Results

During the study period 284 children fulfilled the entry criteria; 272 children (95.8%) had an MCU performed; 12 children (4.2%) did not participate because of parental refusal.

There were 149 boys (54.8%) and 123 girls (45.2%). The median age was 8.3 months (range 0.3 to 59.9 months). Vesicoureteric reflux was detected in 77 children (table 1). Bladder outlet obstruction secondary to posterior urethral valves was present in one child and intrarenal reflux was detected in seven (2.6%). No child developed a symptomatic iatrogenic urinary tract infection after the MCU.

The timing of the MCU, and the grade and presence of vesicoureteric reflux detected are given in table 2. The median time between diagnosis and the MCU was 29 days (interquartile range 23-34 days, range 5-167 days).

Overall, there was evidence of a weak association between timing of the MCU and grade of vesicoureteric reflux (table 2: χ² ₍₁₄₎ = 24.8, p = 0.04). This was not because of a trend towards increased detection and severity of reflux with decreasing time between diagnosis of urinary tract infection and the MCU (χ² ₍₁₎ = 1.3, p = 0.3). Rather, the positive association was because of two cases of grade III-V vesicoureteric reflux in children who had the MCU performed within the first week after diagnosis (Fisher’s exact test; p = 0.02).

Using logistic regression, after adjustment for factors that may have been associated with the presence of vesicoureteric reflux or may have influenced the timing of the MCU—age, gender, a history of fever, vomiting, lethargy, anorexia, diarrhoea, duration of illness, a past history of previous unexplained fever, recorded temperature on admission—there was no evidence of an association between the timing of the MCU after urinary tract infection and vesicoureteric reflux (χ² = 0.58, p = 0.5; odds ratio 0.93, 95% confidence interval 0.76 to 1.12).

Discussion

Despite the common nature of urinary tract infection and the frequency with which MCU is performed in children, we could not find any studies in children which have analysed the association between timing of the MCU after urinary tract infection and the detection and grade of vesicoureteric reflux. The frequency and grade of vesicoureteric reflux observed in this study was similar to previous studies.2-4 Beyond one week after the diagnosis of urinary tract infection there was no evidence of an association between the presence or grade of vesicoureteric reflux and timing of the MCU, even after adjustment for factors that may predict for reflux or clinical indicators that may have influenced the timing of the MCU. The children who had the MCU within one week of diagnosis, however, were more likely to show vesicoureteric reflux, and of a high grade, but because there were only two patients in the early MCU group, we should be cautious not to overinterpret this finding. There may have been other factors which prompted the early MCU in these children which would have made vesicoureteric reflux more likely, but which we could not adjust for in this study. These data may support the notion that acute inflammation does cause transient malfunction of the vesicoureteric junction.

MCU after urinary tract infection appears to be a very safe procedure: there were no episodes of symptomatic infection following the test. Concerns regarding ascending spread of infection may be unfounded as long as appropriate chemoprophylactic cover is provided.

The present study did not involve a random allocation of timing of the MCU and so there is a potential for bias. Apart from randomising the time of the MCU, an ideal study may have involved performing the test on each child at two time intervals after diagnosis—early and late—and comparing proportions in a paired manner. We did not think such a study was feasible because of the invasive nature of even a single MCU. Measurement error has been minimised by blinding the radiologists to the study hypothesis and to the time of infection. With only 12 patients (4.2%) excluded because of parental preference, it would be unlikely that selection bias would significantly alter these results.

The notion that UTI causes disturbance of the vesicoureteral junction and vesicoureteric reflux was suggested by animal experiments in the 1960s. These were primarily designed to test the hypothesis that vesicoureteric reflux was not a primary abnormality of the intramural portion of the ureter, but secondary to urinary infection. Chronic urinary infection was produced using a number of methods in dogs (inoculation of bacteria through nephrostomy tubes,11 12submucosal insertion of bacteria impregnated agar and bladder trauma,13 inoculation of bacteria into a renal artery,13 or periurethral inoculation with instillation of a foreign body into the bladder14) and reflux was subsequently demonstrated with variable success. Later, Roberts and Riopelle15 compared the rate of resolution of reflux in 16 infant female rhesus monkeys. Eight were instilled with coliform bacteria at the time of monthly cystography, and the control group were not. They concluded that vesicoureteric reflux took longer to resolve in the infected group, even though the mean time that reflux was present was not significantly different between the two groups. It is reasonably well established that the introduction of bacteria into the ureter can cause ureteritis and abnormal peristaltic activity in animals16; however, it remains uncertain whether transient urinary infection alone can cause vesicoureteric reflux in animals. Extrapolating these data to children who do not have artificial manipulation of their renal tracts and have transient infection does involve a number of unproven assumptions.

To our knowledge there has been only one clinical study published in this area. The study of Gross and Lebowitz17 addressed a related but slightly different question: are children with urinary tract infection at the time of the MCU more likely to have vesicoureteric reflux than those with sterile urine? They reviewed a case series of children who had an MCU performed within their unit over a one year period. The proportion of children with infected urine at the time of the MCU was no different in children with or without reflux (25/204 with reflux, 40/397 without reflux, χ² ₍₁₎ = 0.66, p = 0.4). Although the study does have limitations the data do suggest that active infection is not associated with vesicoureteric reflux.

Deferring the MCU for four to six weeks after urinary tract infection is widely recommended because of a proposed association bet- ween the timing of the MCU after the infection and the detection of vesicoureteric reflux in children. We could find no evidence of this association beyond one week after infection. An MCU within one week of infection may be more likely to demonstrate vesicoureteric reflux. Acute infective complications did not occur. Deferring the MCU for four to six weeks after infection carries inevitable cost—parents are worried about the presence of reflux and children are on daily antibiotics. This study suggests these costs are not outweighed by the unconfirmed concern for falsely detecting vesicoureteric reflux, provided the MCU is delayed one week after urinary tract infection. We conclude that the MCU can be arranged at any time after the first week of diagnosis according to convenience considerations rather than after four to six weeks, as is commonly recommended.