Vesicoureteral reflux (VUR) is a common abnormality affecting 1% - 2% of the pediatric population [1] . The association between VUR and urinary tract infection (UTI) is well known, with risk for scarring of the renal parenchyma. Several studies have shown familial clustering of VUR, implying that genetic factors play an important role in its pathogenesis. For example, siblings of children with VUR are at higher risk of reflux than the general population, with reported prevalences between 27% and 51% [2] -[4] . Among multiple gestation births the concordance is higher in monozygotic twins than in dizygotic ones (80% vs. 35%) [5] . Moreover, the risk for offspring of parents with previously diagnosed VUR of having VUR themselves is reported to be 66% [6] .

Although subjects with VUR may present with similar initial symptoms, some cases have a benign natural course with no recurrent UTIs, no progressive renal damage and a high rate of spontaneous resolution of the reflux. Others, on the other hand, have recurrent UTIs, deterioration of renal status and persistent reflux [7] . These two phenotypes may have different etiological and genetic backgrounds. Therefore, it is of great interest to establish whether this previously described inherited (familial) form of VUR represents the same disease as the sporadic cases (when only one individual in the family has VUR) or whether it has a more aggressive course. In order to answer this question we assessed clinical features including severity of reflux, frequency of recurrent infections, renal damage, overall kidney function, and natural course in familial VUR compared with sporadic reflux cases. The aim of this study was to increase the understanding of the etiology of VUR. This knowledge can be further used to predict prognosis and individually tailor appropriate treatment.

Figure 1 illustrates the data selection process for the study. Seven hundred and twenty-six children with VUR

treated at Queen Silvia’s Children Hospital (a tertiary referral center) in Gothenburg, Sweden between 1990 and 2004 were identified through medical records. Letters were sent to all but 13 index cases, inquiring as to whether there were other members of the family or close relatives with VUR. The 99 cases that reported more than one individual with VUR were interviewed by telephone. During this conversation we inquired as to which members of the family had diagnosed VUR or had symptoms indicating such a problem, i.e. recurrent UTIs, bladder function symptoms or kidney problems. We also asked about possible consanguinity in the family. In this study, familial VUR cases were defined as patients with one or more first, second or third degree relatives with VUR. These affected relatives were analyzed as a separate group and compared with index and non-fami- lial cases. Patients with VUR secondary to urethral valves, myelomeningocele or high anal atresia with neurogenic bladder were excluded from the study.

For inclusion of family members in the group of relatives with VUR, a previously performed voiding cystourethrography (VCU) showing reflux was mandatory, except for older relatives diagnosed in an era when VCU was not in general use. A history of recurrent febrile UTIs during childhood with or without renal damage suggesting high probability for VUR was accepted in these cases. Relatives with radiologically confirmed diagnosis were marked as certain VUR cases in the pedigrees (n = 55), while cases with a strong history but with no radiological investigations available were labelled probable affected cases (n = 49) (Figure 2). Fifty-seven index patients who reported other members of the family with vague history of diffuse urinary tract problems, such as cystitis during adulthood, were excluded.

Owing to the fact that some individuals with VUR in our material belonged to the same family, the total number of families with hereditary reflux was ultimately 66. These 66 families were invited to participate in our study, and informed consent from patients and family members was obtained to examine medical records regarding their VUR and kidney status. The families were also asked to provide blood samples for a genetic study [8] .

To clarify the relationship and analyse the pattern of inheritance, pedigrees were constructed for each family. VUR in all affected individuals was characterised regarding presenting symptoms, grade of reflux, recurrent UTIs, kidney damage, natural course of the reflux, surgical treatment and additional anomalies. Data was also collected from medical records on subjects who did not report any relatives with VUR (controls), a total of 361 individuals. In these patients we recorded presenting symptoms, grade of reflux at presentation, recurrent UTIs (yes or no) and renal damage.

When recording the recurrent UTIs, only the febrile infections (>38.5˚C) with positive urinary culture (at least 100,000 cfu/ml) were counted in the hereditary groups. In the group of controls such data about symptoms and urinary culture were not always available.

In the characterisation of the subjects, grade of reflux was registered from VCU investigations according to the international grading system, both at presentation and at the latest follow-up before eventual surgical treatment [9] . In case of bilateral VUR, the patient was classified by the more severe side. Resolution of VUR was defined as grade II or less. DMSA and MAG-3 scintigraphy results were used for evaluation of renal damage, including split function, parenchymal abnormalities and size of kidneys. Focal damage was defined as one or more areas with reduced uptake or indentation of the renal outline. Generalised damage was classified as a small kidney with reduced tracer uptake or diffuse parenchymal anomaly. Reference values for split function were between 45% and 55%. Usually more than one scintigraphy was performed during the follow-up time. In this study we present the results from the latest available investigations regarding renal abnormality.

Glomerular filtration rate (GFR) was assessed using 51Cr-EDTA. In the cases where no direct GFR measurement was available, indirect GFR estimation was performed according to the formula of Schwartz [10] from serum creatinine and the height (cm) of the patient. The reference value used for normal GFR was 110 mmol/l/ 1.73 m² after two years of age according to Bröchner-Mortensen [11] and GFR < 80% (<2SD) of expected GFR was considered subnormal. Before two years of age, steady state of GFR is not achieved, and for this age group we operationally used the equation developed by Winberg for estimation of expected clearance [12] . Deterioration of renal status at follow up was defined as loss of ³7% of split function seen on scintigram or decrease of GFR with ³12%.

Statistical methods. In the statistical analysis of VUR phenotype in the group of relatives, only the 55 cases (of 104) were included, i.e. those who had performed VCU, whereas in the index and control groups VCU was performed in all cases and thus included. For categorical variables n (%) is presented, and for continuous variables Median (Min-Max). For comparisons between groups the chi-square test was used for non-ordered categorical variables, while Kruskal-Wallis test was used for continuous variables. For pairwise comparison be-

tween groups, Fisher’s exact test was used for dichotomous variables, Mantel-Haenszel chi-square test was used for ordered categorical variables, chi-square test was used for non-ordered categorical variables and Mann- Whitney U-test for continuous variables. All tests were 2-tailed and conducted at the 5% significance level.

Approval from the Regional Ethical Review Board in Gothenburg was obtained