Retrospective study of children with renal scarring associated with reflux and urinary infection

Criteria for classification

Delay in diagnosis of urinary tract infection was recorded if no urine sample was cultured when there was a period of acute symptoms related to the urinary tract or high fever lasting for at least five days14; when there were ill defined symptoms (such as abdominal pain or screaming attacks) for at least one month; when there were one or more previous episodes with symptoms similar to those present when urinary tract infection was first diagnosed; or when there was delayed bladder control, defined as any daytime wetting at age 5 years or over or night time wetting at 7 years or over.

Delayed or inadequate treatment was recorded when the diagnosis had been delayed (when inevitably treatment was also delayed); more than four days had elapsed after collection of urine before the start of effective treatment14; or low dose antibacterial prophylaxis had not been prescribed either between treatment of the presenting infection and investigation or after the diagnosis of reflux.

Delayed or inadequate investigation was recorded when cystography was excluded from the initial investigation; cystography was delayed for more than one month in a child not receiving prophylaxis; adequate imaging of the kidneys was omitted from the initial investigation; investigation was not performed until after infection had recurred15; or a family history of vesicoureteric reflux or reflux nephropathy was disregarded.

These criteria were based on standard practices, recommendations,*RF 14-16* and experience.^5,8

Renal scarring was diagnosed by intravenous urography according to the standard criteria and classified radiologically as previously described.^2,17 In type A there were one or two scars only; type B more than two scarred areas; type C generalised parenchymal thinning with deformity of all papillae; and type D a small, poorly functioning, end stage kidney. Table I shows the types of scarring in pairs of kidneys in relation to sex. Type A scarring was more common in girls than boys, and seven boys and three girls had type C scarring. Impaired uptake of dimercaptosuccinic acid labelled with technetium-99m was also shown in all affected kidneys.

Diagnosis of urinary tract infection

Diagnosis was delayed in 41 of the 52 children, 21 boys and 20 girls (table II). The longest recorded delay was four years and 11 months. All 13 patients in whom symptoms started in the first six weeks of life had a delay in diagnosis ranging from seven days to two years. Symptoms in infancy were mainly non-specific, whereas in older children they were more often related to the urinary tract. Fever was the presenting symptom in 34; in two the symptoms were due to hypertension or renal failure. In three children with longstanding bed wetting, urinary infection was first diagnosed by the school doctor at ages 5, 7, and 9 years, when investigation showed bilateral renal scarring. No delay in diagnosis was recorded in 11 children, three boys and eight girls.

Treatment of urinary tract infection

Appropriate treatment, including the prescription of prophylaxis, was delayed in 49 children. In 45 of them effective treatment of the presenting infection was delayed (table II). In 41 this was secondary to delay in diagnosis. In five of these 41 and two whose condition was diagnosed immediately there was evidence that treatment was not instituted for between six days and eight months. In two others the prescribed antibiotic failed to clear the infection; one progressed to septicaemia, and the infection persisted for three weeks in the other. Thus treatment was delayed or ineffective in at least nine children known to have urinary infection at the time.

Prophylaxis - Low dose antibacterial prophylaxis was prescribed immediately after treatment of the acute infection in 15 children. In nine others it was started four to 12 months later, after one or more recurrences of infection and before investigation. In 28 children (in all of whom infection recurred) no prophylaxis was prescribed until after reflux had been diagnosed (more than a month later in 25 of them). Even after reflux had been identified there was further delay in prescribing prophylaxis of two to four months in three of them and more than one year in four others.

Recurrence of infection - In 31 children there was a clear history of recurrent symptoms or persistent day or night wetting before diagnosis. After confirmation of urinary infection, 37 of the 52 children had one or more further infections before starting prophylaxis.

Investigation

After the infection had been diagnosed imaging was delayed or inappropriate in 33 of the 52 cases reviewed.

Interval to diagnosis of vesicoureteric reflux - In 15 children, three without prophylaxis, reflux was identified within one month of diagnosis of the infection. In 37, 25 without prophylaxis (table II), cystography was performed two months to six years and seven months after diagnosis of the infection. Reflux was diagnosed in the first six weeks of life in four infants, by the age of 2 years in a further 21, between 2 and 5 years in 13, and aged 5 years or over in 14. In 26 children ultrasonography was the initial investigation. Either dilatation or hydronephrosis was noted in 12, but in 14 (five of whom with symptoms that had started in the first 6 weeks no ureteric dilatation or renal abnormality was reported.

Antenatal ultrasonography, family history, and new scar development - The single abnormal result of ultrasonography of the fetal renal tract found in the 32 mothers who underwent antenatal ultrasonography was reported as urinary tract dilatation only. Four mothers were known to have reflux nephropathy. Four of their five children had bilateral reflux and bilateral renal scarring when first investigated aged 6 weeks to 7 years, after urinary tract infection (the fifth, not included in this study and with no history of infection, had bilateral reflux and normal kidneys). Ten children were reported to have structurally normal kidneys on their initial intravenous urograms (confirmed retrospectively in eight of them). In five others only one kidney was scarred initially, though subsequently both kidneys were affected. In the 37 other children the first intravenous urogram already showed bilateral scarring.

Relation between scar type and delay in diagnosis of urinary infection - All the children included in the study had bilateral renal scarring (table III). Among the 11 children without any delay in diagnosis the types of scars seen were AA (two), AB (seven), AC (one), and BD (one). There was up to six months' delay in diagnosis in 24 children, 13 of whom had severe scar types (AD, BB, BC, BD, or CC). Among the 17 children in whom there was delay of more than six months 11 had more severe combinations of scar types. There was a significant relation in the type of renal scarring between no delay and delay of more than six months (X²=6.32; P=0.01, Yates's correction) and also a significant trend (X²=7.43; df=1; P=0.01, Yates's correction; see table III).

Detection and treatment of reflux

The symptoms of urinary infection are well known to be diverse and often non-specific, particularly in infants and young children, in whom the growing kidney is most susceptible to infective damage and when incompetence of the vesicoureteric valve is more likely to be present. Failure to consider the diagnosis was clearly an important factor contributing to the delay in this series. An alternative clinical diagnosis such as a viral illness had sometimes been made, and, particularly in young children, urine collection was often not attempted. In some instances there was delay in consulting the doctor - for example, in families who moved house frequently or because of language difficulties, family illness, or stress. Only two thirds of the children presented with a febrile illness. Similar delays in diagnosis were also encountered by the South Bedfordshire General Practitioners Group18 and by Jadresic et al.19 In our series there was a significant relation between the delay interval from onset of symptoms to diagnosis of urinary infection and the severity of renal scarring.

The exact timing of the start of treatment was not always easy to obtain retrospectively, but it seemed that microbiological confirmation after several days was usually awaited before treatment was started. (The use of nitrite sticks, urinoscopy, or microscopy would enable rapid preliminary diagnosis.14 The records conveyed little sense of urgency about the children's management, although the importance of delay in starting treatment as a factor in the acquisition of scars has been emphasised previously.*RF 3-5* For example, Winberg et al found that the incidence of scarring was four times greater in a group of girls with delayed treatment than in a comparable group treated immediately.3 There is also strong supportive evidence from animal studies.*RF 11- 13*

The prescription of prophylactic antibacterial treatment after a urinary infection and until the cause has been identified has been recommended in the guidelines published by the Royal College of Physicians.14 This was done in only 15 of the 52 children. Even in the 29 in whom prophylactic treatment was delayed until after the diagnosis of reflux there was a further delay in its introduction in seven.

Prevention of scarring

In a later study by the South Bedfordshire General Practice Group 11 out of 23 children with renal scarring in their practices received suboptimal management, in that investigation was delayed.20 In our study the most important imaging delay was in the diagnosis of reflux; in 37 (71%) three quarters of them without prophylaxis, the delay was over two months and up to six and a half years. It was unfortunate that 14 of the 25 in whom initial ultrasonography was performed were incorrectly reassured because no abnormality of the renal tract was reported. The main purpose of investigating children with urinary infection is to identify those with renal abnormalities or at risk of developing them. Views suggesting that a cystogram may not be necessary or that ultrasonography is a sufficient examination to exclude renal scarring clearly need revision.

Ideally reflux should be recognised before infection occurs and a search in siblings and offspring of index patients with severe reflux or reflux nephropathy offers an opportunity to do this^21,22 - an opportunity that was missed in the four families with known maternal reflux nephropathy in our study. The discovery of reflux or reflux nephropathy in the five children of these four mothers supports the findings of Bailey et al21 and Aggarwal and Verrier-Jones,22 and the strong recommendations of Bailey et al that infants born to parents with reflux nephropathy or severe reflux or who have affected siblings should be investigated as soon after birth as possible. At least 14 other children in our study had siblings with vesicoureteric reflux, details of whom will be reported separately.

Whether some or all of the scarring in the children in this study might have been prevented or limited can only be a matter for speculation in a retrospective study. Some of the children probably had abnormal kidneys at birth, but on the initial intravenous urogram of 13 children one or both kidneys were structurally normal, bilateral scarring developing subsequently. Surprisingly an abnormality of the fetal renal tract was reported in only one of the 32 mothers undergoing antenatal ultrasonography, repeated in some as late as 32 weeks' gestation. Vesicoureteric reflux in infancy, however, is not always preceded by dilatation of the fetal urinary tract detected on antenatal ultrasonography,23 and Gordon et al have shown that renal scars can be acquired early in infancy after infection of a refluxing urinary tract.24