The child with urinary tract infection : a dilemma for the paediatrician

Professor C C de Silva was born on the 25th of February 1904. He received his primary and secondary education at St Thomas College, Mount Lavinia. After one year at the Ceylon Medical College, he proceeded to the United Kingdom to complete his medical degree. He wasted no time in passing the MRCP and MD examinations. He later settled down to a general practice in Colombo. In 1949 he was appointed to the newly created Chair of Paediatrics of the then University of Ceylon. He accepted the challenge of organizing and developing the new Department of Paediatrics from its inception. He pioneered and organized undergraduate teaching in Sri Lanka. He was also responsible for the introduction of post-graduate training and the initiation of a DCH programme.


Introduction
Urinary tract infections (UTIs) are some of the most common and serious infections of childhood, particularly in the early years of life.Apart from ___________________________________________ 1 Senior Lecturer and Head of Department, Faculty of Medicine, University of Peradeniya.causing a potentially serious acute illness, UTIs can damage the kidneys and cause renal scarring, leading to hypertension and reduced renal function in later life 1,2 .Interestingly, UTIs in the pre-antibiotic era were primarily treated by disinfecting the colon.The scientific rationale of this therapeutic approach was the suspicion that bowel flora was responsible for causing UTIs.This is true even in the 21st century.Sterilization of the bowel was performed by using large enemata and calomel.Moreover, routine appendicectomy was performed if there was any suspicion of involvement of the appendix.If cystitis was suspected, washing out of the bladder was performed.In cases of pyelonephritis, urine was alkalinized by using large doses of citrate of potash.In more chronic cases they recommended use of a vaccine prepared from the urine of the patient.In resistant cases the last option was to improve the child's general health with a change in environment 3 .The acute stage morbidity and mortality of patients with UTIs were significantly high during this period.In 1931, John Thompson reported 3 deaths out of 24 cases of pyelitis in infancy.Jeffreys reported 6 deaths out of 60 cases, in all ages, treated at Great Ormond Street Children's Hospital, London.
With the discovery of antibiotics in the early 1940's, there was a drastic reduction in the acute morbidity and mortality of childhood UTIs.Presently, the major concerns of the medical fraternity are focused on the long term effects of UTI 4 .Recent technological advances in imaging techniques, such as ultrasonography, x-rays and radioisotope scanning, have highlighted the long -term effects of UTI 5 .There are many case reports that highlight new scar formation following UTIs detected using both intravenous urography and DMSA scans 6,7 .Several independent groups have confirmed link between UTI, vesicoureteric reflux (VUR) and renal scarring 4,8 .Smellie et al have also demonstrated a link between delay in diagnosis and treatment and the development of new renal scars.The risk of developing UTI before the age of 14 years is approximately 1% in boys and 3-5% in girls 2 .Correct diagnosis and management of UTIs is of utmost importance in order to prevent potential renal parenchymal damage that can lead to chronic renal failure in later life 4,9 .
VUR is a risk factor contributing to the development of UTIs and is present in a third of young children with UTI 10,11 .It is also a risk factor for renal scarring, with the propensity to progress towards reflux nephropathy 12,13 .It is also associated with renal dysplasia and other developmental abnormalities of the urinary tract including detrusor instability.Evidence suggests that an autosomal dominant mechanism is responsible for causing reflux and renal scarring in certain children 14 .With increase in knowledge, a consensus statement was produced in 1991 by the Royal College of Physicians to detect and manage UTIs in children 15 .Paediatricians in many countries have followed these guidelines.I became particularly interested in the management of childhood UTI in the early 1990's.
Although UTI is a common problem, lack of awareness of the presenting features, limited access to laboratory facilities and indiscriminate use of antimicrobial therapy may have resulted in continuing infections leading to scarring of kidneys.These effects are witnessed by adult nephrologists who see many progressing to end stage renal failure today.Establishing new nephrology units to serve increasing numbers of adult patients seeking treatment for renal failure, long waiting lists for costly procedure of renal transplantation plus frequent paper advertisements appealing for financial assistance may reflect our poor practice in the past.This evening I wish to present to you the findings of a prospective study which was undertaken at the Paediatric Unit, Peradeniya, from 1993 onwards in order to evaluate the impact of UTIs in the acute stage and its long-term effects.
The objective of the study was to establish data for UTIs in regard to presenting features, usefulness of urine microscopy in the diagnosis, antibiotic sensitivity pattern of the causative organisms, pattern of underlying urinary tract anomalies, predisposing factors for renal scarring and long-term outcome.

Patients and Method
Patients admitted to the paediatric ward and patients referred to the renal clinic with bacteriologically confirmed UTIs were studied.Their details were entered in a computer database and followed up in a special weekly clinic under personal supervision.Parents and guardians of children were educated about the disease and its long-term management.An educational booklet was provided free.
Infection of the urinary tract was identified by the growth of a significant number of organisms of a single species in a properly collected sample of urine, in the presence of symptoms.It was difficult to differentiate clinically between upper and lower UTIs.Thus, all febrile infections were considered to involve the upper urinary tract with potential for scarring.It was necessary to have a high degree of suspicion since their presenting features were nonspecific.All infants and children presenting with unexplained fever and neonates with septicaemia were suspected to have a UTI.Infants and young children who presented with recurrent fever, prolonged diarrhoea, vomiting, abdominal pain, poor weight gain and children having urinary symptoms were investigated to exclude UTI.Confirmation of the diagnosis was made on positive urine cultures in properly collected samples of urine.Almost always attempts were made to collect 2 samples of urine before instituting antimicrobial therapy.Moreover, clear instructions were provided to parents and nursing staff regarding method of collection and transport of specimens in order to minimise the chances of contamination.Urinalysis was helpful in providing immediate information and helped initiation of treatment.Clean catch midstream specimens of urine were collected in most cases after washing the genitalia with soap and water.Specimens were collected into sterile, screw capped glass bottles.In neonates and infants, urine samples were obtained by suprapubic aspiration or by clean catch specimens.Collection of specimens into a sterile bag strapped to the perineum was discouraged as risk of contamination was high.Urine cultures were repeated if contamination was suspected and if colony counts were equivocal.Definitions are given in Table 1.

Treatment
The patient's age, general condition and likelihood of compliance with medication helped in deciding between outpatient therapy and hospitalization.Antibiotics were given parenterally to infants and children with complicated infections (Table 2).
Treatment was initiated using a second or a third generation cephalosporin or co-amoxyclav until culture reports were available.Change over to an oral antibiotic was done if the child showed clinical improvement.Babies under 3 months and children with a positive blood culture were treated with parenteral antibiotics for the entire duration.Children above 3 months with a simple infection were treated with oral antibiotics (Table 2).Cephalosporins and co-amoxyclav that reach the renal parenchyma and blood stream in adequate therapeutic concentrations were selected.Nitrofurantoin and nalidixic acid were not considered since they did not reach adequate tissue levels.Repeat urine culture and ultrasound scan of the urinary tract were performed in children who did not show an improvement within 48 hours.
Treatment was given for 10-14 days for children with complicated infections and 7-10 days for simple infections.All the children were investigated to identify underlying anomalies leading to recurrent infections and renal damage.Guidelines for evaluation of patients varied with age (Figure 1).DTPA scans were performed in children with evidence of hydronephrosis in the absence of reflux.This was useful in assessing the function of kidneys and drainage of the collecting system.In selected cases, X-rays of the kidneys and the bladder area were taken to exclude radio-opaque calculi.Cystoscopy and intravenous pyelograms were rarely performed.Urine cultures were routinely repeated at cessation of antibiotic therapy.A detailed analysis was made of all the urine culture reports between two yearly periods for a comparative analysis.Prophylactic antibiotics were given to all children with UTIs until the imaging studies were completed.
If there was a reflux or renal scarring prophylaxis was continued until five years.Children with recurrent infections were also given prophylactic antibiotics for a period of two years even if the urinary tract was normal.Nitrofurantoin, nalidixic acid, cephalexin and cotrimoxazole were used as prophylactic drugs and given as a single bedtime dose (Table 3).Plenty of fluids were encouraged.Frequent voiding and the practice of double micturition was recommended to minimise stasis of urine in the bladder.High fibre diet was encouraged.Attention to undergarments and perineal hygiene, including foreskin of boys, was explained to parents.Although circumcision was not routinely done it was considered in children having a reflux with recurrent infections.Asymptomatic bacteriuria is a benign condition.Organisms commonly identified are usually low virulent E. coli.In many instances eradication of low virulent bacteria is followed by symptomatic infection with more virulent bacterial strains.Hence antibiotic therapy was not prescribed.
Among children with UTIs, VUR is seen in about 30-50%, while in infants it is 40-50%.Presence of moderate to severe reflux is known to predispose towards reflux nephropathy.International reflux study classification helps to grade reflux from I-V on severity 17 .Basis for classification is the appearance of the urinary tract in micturating cystourethrogram (MCUG).Children with high grade reflux and concomitant infections are at a higher risk for renal scarring.Reflux generally resolves or improves with bladder growth and maturation 18,19 .Lesser grades of reflux are more likely to resolve than higher grades.A majority of children were managed on long term antibiotics with the expectation of spontaneous resolution of reflux.All patients were closely monitored for breakthrough infections.The prophylactic antibiotics were discontinued in children above five years despite the presence of mild to moderate reflux.According to published data risk of recurrent infections and renal scarring are low after four to five years 20 .Surgical repair was considered in patients with Grade V reflux or if breakthrough infections developed in patients with Grade IV reflux.After surgery, prophylactic antibiotics were continued for six months.Treatment of primary VUR is outlined in Table 4. Results

21
Over a period of ten years from 1993 onwards 808 children with confirmed UTIs were registered in our clinic.447 (55%) were males and 361 (45%) were females.Male: female ratio was 1:2.672 (83%) of them were below the age of 5 years.There were more males than females below one year of age.Sex ratio was almost equal in older children (Figure 2).Fever was the main presenting feature in 81%.It is interesting to note that only around one third of children presented with symptoms relating to the urinary tract (Table 5).Ultrasound scans were performed in all 808 children, of which 26% showed abnormalities.41.6% of patients were subjected to MCUG of which 30% were abnormal.48% of them had undergone DMSA scans of which 24% showed scarring of kidneys.Abnormalities were seen in 24.6% of DTPA scans performed in 16% (Table 7).Analysis of children with renal scars 23 22.6% of children in the study group had detectable renal scars on DMSA scans.52% of them were males and 48% were females.76% had scarring of upper poles of kidneys.24% of them had multiple scars with 7% having bilateral involvement (Table 9).Over 70% were below 5 years and 42% were below 2 years (Figure 3).92% of children were detected during their first presentation.Only 66% presented with specific symptoms.45% of children with renal scars had underlying abnormalities, and the main abnormality detected was reflux in 72% (Table 10).

Figure 2 Age and sex distribution
E.coli was the commonest organism cultured from 95%.Long term follow up study of children with vesicoureteric reflux 24 Fifty nine patients with initial reflux confirmed by MCUG were studied.They were followed up for 6-12 years with the average being 8 years.Male: female ratio was 1.6:1 and mean age at presentation was 1.9 years.Their long-term outcome was assessed in regard to occurrence of UTIs during and after discontinuation of prophylaxis, adverse effects to prophylactic drugs, clinical and laboratory evidence of renal failure, such as blood pressure measurements, physical growth parameters and proteinuria.A majority had Grade III VUR (Figure 4).Reflux was secondary to posterior urethral valves in 3%.52% of children with reflux had detectable renal scars.15% of this group underwent surgery, of which over 75% were due to anomalies other than reflux such as phimosis, posterior urethral valves and hypospadias.Blood pressure measurements were taken at the end of the follow up period (Table 11).One patient was on long-term antihypertensives.All other patients recorded systolic blood pressures below 95th percentile for age.Urine was free of proteins in a majority.3 patients had laboratory evidence of established renal failure.Two of them had presented before the age of two years with posterior urethral valves and are awaiting renal transplant.The other patient whose renal failure was due to a neurogenic bladder underwent a renal transplant at the Teaching Hospital, Peradeniya, with the help of Dr. Oswald Fernando from the UK.This is the first paediatric renal transplant done in our country.

Discussion
In this study 808 children with culture positive UTIs were followed up over a ten year period.Special emphasis was given to the accuracy of initial diagnosis.In spite of stringent entry criteria, the large number of cases studied illustrates how common UTIs are in paediatric practice.More than 75% in this study were below five years.Similar observations have been made locally by Lamabadusuriya 26 , de Silva et al 27 and David 28 who showed that more than 55% of children in their studies were below five years.This indicates that young children are more at risk.This study group shows a male preponderance below one year, which is in keeping with the figures from the West.Lamabadusuriya in his study of 165 patients followed up in private sector hospitals and de Silva et al in their study made similar observations.Fever was the commonest presenting feature in 70%.Only 30% presented with symptoms relating to the urinary tract.Observations made by Lamabadusuriya, de Silva and David in their studies are similar.Therefore, in children presenting with fever without associated symptoms, a high degree of suspicion is necessary for diagnosis.27% did not show any significant microscopic abnormalities indicating that the "gold standard" for establishment of diagnosis is a positive urine culture.E. coli was the commonest organism isolated from 95% in our study.
Co-amoxyclav, which had the highest level of sensitivity, along with cephalosporins, remain the most suitable tissue penetrating antibiotics for treatment of acute episodes.Nalidixic acid and nitrofurantoin have retained their high levels of sensitivities.These antibiotics therefore remain suitable for the treatment of lower UTIs and for prophylaxis.However, ciprofloxacin and mecillinam had low levels of sensitivities indicating that resistance had built up against these two drugs.Therefore prescribing of these two drugs is not recommended.Gentamicin, which had a low level of sensitivity in 1997, had attained a somewhat acceptable sensitivity level in 2002.Use of gentamicin for the treatment of acute episodes has diminished over the last few years due to nephrotoxicity and availability of safer alternatives such as cephalosporins.This decrease in usage is the most likely cause for its increased sensitivity in 2002.
The change in sensitivity pattern of cotrimoxazole over the 5-year period was noteworthy.This emphasizes the importance of periodic revision of antibiotic sensitivity patterns to enable appropriate treatment.
The study shows that ultrasound scan was not sensitive in detecting VUR.Gunaratna et al 29  Over the years evidence confirms that renal scarring is an unlikely complication of UTIs in patients over 5 years.Hence the international consensus is to carry out minimal imaging for this group 20,32 .However, in the Sri Lankan context with limited diagnostic facilities, it may be prudent to carry out a DMSA scan to detect renal scarring even in patients presenting above 5 years.The presence of renal scars can predispose the child to hypertension and chronic renal failure 31 .Our patients with reflux were managed conservatively with regular monitoring and administration of prophylactic antibiotics except in 6 patients.Complications such as hypertension and progression to end stage renal failure were rare.These findings reflect recent reports of excellent outcome in children with reflux managed conservatively.
Due to concern of reflux nephropathy, the Royal College of Physicians published guidelines in 1991 recommending imaging in every child after the first UTI 15 .The working group advised all patients to have an ultrasound scan and that the DMSA scan should be carried out on those below 7 years.MCUG was considered mandatory in children below 1 year.
Without good evidence of benefits from treatment plus poor compliance to long-term low dose prophylaxis, the presumed benefits of imaging and long-term treatment may not be as good as originally expected.Stark 33 , in 1997, challenged the validity of imaging tests and said they are excessively costly and invasive without any benefit.
Tertiary specialists were heavily represented in the research-working group of the Royal College of Physicians.Their views and experiences represented the consensus on UTIs and reflux nephropathy.In contrast, general practitioners and paediatricians, who see most children during their first UTI, were in the minority.However, published guidelines provided a useful step in evaluating current practice.These guidelines are followed extensively in UK and Sri Lanka.However, ironically and most unfortunately, studies suggest that many children do not have the recommended management.
Although there was no formal attempt to inform general practitioners of the guidelines, some general practitioners are now referring large numbers of children for imaging investigations and a paediatric opinion following simple, non-febrile UTI.This has generated a massive workload for radiology departments and exposed large numbers of children to significant radiation without much evidence of benefit.The proportion of children seen with evidence of renal damage is lower than in earlier studies.This may be because UTI is diagnosed early in infants when they are referred to hospital, or it may reflect the fact that many more straightforward cases are being referred for further investigation.Unfortunately, there is continuing evidence of delay in diagnosis of UTI in infants and toddlers in primary care where urine collection is perceived as difficult and some children have several consultations before the diagnosis of UTI is considered and further delay before it is confirmed.
In conclusion, it is now clear that evidence based recommendations, even for the developed countries have yet to be reached.The situation is worse for developing countries where the facilities are not readily available.This poses a therapeutic dilemma for the practising paediatrician.In Sri Lanka certain towns have collection centres for urine cultures.These are sent to established laboratories in Colombo.Accuracy of results is likely to suffer adversely during transport, as the diagnosis of UTI is heavily dependant on the results of the urine culture report.Thus the paediatrician is left in a dilemma with regard to the extent of imaging of the urinary tract.On one hand imaging is costly, invasive and involves irradiation of a vital area of the child which can be totally unnecessary if the urine culture is falsely positive.On the other hand lesser investigations may miss significant underlying anomalies.It is worth emphasising here that infants and young children with neglected UTIs may end up with renal failure.Hence it is a worthy asset to improve the microbiological infrastructure facilities in the lesser developed districts of Sri Lanka which would facilitate the proper diagnosis of children with UTIs, particularly the very young patients.Until such time children with suspected UTIs could be referred to centres where they could be properly investigated and managed, thus reducing the potential risk of end stage renal failure.
II, III Antibiotic prophylaxis up to 5 years of age Grade IV Antibiotic prophylaxis up to 5 years of age Surgery if breakthrough infections occur Grade V Antibiotic prophylaxis until surgery In all patients All children were followed up under personal supervision.Patients were monitored for their physical growth, renal function and blood pressure.Urine cultures were done promptly if they were symptomatic.Children with renal scars were given continuous care up to adulthood.Children without underlying abnormalities were given the opportunity to attend the clinic if they became symptomatic.All patients with reflux who have completed 5 years of follow up were reviewed and assessed for long-term outcome.

Figure 3 .
Figure 3. Age distribution of children with scars

Figure 4
Figure 4 Grading of VUR

Table 2
Figure 1 Evaluation following initial UTI.

Table 6
).A change was evident in sensitivity levels of cotrimoxazole and cephalexin.These had low sensitivities in 1997 and relatively higher sensitivities in 2002.In 1997 sensitivity tests were not performed for co-amoxyclav and cefradine.

Table 9
Degree of renal scarring