features are raised intra-vesical voiding pressure, ureteric reflux and

490 August, 1965

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

MAcGREGOR: Pyelonephritis in Children

sacculation of the bladder (Williams and Sturdy,1961; Gross and others, 1963). Except forbladder neck hypertrophy itself, all thesefeatures can be due to other causes, and evennow bladder neck obstruction is not accepted byall surgeons (Johnston, 1963), but most considerit real and important (Williams and Sturdy,1961; Gross and others, 1963; Williams, 1965),with an onset usually under three years old, andpresent, according to one estimate (Spence andothers, 1964), in 25% of recurrent infections,though this may be too high a figure (Williams,1965). The "spinning-top appearance" of theurethra during a voiding cystogram, previouslythought to indicate bladder neck obstruction, isprobably a normal appearance. Likewise the"wide bladder neck deformity" in girls, re-garded as an abnormality predisposing to in-fection by some workers (Forsythe and Wallace,1958), may be a normal variant (Burrows andAllen, 1964).

Surgical Treatment of Ureteric Reflux andBladder Neck ObstructionAbout ureteric reflux there is not yet sufficient

evidence to justify invariable corrective surgery(Scott, 1962; TurnerAWarwick, 1962; Macaulay,1964), nor has time enough elapsed to assessfully the results of corrective procedures(Johnston, 1963). Most surgeons advocate firsta period of conservative management, with anti-biotics and the use of "triple micturition" (Grossand others, 1963). For bladder neck obstruction,graduated urethral dilatations are urged bysome surgeons (Palken and Kennelly, 1960;Knappenberger, 1963). The presence of dilata-tion of the upper urinary tract is a widelyaccepted indication for operation (Turner-Warwick, 1962; Gross and others, 1963), mostoften a reimplantation of the ureters into thebladder wall, as well as a plastic procedureupon the bladder neck. Postoperative follow-up for several years in one series showed thatinfection was unlikely to return if there hadbeen no pre-operative dilatation of the upperurinary tract (Gross and others, 1963). Anotherexperienced view is thaat successful reimplanta-tion of the ureters will lead in most instances tocontrol of infection without chemotherapy(Williams, 1964), and that results justify blad-der neck surgery alone in a relatively smallnumber of girls. Unless reflux coexists, pye-lonephritis is not usual with bladder neckobstruction (Williams, 1965).

Techniques of InvestigationExcretion pyelography. This will display

dilatation of the calyceal system in albout 13%of persisting childhood urinary infections(Kunin and others, 1964). Blunting or shorten-ing of the calyces, the size of the kidneys, andthe presence of coarse localised depressionsupon them, may also be revealed. For smallchildren tomography (one or two cuts) mayhelp to evaluate the pyelogram. Concentrationof dye may be remarkably good even withseverely damaged kidneys (Hodson, 1959).Weakening of the picture in later films mayindicate dilution from ureteric reflux (;Hodson,1964). Tonelessness of the lower ureters (a"flalbby dilatation"), or a ureter visible in itsentire length may be pyelographic pointers tothe presence of reflux (Marshall, 1962; Turner-Warwick, 1962; Johnston, 1963). Kinking maybe visible at the upper end of such a ureter(Williams, 1964). The "ureteric spurt reaction".a jet of urine from the ureteric orifice acrossthe bladder, may be noted in the I.V.P. ofchildren with urinary infections (Nevin, Cline,and Haug, 1952). A nephrogram may be ob-tained by the rapid intravenous injection of50 ml. of 70% organic iodine solution .(Hodson,1959). A pyelogram is advocated in all firsta,ttacks of urinary infections, as soon as theinfection is controlled (British Medical Journal,1964; Spence and others, 1964).

Cysto-urethrography. A voiding or micturat-ing cystogram is necessary for the diagnosis ofureteric reflux. If X-ray screening is possibleduring the procedure, more useful results willbe obtained (Allen and Burrows, 1964). Theprocedure is indicated in recurrent infections,or if reflux is suspected on the pyelogram, orif renal scarring is visible (British MedicalJournal, 1964). The procedure is often moreuseful than a pyelogram as a first investigationin neonates (Smellie and others, 1964).Reflux may be difficult to demonstrate. Theearliest indication may be fleeting filling ofthe lower ureter (best seen in an oblique-view)during micturition. More obvious reflux pro-duces distension of the pelvis of the kidneyduring voiding (Johnston, 1963). At all ages,a cystogram is more often revealing in urinaryinfections than a pyelogram. Thus, in one seriesof 500 children submitted to both procedures,the cystogram was abnormal on 237 occasionsand the pyelogram on only 27 (Allen andBurrows, 1964). Another large series reportedan albnormal IVP in 22% and an abnormalcystogram in 44% of cases (Lancet, 1962).Using the two procedures together, the needfor retrograde pyelography is virtuallyeliminated (Forsythe and Wallace, 1958

August, 1965 491

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

POSTGRADUATE MEDICAL JOURNAL

Hodson, 1959; Burrows and Allen, 1964).Cineradiography. This is a refinement of the

voiding cystogram, but provides little informa-tion that cannot be obtained by other techniques(Clarke, 1960; New England J. Med., 1963;Smellie and others, 1964).

Cystoscopy. An essential investigation in thediagnosis of bladder neck obstruction.

Estimation of residual urine. Aid may beobtained by introducing floating lipiodol intothe bladder at the time of the cysitogram. Ifthere is no residue, all should be excreted with-in 24 hours. With reflux, radio opaque materialmay ascend to the pelvis of the kidney (NewEngland J. Med., 1963).

Estimation of voiding pressures within thebladder. This is useful for assessing the resultsof operation; measurement can be made througha small urethral polythene tube (New EnglandJ. Med., 1963; Gross and others, 1963). Normalintravesical pressures are 0-18 mm. Hg. resting,rising to 70-100 mm. Hg. voiding (Melick,1962a). Other authors record lower voidingpressures of 15 to 20 mm. Hg., and regardedpressures of 40-50 mm. Hg. or more as in-dicative of obstruction (Gross and others, 1963;Spence and others, 1963). Reflux may reducethis.

Detection of reflux by a radioactive isotope.By using intravenous 1311-hippuran, a nephro-gram can be obtained with a scintillationcounter during the passage of ;the dye throughthe kidney, and again during micturition ifthere is reflux (Dodge, 1963).

Aortogram. This investigation is of specialvalue if hypertension is present, or in othercases where chronic disease may be unilateral.In one centre an aortogram is regarded as thenext step if excretion pyelography and a voidingcystogram prove insufficient (Hodson, 1959).

Renal biopsy. The results of this aregenerally disappointing in pyelonephritis (seebelow).

It is generally agreed that incomplete studyof these cases, which may lead to prematureor misplaced surgery, is dangerous (NewEngland J. Med., 1963).

BacteriologyA pure growth of organisms is obtained

from the urine in some three quarters of child-hood urinary infections. Of these E.Coliaccounts for the majority, 60-80%. B.Proteus,often in mixed growth, is found in 12-17%,cocci (Staph. aureus and enterococci) in 6-8%,and pseudomonas in 4% (Gardborg, 1959;Pryles, 1960; Deluca and others, 1963; Kunin

and others, 1964). Pseudomonas and proteusare both related to instrumentation or to sur-gery (Kleeman and others, 1960), or to previousdrug treatment (Turck and others, 1963).Staphylococcus aureus has been recorded as thesecond commonest organism in pure growth(Pryles, 1960). In recurrences, a new strain oforganism is commonly found; estimates forthis even vary from 30-87% (Kunin and others,1964; Spence and others, 1964). Treatmentfrequently leads to replacement by a differentorganism in the urine (Gardborg, 1959; Turckand others, 1962). Renal biopsy, presumablybecause of the patchy nature of the inflam-matory process, has not yielded usefulbacteriological information; results have beendescribed as "mystifying and sparse" (Kleemanand others, 1960).

PathologyThe route of infection of the urinary trace in

most cases is regarded as ascending. Haema-togenous infection which certainly occurs instaphylococcal cases, must be accepted as arare possibility (Kleeman and others, 1960).Experimentally, haematogenous infection bygram-negative organisms will onlly occur ifabnormal renal conditions, such as uretericligature, are present (Brumfitt and Percival,1964). Organisms can ascend a stagnant filmof fluid, and they can be found, according toone paper, at 5 cms. depth in the male urethra(Spence and others, 1964), so ascent to thebladder would not be difficult. Urethritis iscommon in many adult women (Brumfitt andPercival, 1964), and cystourethritis precedesacute pyelonephritis in many young women;transient reflux up the ureter has been observedmany times in the former condition (Hanley,1964). The concepit in some cases of lowerurinary infection without pyelonephritis is gain-ing ground (Mackinnon, 1964). To allow ofmultiplication of organisms within the bladdera disturbance of dynamics is usually necessary,especially deficient emptying (Brumfitt andPercival, 1964; Spence and others, 1964).When inflammation of the renal pelvis is

present, renal parenchymal infection is almostinvariable (Weiss and Parker, 1939). Acutepyelonephritis is primarily an inflaimmation ofthe interstitial tissues of the kidney. Startingin the medulla, the septic process soon burstsinto the tubule, and leucocytes and organismsare released into the urine. Glomeruli areremarkably resistant to the suppurative process,although they too become involved in an" i n v a s i v e glomerulitis". Haematogenous

August, 1965492

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

MAcGREGOR: Pyelonephritis in Children

staphylococcal pyelonephritis, the so-calledcarbuncle of the kidney, gives rise to constitu-tional symptoms with local tenderness, butoften at first no pyuria. Such infections usuallyheal completely.

Features by which to differentiate kidneysdamaged by healed or chronic pyelonephritis'from other renal disturbances have been muchdiscussed, and absolute criteria are hard to find'(Weiss and Parker, 1939; Kimmelstiel andothers, 1961). In general, such kidneys areirregularly contracted, often nodular, withadherent thickened capsules and flat U-shapedscars. The pelvis may be thickened and infil-'trated. Renal papillae are atrophic, andcalyces dilated and shortened. Within, there isa characteristic patchiness of involvement, withhyperplasia of unaffected nephrons. Theinterstitial tissues are much inifilltrated withlymphocytes and plasma cells but the presenceof many polymorphs is essential to the inflam-'matory diagnosis. Many tubules show dilatationand atrophy and contain colloid material, so-called "thyroid-like" areas, which are typical ofpyelonephritis. The glomeruli are relativelywell preserved, even in scarred areas; peri-glomerular fibrosis is common, and whenadvanced renal failure existed, focal necrosisof the kind descri'bed as "alterative glomerulitis"is frequently seen. Hyperplastic arteriosclerosis,("onion-peel arteritis") of varying severity isregularly seen, and in many such kidneys thereis an intermediate or "barrier" zone of relativepreservation between an overlying necrotic andscarred cortex, and the contracted medullawith dilated calyces. Most of these changesare non-specific and the presence of abundantnolymorphs and of thyroid-like areas are pro-bably the most reliable registers ofnyelonephritis. "Chronic active pyelonephritis"is a useful concept to denote the scarredpyelonephritic kidney in which some acuteinflammatory change is still evident. A lesscommon group of chronic pyeloneohritic'kidneys are smaller and evenly and diffuselycontracted (Kincaid-Smith and others, 1958).They may show adenoma-like nodularity (Weissand Parker, 1939). These are seen in youngerpeople, often with growth failure and renalrickets. They probably represent inflammationthat has begun in very early life. ("Congenitalhyponlastic kidneys," which are bilaterallysmall but of normal structure. are rarely if everseen (Weiss and Parker, 1939)). In neonatalpyelonephritis, histological appearances may beconfusing because of the presence of glomerular

crescents which suggest nephritis. Dysplasticlesions of the nephron may contribute to theincidence at this age (Porter and Giles, 1956).

There is no other type of nephropathy inwhich fluctuation in renal function occurs sofrequently; patients may approach and recedefrom renal failure in a most variable manner,according to the state of activity of theinflammatory process. The residual renal units,though reduced in number, function normally;the consequence is decreasing regulatory flexi-bility with increasing azotaemia. Hyper-chloraemic acidosis, or the salt-wastingsyndrome, are rare biochemical accompani-ments of chronic pyelonephritis.

TreatmentIn the treatment of primary uncomplicated

infections immediate good results can beobtained from most drugs, if the organism issensitive (Pryles, 1964), but there is a 10-20%failure rate no matter what chemotherapeuticis used. Reliable data for judging the claimsof one therapeutic regime over another are stillscant. The need to obtain adequate renal tissuelevels of drugs, as well as urinary levels, isbeing emphasised lately (Brumfitt, 1964),though the value of this has been questioned inthe face of considerable successes withmandelamine and furadantin which do notachieve tissue levels at all (Kleeman and others,1960). But for many gram negative organisms,the levels of antibiotic reached in the serum byconventional therapeutic dosage barely exceed,and seldom fully overreach, the minimuminhibitory concentration range (Brumfitt, 1964).The correlation between clinical effects of adrug in urinary infections and the results of invitro bacterial sensitivity tests is not close;clinicians do not always pay regard to them(Pryles, 1964; Spence and others, 1964). Thedisc method is unreliable; the tube dilutiontechnique is better, except for sulphonamidesfor whom "the patient himself is tihe bestsensitivity test." (Pryles, 1964). Aittention mustbe paid to pH adjustment of the urine in orderto get the best results from drugs (Brumfitt andPercival, 1962). This is especially important forstreDtomycin which needs an alkaline urine,and for tetracycline which is more effective withan acid urine. The naturally occurring anti-bacterial substances in urine are thought to bemore active at a low pH (Kleeman and others,1960). Minimum and optimum lengths oftreatment for acute infections have not beendecided upon. In a small group of infantsbetter results were obtained from a 6 months

August, 1965 493

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

494 POSTGRADUATE MEDICAL JOURNAL August, 1965

than a 2 weeks course of drugs (Stansfeld andWebb, 1954), but when acute pyelitis wastreated for 14 days in each month for a totalof two years, there was still a 7%/0 recurrencerate afterwards (Campanacci and others, 1963).In a controlled study of children with acutepyelonephritis, a relapse rate of 60% was foundafter 2 weeks treatment, 40% after 4 weeks,and 25% after 6 weeks (Spence and others,1964). In persistent infection there is provedeffectiveness in reducing recurrences from giv-ing continued treatment for 6 months, in pre-ference to short courses lasting 3 to 30 days(Deluca and others, 1963). If treatment inpersistent cases is going to work, the urine isalways sterile after 6 weeks. (Turck and others,1962). A sulphonamide drug is still the mostwidely used for first attacks, and furadantinfor chronic uncontrolled infections. Ampicillinis receiving much study and is the first choicefor Proteus and Enterococci (Naumann, 1964);many strains of E.Coli show in vitro resistanceto therapeutic levels of this drug, unless it isgiven parenterally (Brumfitt, 1964).

REFERENC(ESALLEN, R. P., and BURROWS, E. H. (1964): Micturat-

ing Cysto-urethrography in Investigation ofUrinary Tract Diseases in Children. Arch. Dis.Childh., 39, 95.

BRENNER, B. M., and GILBERT, V. E. (1963):Elevated Levels of Lactic Dehydrogenase GlutamicOxalacetic Transaminase and Catalase in InfectedUrine. Amer. J. med. Sci., 245, 31.

British Medical Journal. (1964): Pyelonephritis inChildhood, 'ii, 1215.

BRUMFIrr, W. (1964): An Assessment of Ampicillinin the Treatment of Urinary Tract Infection.Postgrad. med. J. Suppl., 40, 55.

BRUMFITr, W., DAVIES, B. I., and ROSSER, E.(1961): Urethral Catheter as a Cause of UrinaryTract Infection. Lancet, ii, 1059.

BRUMFITr, W., and PERCIVAL, A. (1962): Adjustmentof Urinary pH in the Chemotherapy of UrinaryTract Infections, Lancet, i, 186.

BRUMFITr, W., and PERCIVAL, A. (1964): Patho-genesis and Laboratory Diagnosis of Non-tuber-culous Urinary Tract Infection. J. Clin. Path., 17,482.

BRUMFITrI, W., PERCIVAL, A., and WILLIAMS, J. D.(1964): Problems in the Treatment of UrinaryTract Infection. Brit. J. clin. Pract., 18, 503.

BURKE, J. B. (1961): Pyelonephritis in Infancy andChildhood. Lancet, ii, 1116.

BURROWS, E. H., and ALLEN, R. P. (1964): UrethralLesions in Infancy and Childhood Studied byMicturition Cysto-Urethrography. Brit. J. Radiol.,37, 187.

CAMPONACCI, D., BoNoMINI, V., and ZUCHELLI, P.(1963): Chemotherapy in Acute and ChronicPyelonephritis, Lancet ii, 601.

CATrELL, W. R. and LEFFORD, M. J. (1963):Bacteriological Examination of Urine. Brit. med.J., i, 97.

CLARKE, S. H. C. (1960): Investigation into Methodsof Collecting Urine. Brit. med. J., ii, 1491.

DELUCA, F. G., FISHER, J. H., and SWENSON, 0.(1963): Review of Recurrent Urinary Infection inChildhood. New Engl. J. Med., 268, 75.

DODGE, E. A. (1963): Vesico-ureteric Reflux: Diag-nosis with Iodine 131, Lancet, i, 303.

DUNN, P. M., HINE, L., and MACGREGOR, M. E.(1964): Search by Clinical Methods for PersistentI Trinary Infection in Children. Brit. med. J.. i.1081.

EDWARDS, D. (1964): Cysto-urethrography in 'RecentAdvances in Radiology', edited T. Lodge, p. 179,London: J. & A. Churchill.

FORSYTHE, W. I., and WALLACE, I. R. (1958): TheInvestigation and Significance of Persistent andRecurrent Urinary Infections in Children. Brit. J.Urol., 30, 297.

GADEHOLT, H. (1964): Quantitative Estimation ofUrinary Sediment, with Special Regard to Sourcesof Error. Brit. med. J., i, 1547.

GARDBORG, 0. (1959): Bacteriological Findings andResistance to Antibiotics in Chronic and RecurringUrinary Tract Infections in Childhood. ActaPaediat. ,(Uppsala) Suppl., 118, 62.

GROSS, R. E., RANDOLPH, J., and WISE, H. M. Jnr.(1963): Surgical Correction of Bladder NeckObstruction. New Engl. J. Med., 268, 5.

GUTTMAN, D. and STOKES, E. J. (1963): Diagnosis ofUrinary Infection. Brit. med. J., i, 1383.

HANLEY, H. G. (1964): Pyelonephritis and AscendingInfection from the Lower Urinary Tract., J. Urol(Baltimore), 91, 1.

HODSON, C. J. (1959): Radiological Diagnosis ofPyelonephritis. Proc. roy. Soc. Med., 52, 669.

HODSON, C. J., DREWE, J. A., KARN, M. N., andKING, A. (1962): Renal Size in Normal Children.Arch. Dis. Childh., 37, 616.

HODSON, C. J., and EDWARDS, D. (1960): ChronicPyelonephritis and Vesico-ureteric Reflux. Clin.Radiol., 11, 219.

HODSON, C. J. (1964): The Urinary Tract in 'RecentAdvances in Radiology', edited T. Lodge, p. 156,London: J. & A. Churchill.

HOLLAND, N. H., and WEST, C. D. (1963): Preventionof Recurrent Urinary Tract Infection in Girls.Amer. J. Dis. Child., 105, 160.

HOUSTON, I. B. (1964): Pus Cell and Bacterial Countsin the Diagnosis of Urinary Tract Infe4tions inChildhood. Arch. Dis. Childh., 38, 600.

HuTH, M. S. R., CHALMERS, J. A., MACDONALD, J. S.,and DE WARDENER, H. E. (1961): Pyelonephritis.Lancet, i, 351.

JAMES, U. (1959): Urinary Infection in the Newborn.Lancet, ii, 1001.

JOHNSTON, J. H. (1963): Vesico-ureteric Reflux in theChild. Clin. Paediat, 2, 72.

JONES, N. F., and MILLS, I. H. (1964): ReversibleRenal Potassium Loss with Uxinary Tract Infec-tion. Amer. J. Med., 37, 305.

KASS, E. H. (1956): Asymptomatic Infections of theUrinary Tract. Trans. Ass. Amer. Phycns., 69, 56.

KASS, E. H. (1960): Biology of Pyelonephritis, editedQuinn, E. L., and Kass, E. -H. London: J. & A.Churchill.

KIMMMELSTIEL, P., KIM, 0. J., BERES, J. A., andWELLMAN, K. (1961): Chronic Pyelonephritis.Amer. J. Med., 30, 589.

KINCAID-SMITH, P. (1964): Ampicillin in Bacteriuriaand Pyelonephritis of Pregnancy. Postgrad. Med.J. Suppl. 40, 74.

KINCAID-SMITH, P., MCMICHAEL, J., and MURPHY,

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

August, 1965 MAcGREGOR: Pyelonephritis in Children 495

E. A. (1958): The Clinical Course and Pathologyof Hypertension with Papilloedema. Quart. J. Med.,27, 117.

KINCAID-SMITH, P., BULLEN, M., MILLS, J., FUSSELL,U., HUSTON, N., and GooN, F. (1964): The Reliabil-ity of Screening Tests for Bacteriuria in Pregnancy.Lancet, ii, 61.

KLEEMAN, C. R., HEWITr, W. L., and GAZE, L. B.(1960): Pyelonephritis. Medicine (Baltimore), 39, 3.

KNAPPENBERGER, S. T. (1963): Importance of UrethralDilatations in Treatment of Chronic Urinary TractInfections in Female Children. J. Urol (Baltimore),89, 95.

KUNIN, C. M., DEUTSCHER, R., and PAQUIN, A.(1964): Urinary Infection in School-children.Medicine (Baltimore), 43, 91.

KuNIN, C. M., SOUTHALL, I., and PAQUIN, A. J.(1962): Urinary Tract Infections in School-children. New Engl. J. Med., 266, 1287.

Lancet, (1962a): Obstructive Nephropathy, i, 31.Lancet, (1962b): Bacteriuria in School-girls, ii, 598.Lancet, (1963): Chemotherapy of Urinary Infections,

i, 148.LEATHER, H. M., WILLS, M. R., and GAULT, H. M.

(1963): Bacterial Pyrogen in the Diagnosis ofPyelonephritis. Brit. med. J., i, 92.

LEIGH, D. A., and WILLIAMS, J. D. (1964): Methodfor Detection of Significant Bacteriuria in LargeGroups of Patients. J. Clin. Path., 17, 498.

LINCOLN, K., and WINBERG, J. (1964): QuantitativeEstimation of Cellular Excretion in UnselectedNeonates. Acta Paediat. (Uppsala), 53, 447.

LINDMEYER, R. I., TURCK, M., and PETERSDORF, R. G.(1963): Factors Determining the Outcome ofChemotherapy in Infections of the Urinary Tract.Ann. intern. Med., 58, 201.

LITTLE, P. J. (1962): Urinary White Cell Excretion.Lancet, i, 1149.

LITTLE, P. J. (1964): A Comparison of Urinary WhiteCell Concentration with the White Cell ExcretionRate, Brit. J. Urol., 36, 360.

LITTLE, P. J., and DE WARDENER, H. E. (1 962): Useof Prednisolone Phosphate in the Diagnosis ofPyelonephritis in Man. Lancet, L, 1145.

MACAULAY, D., and SUTToN, R. N. P. (1957): ThePrognosis of Urinary Tract Infection in Child-hood. Lancet, ii, 1318.

MACAULAY, D. (1964): Recurrent Urinary Infectionin Girls. Lancet, ;;, 1319.

MCGRACKIE, J., and KENNEDY, A. C. (1963): Simpli-fied Quantitative Methods for Bacteriuria andPyuria, J. clin. Path., 16, 32.

MACKINNON, K. J. (1964): Ampicillin in Bacteriuriaand Pyelonephritis; discussion. Postgrad. med. J.,Suppl. 40, 79.

MCLEOD, J. W., MASON, J. M., and PILLEY, A. A.(1963): Prophylactic Control of Urinary Infectionfrom Catheterisation. Lancet, i, 292.

MARSHALL, F. C. (1962): Excretion UrographicChanges which Suggest Reflux, J. Urol. (Baltimore),87, 681.

MELICK, W. F., BRODEUR, A. E., and KARELLOS, D. N.(1962a): Abnormal Hydrodynamics of Bladder andUreter as a Cause of Reflux, J. Urol. (Baltimore),88, 38.

MONTGOMERIE, J. Z., and NORTH, J. D. K. (1963):Evaluation of Pyrogen Test in Chronic Pyelone-phritis. Lancet, i, 690.

MONZON, 0. T., ORY, E. M., DOBSON, H. L., CARTER,E., and Yow, E. M. (1958): A Comparison ofBacterial Counts of the Urine Obtained by Needle

Aspiration of the Bladder Catheterisation, and Mid-stream-voided Methods. New Engl. J. Med., 259,764.

NAUMANN, P. (1964): Treatment of Urinary TractInfections with Ampicillin, Postgrad. med. J.Suppl. 40, 62.

New England Journal of Medicine (1963): LowerUrinary Tract Obstruction in Children., 268, 52.

NEVIN, I. M., CLiNE, F. A., and HAUG, T. M. (1962):Forceful Ureteric Spurt: a Common RoentgenManifestation of Urinary Tract Infection in Child-ren. Radiology, 79, 933.

OSBORN, R. A., and SMITH, A. J. (1963): Comparisonof Quantitative Methods in the Investigation ofUrinary Tract Infection, J. clin. Path., 16, 46.

O'SULLIVAN, D. J., FITZGERALD, M. G., MEYNELL,M. J., and MALINS, J. M. (1960): A SimplifiedMethod for the Quantitative Bacterial Culture ofUrine, J. clin. Path., 13, 527.

PALKEN, M., and KENNELLY, J. M. (1960): RecurrentUrinary Tract Infection in Girls, J. Urol. (Balti-more) 83, 749.

PERCIVAL, A., BRUMFITr, W., and DE LouVoIs, J.1964): Serum Antibodies as an Indication ofClinically Inapparent Pyelonephritis. Lancet, ii,1027.

POIRIER, K. P., and JACKSON, G. G. (1957):Characteristics of Leucocytes in the Urine Sedi-ment in Pyelonephritis. Amer. J. Med., 23, 579.

PORTER, K. A. and GILES, H. MCC. (1956): APathological Study of Five Cases of Pyelonephritisin the Newborn. Arch. Dis. Childh., 31, 303.

PRESCOTT, L., and BRODIE, D. E. (1964): A SimpleDifferential Stain for Urinarv Sediment. Lancet ii,940.

PRYLES, C. V. (1960): The Diagnosis of UrinaryTract Infection. Pediatrics 26, 441.

PRYLES, C. V. (1964): Anti-microbial Therapy ofUrinary Tract Infections in 'Current PediatricTherapy'. Edited Gellis, S. S., and Kagan, B. M.Philadelphia: W.B. Saunders.

PRYLES, C. V., and NEUMANN, C. G. (1962):Pyelonephritis in Infants and Children. Amer. J.Dis. Child., 104, 215.

ROSENHEIM, M. L. (1963): Problems of ChronicPyelonephritis. Brit. med. J., i, 1433.

ScoTT, J. E. (1962): Experimental Asplects ofVesico-ureteric Reflux, Proc. roy. Soc. Med., 55,422.

SENECA, H. (1964): Current Therapy of Infections ofthe Renal Excretory System. J. Amer. Geriat. Soc.,12, 1100.

SIMMONS, N. H., and WILLIAMS, J. D. (1962): ASimple Test for Significant Bacteriuria. Lancet, i,1377.

SMALLPIECE, V. (1958): Prognosis of Urinary Infec-tions in Childhood. Lancet, i, 103.

SMELLIE, J. M., HODSON, C. J., EDWARDS, D., andNORMAND, C. S. (1964): Clinical and RadiologicalFeatures of Urinary Infection in Children. Brit.med. J., ii, 1222.

SPARK, H., TRAVIS, L. B., DODGE, W. F., DAESCHMER,C. W., and HoPPs, H. C. (1962): Prevalence ofPyelitis in Children at Autopsy. Pediatrics, 30,737.

SPENCE, H. M., MURPHY, J. J., McGOVERN, J. H.,HENDRON, W. H., and PRYLES, C. V. (1964):Urinary Tract Infections in Infants and Children.J. Urol. (Baltimore) 91, 623.

STANSFELD, J. M. (1954): Chronic Pyelonephritis inChildren. Proc. roy. Soc. Med., 47, 631.

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from

496 POSTGRADUATE MEDICAL JOURNAL August, 1965

STANSFELD, J. M. (1962): The Measurement andMeaning of Pyuria. Arch. Dis. Childh., 37, 257.

STANSFELD, J. M. (1963): Personal communication.STANSFELD, J. M., and WEBB, J. H. (1954): A Plea

for Longer Treatment of Chronic Pyelonphritis inChildren, Brit. med. J., i, 616.

STEPHENS, F. D., and LENAGHAN, D. (1962): Anatom-ical Basis and Dynamics of Ureteral Reflux, J.Urol. i(Baltimore), 87, 669.

STEELE, R. E., LEADBETTER, G., and CRAwFORD, J.(1963): Prognosis of Childhood Urinary TractInfection, New Engl. J. Med., 269, 883.

STERNHEIMER, R., and MALBIN, B. (1957): ClinicalRecognition of Pyelonephritis, with a New Stainfor Urinary Sediments, Amer. J. Med., 11, 312.

SWEET, A. Y., and WOLINSKY, E. (1964): An Outbreakof Urinary Tract and Other Infections due toE. Coli, Pediatrics., 33, 865.

TURCK, M., BROWDER, A., LINDMEYER, R. I., BROWN,N. K., and PETERSDORF, R. G. (1962): Failure ofProlonged Treatment in Chronic Urinary Infectionwith Antibiotics, New Engl. J. Med., 267, 999.

TURNER-WARWICK, R. T. (1962): The Clinical Prob-

lem and Investigation of Ureteric Reflux, Proc. roy.Soc. Med., 55, 419.

VERTANEN, S., OKSANEN, T., and PErONEN, T. (1962):Colony Count and Diagnosis of Urinary TractInfection in Infants and Children, Ann. Paediat.Fenn., 8, 269.

WEISS, S., and PAKER, F. (1939): Pyelonephritis: itsRelation to Vascular Lesions and to Arterial Hyper-tension, Medicine (Baltimore), 18, 221.

WILLIAMS, D. I. (1962): Reflux in Double Ureters,Proc. roy. Soc. Med., 55, 423.

WILLIAMS, D. I. (1963): Personal communication.WILLIAMS, D. I. (1964): Vesico-uretic Reflux in

'Recent Advances in Surgery', edited Taylor, S.P. 110. London: J. & A. Churchill,

WILLIAMS, D. I. (1965): Recurrent Urinary Infectionin Girls, Lancet, i, 109.

WILLIAMS, D. I., and STURDY D. E. 1(1961): RecurrentUrinary Infection in Girls, Arch. Dis. Childh., 36,130.

WOODRUFF, J. D., and EVERETr, H. S. (1954): Prog-nosis in Childhood Urinary Tract Infections inGirls, Amer. J. Obstet. Gynec., 68, 798.

copyright. on June 3, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.41.478.485 on 1 August 1965. D

ownloaded from