- Email: [email protected]
Medical versus surgical treatment in children with severe bilateral vesicoureteric reflux and bilateral nephropathy: a randomised trial
ARTICLES
Medical versus surgical treatment in children with severe bilateral vesicoureteric reflux and bilateral nephropathy: a randomised trial Jean M Smellie, T Martin Barratt, Cyril Chantler, Isky Gordon, Nina P Prescod, Philip G Ransley, Adrian S Woolf
Summary Background Nephropathy associated with vesicoureteric reflux (VUR) and urinary tract infection can result in end-stage renal failure, hypertension, or both. Whether long-term VUR contributes to these outcomes is unknown. We compared, in a randomised trial, medical with surgical management of children with bilateral severe VUR and bilateral nephropathy. Methods We stratified by age and glomerular filtration rate (GFR) 25 boys and 27 girls aged 1–12 years and randomly assigned them to medical or surgical management. At enrolment and 4 years’ follow-up we estimated GFR from the plasma clearance of 51Cr-labelled edetic acid (EDTA), and did intravenous urography. We also did a metastable 99mTc-labelled dimercaptosuccinic acid (DMSA) assay and contrast cystography. The change in GFR at 4 years, expressed as a percentage change between enrolment and 4 years, was available for 26 of 27 patients in the medical and 24 of 25 in the surgical group. We assessed GFR in 48 patients 10 years after enrolment. Findings Mean GFR at enrolment was 72·4 mL/min per 1·73 m2 (SD 24·1) in the medical and 71·7 mL/min per 1·73 m2 (22·6) in the surgical group. The mean percentage change in GFR at 4 years was ⫺2·4% (SE 4·5) versus 4·7% (5·0) in the medical and surgical groups, respectively. The difference in change in GFR at 4 years between the two groups was not significant (7·1%, 95% CI ⫺6·4% to 20·6%). Interpretation Our data do not lend support to the view that the outcome for renal function is improved by surgical correction of VUR in children with bilateral disease. Lancet 2001; 357: 1329–33 See Commentary page 1309
Introduction Vesicoureteric reflux (VUR) can be associated with other abnormalities of the bladder or lower urinary tract. 1–2% of apparently healthy children have this disorder, and it is present in 30–40% of children with urinary tract infection, of whom about 30% have radiological evidence of renal scarring. Such scarring is a coarse irregular nephropathy resulting from congenital dysplasia, acquired postinfectious damage, or both.1 This nephropathy is a major cause of severe hypertension in children and young adults. It occasionally progresses to chronic renal failure,2 which accounts for about 25% of UK children with end-stage renal failure, requiring regular dialysis or transplantation.3 Several factors might lead to this progression, including inadequate renal growth, further scarring, associated hypertension, and secondary glomerulosclerosis with proteinuria. Common practice has been to correct VUR surgically, on the assumption that this disorder predisposes to urinary tract infection and facilitates the spread of infection to the renal parenchyma. However, findings from comparative trials of medical and surgical management of children with severe VUR have shown no significant difference in renal growth or acquisition of new scars for 10 years.4–9 The optimum management of children with bilateral severe VUR and bilateral nephropathy, the group thought to have the highest risk of developing end-stage renal failure,2 is not known. We, therefore, undertook a randomised trial of medical management versus surgical correction in 53 such children, with a glomerular filtration rate (GFR) at enrolment down to 20 mL/min per 1·73 m2 body surface area. We aimed to assess whether the expected decline in renal function could be prevented by surgical correction of VUR. In this study we report the clinical, renal function, and imaging outcomes at 4 years, and the clinical outcome and GFR at 10 years.
Methods
Great Ormond Street Hospital for Children NHS Trust and Institute of Child Health, University College London, London WC1N 1EH, UK (J M Smellie FRCP, Prof T M Barratt FRCP, I Gordon FRCP, N P Prescod BSc, P G Ransley FRCS, Prof A S Woolf FRCPCH); and King’s College, Guy’s and St Thomas’ Hospitals’ Medical and Dental Schools (J M Smellie, Prof C Chantler FRCP, N P Prescod) Correspondence to: Prof T Martin Barratt (e-mail: [email protected])
THE LANCET • Vol 357 • April 28, 2001
Patients Children were eligible for enrolment if they were attending the nephrology or urology clinics at Great Ormond Street or Guy’s Hospitals, London, UK, between Nov 1, 1985, and Dec 31, 1989, and had bilateral severe primary VUR (International Reflux Study in Children grades III–V),10 bilateral renal scarring diagnosed on intravenous urography with corresponding defects in metastable 99mTc-labelled dimercaptosuccinic acid (DMSA) uptake, and a history of bacteriologically confirmed urinary tract infection. The criteria for enrolment, were: age from 1 to 12 years; bilateral severe VUR diagnosed by contrast micturating cystourethrography, GFR of 20 mL/min per 1·73 m2 or more, no current hypertension (or hypertension controlled by antihypertensive therapy for more than 6 months), and no current urinary tract infection. Children younger than 1 year were excluded from the trial because of the confounding effect of the normal increase in GFR at that age11 and of the high surgical complication rate. We excluded children with bladder outflow obstruction, overt neuropathic bladder, previous surgery to the uretero-vesical junction, a duplex system or one kidney, or one kidney contributing less than 10% to overall renal function.
1329
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Study design The children were taken from a large group of unknown size, who might have been suitable for the trial, and were referred by consultant nephrological and urological staff to an investigator (JMS) for review. 53 children fulfilled the criteria and parental written informed consent for the trial was obtained for every child. We stratified children by age (1–5 or 6–12 years) and GFR (20–39, 40–79, or ⭓80 mL/min per 1·73 m2). Envelopes for the six categories were prepared containing equal numbers of cards for each treatment arm. An investigator (NPP) took a concealed card from the appropriate envelope to randomly assign children to a group. Children in the medical management group took antibacterial prophylaxis with co-trimoxazole (sulfamethoxazole 5–10 mg per kg daily and trimethoprim 1–2 mg per kg daily), trimethoprim (1–2 mg per kg daily) alone, or nitrofurantoin (1–2 mg per kg daily). These children had double micturition at bedtime, and we carefully noted bladder and bowel habits. We treated those with bacteriologically confirmed urinary tract infection with appropriate antibacterial drugs. An investigator (JMS) supervised the children in special clinics every 3 months for the first 4 years and at longer intervals thereafter. Surgical correction of VUR was done as soon as possible after randomisation and completion of the preliminary investigations (median 0·19 years, range 0·01–0·79). Individual surgeons used their preferred technique, usually a Cohen ureteric advancement.12 Antibacterial prophylaxis was stopped for patients in the surgical group after 6 months if VUR was shown to have been corrected by micturating cystourethrography or indirect radionuclide cystogram. The trial protocol was approved by the ethics committees of both hospitals. Patient assessment Before the children were randomly assigned to groups we measured height, weight, and systolic blood pressure, with standard deviation scores (SDS) derived from published data.13–15 We also did urine microscopy and culture, and dipstick test for protein. We measured GFR and did contrast micturating cystourethrography, a renal 99mTcDMSA scan, and limited intravenous urography (one to two x-ray films only). These measurements were used as enrolment values if they had been done less than 6 months before randomisation. We did the DMSA scan, limited intravenous urography, cystogram, and GFR again about 4 years after initial measurements (GFR was measured between 3·7 and 4·9 years). We graded VUR between 0 and V, in accordance with International Reflux Study in Children Criteria.10 For patients in the surgical group we did a postoperative 99mTcdiethylenetriamine penta-acetic acid (DTPA) scan, with micturating cystourethrography or indirect radionuclide cystogram. These were only repeated if we saw VUR or evidence of uretero-vesical junction obstruction (in the latter years of follow-up we sometimes did an indirect radionuclide cystogram instead of micturating 99m 99m cystourethrography with Tc-DTPA or Tcmercaptoacetyltriglycine [MAG3]). We measured renal length with intravenous urography, related this measurement to height, and expressed the relation as SDS.16 We calculated renal growth as the difference between SDS at 4 years and enrolment. Growth of ⫺0·5 or more was satisfactory. Renal scarring on intravenous urography and 99mTc-DMSA image defects were independently assessed at entry and after completion of the trial by at least two investigators. Radiological scarring was typed from A to D: type A indicated 1 or 2 scars, type B
1330
more extensive scarring with some normal parenchyma, type C an irregularly thinned rim of parenchyma that surrounded deformed renal papillae and calyces, and type D a small deformed kidney.1 Renal function We estimated GFR from the plasma clearance of 51 Cr-edetic acid (EDTA) by single exponential analysis of blood samples at 2 and 4 h after injection17 and corrected it to 1·73 m2 of body surface area with the Dubois formula.18 The coefficient of variation of replicate estimates of GFR by this method in the same individual is 3·9%.17 The difference in each child between the GFR at 4 years (GFR4) and the initial GFR (GFR0) was expressed as a percentage of GFR0: ∆GFR4%=100⫻(GFR4⫺GFR0)/GFR0 We measured GFR about 10 years after entry (GFR10) in 44 children and calculated the change from GFR0 in the same way as at 4 years (∆GFR10%). In the four children who had reached end-stage renal failure, GFR was not measured but was deemed as being 5 mL/min per 1·73 m2 for statistical purposes. Departures from protocol (figure 1) We withdrew one child in the medical group because she was judged on review after randomisation, to have only unilateral renal damage, and was lost to follow-up after 2 years. Two children in the surgical group proved to have unilateral VUR: one was detected on subsequent review of micturating cystourethrography, and the other (who had previously had bilateral severe VUR) on cystography during surgery. These two children had only unilateral ureteric reimplantation and no reflux was seen on postoperative cystography. Both had bilateral nephropathy and were retained in the trial on the basis of intention to treat. We excluded one child in the surgical group from the 4-year analysis because the estimate of GFR at 4 years of 174 mL/min per 1·73 m2 was deemed erroneous because of a high volume of distribution of the 51Cr-EDTA with flanking values of 102 and 113. GFR at 4 years was not available for another child in the medical group because she had emigrated. Both were available for follow-up at 10 years. Statistical analysis The primary endpoint was a comparison of mean percentage change in GFR at 4 years in the medical and surgical groups. We calculated the difference between group means and 95% CI. Other endpoints were somatic growth, blood pressure, recurrence of urinary tract infections, renal growth, and additional renal scarring. We calculated between-group differences on other continuous variables and the 95% CI. For dichotomous variables, relative risk and 95% CI were calculated with EPI INFO version 6.04. Observed differences and relative risks were not regarded as significant if they fell within the 95% CI.
Results Entry data 52 children were followed up: 28 were randomly assigned to medical management (one child was withdrawn) and 25 to surgery (figure 1). There was a family history of VUR in 11 children. The antecedent histories of the patients have been published.19,20 16 children had a radiologically normal kidney (two normal kidneys in eight) before enrolment with bilateral disease. Two children, one in each group, had hypertension controlled by drugs. At enrolment there were no differences between the two groups at entry for sex, age, SDS for systolic blood pressure and height, or GFR (table 1). Table 2 shows VUR grade, scar type, and kidney length SDS in the 104 kidneys and ureters. A closely similar
THE LANCET • Vol 357 • April 28, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Medical (n=54)
Surgical (n=50)
Enrolment
4 years’ follow-up
Enrolment
4 years’ follow-up
VUR grade 0 I II III IV V
0 0 0 8 29 17
11 9 6 13 11 4
1 0 2 15 19 13
46 1 0 1 2 0
Renal scar type A B C D
16 25 8 5
16 21 8 9
14 26 6 4
14 24 6 6
Renal length SDS >⫺1·0 ⫺1·0 to⫺2·0 <⫺2·0 Missing data
38 9 3 4
30 13 6 5
32 10 6 2
32 8 6 4
Children aged 1–12 years with bilateral VUR and nephropathy and history of UTI Excluded if other urological abnormality GFR <20 mL/min per 1·73 m2 or uncontrolled hypertension 53 randomised
28 assigned to medical management
25 assigned to surgical management
1 withdrawn 1 not available for primary endpoint at 4 years
1 not available for primary endpoint at 4 years
26 analysed for primary endpoint
24 analysed for primary endpoint
Renal growth ∆SDS ⭓⫺0·5 (normal) <⫺0·5 to ⫺1·0 (slow) <⫺1·0 none (slow) Missing data
34 8 5 5
33 6 5 6
∆=change. SDS=standard deviation score.
Table 2: Data on kidneys and ureters for all patients 26 followed-up at 10 years
22 followed-up at 10 years
Figure 1: Trial profile
distribution between the treatment groups was noted. Outcome at 4 years One or more bacteriologically confirmed urinary tract infections developed in 11 children in the medical and in six in the surgical group (relative risk 1·69, 95% CI 0·74–3·86). One child in the medical and two in the surgical group developed hypertension (systolic blood pressure >2 SDS) that required treatment. The change in SDS mean height was 0·25 (SE 0·14) in the medical and 0·47 (0·17) in the surgical group (difference 0·22, 95% CI ⫺0·22 to 0·66). VUR was successfully corrected in 47 of the 50 ureters in the surgical group with no complications apart from transient dilatation of the ureter in one child. There was minimum or no VUR (grade 0–I) at 4 years in 20 of the 54 ureters in the medical group (table 2). 26 children in the medical and 24 in the surgical group reached the primary endpoint of measurement of GFR at 4 years. Neither the absolute values nor the percentage change were significantly different at 4 years (table 3). Recalculation of the data, including the one GFR measurement deemed to be erroneous, did not affect the results. At entry all kidneys were scarred to a varied extent (table 2). After 4 years no new focal scars were seen, but
Medical
Surgical
Difference*
27 72·4 (24·1)
25 71·7 (22·6)
·· ··
⫺0·13 (⫺1·80 to 2·06) ⫺0·30 (⫺3·80 to 1·30)
4 years Number GFR ∆GFR
26 70·2 (26·3) ⫺2·4% (4·5)
24 73·7 (24·9) 4·7% (5·0)
·· 3·5 (⫺11·0 to 17·9) 7·1% (⫺6·4 to 20·6)
73 (28–130) 3 12 10
10 years Number GFR ∆GFR
26 68·3 (29·8) ⫺6·3% (5·9)
22 74·1 (35·6) 2·6% (7·6)
·· 5·8 (⫺11·0 to 22·6) 8·9% (⫺10·3 to 28·2)
Medical
Surgical
Number (boys)
27 (13)
25 (12)
Age (years) 1–5 6–12
4·7 (1·2–12·2) 15 12
4·4 (1·0–11·7) 13 12
Systolic blood pressure (SDS) Height (SDS)
0·04 (⫺1·03 to 2·07) ⫺0·10 (⫺3·10 to 1·20)
GFR (mL/min per 1·73 m2) 20–39 40–79 >79
76 (24–108) 3 12 12
Values are median (range), unless otherwise indicated. GFR=glomerular filtration rate. SDS=standard deviation score.
THE LANCET • Vol 357 • April 28, 2001
Outcome at 10 years After 4 years the children were followed up without protocol constraints, with some children cared for at other centres. Three patients in the medical group had their urteters reimplanted but their results were included in the analysis on the basis of intention to treat. Two further patients in each group developed hypertension that required treatment with drugs. Four children reached end-stage renal failure, and GFR results were obtained in 44 other children between 9·2 and 13·0 years after enrolment (26 from the
Entry Number GFR
Characteristics
Table 1: Characteristics at enrolment
there was progressive thinning of the parenchyma and contraction of scarred areas in 15 kidneys (seven patients from the medical and eight from the surgical group). This contraction was sufficient to change the scarring type from B or C to C or D in four kidneys from the medical and two from the surgical group. Renal length measurements were available in 25 of the 27 children in the medical and 22 of the 23 in the surgical group. Data were missing for three children because intravenous urography was poor quality, had been omitted for clinical reasons, or the child was unavailable for measurement. In both groups renal growth of ⫺0·5 change in SDS or more was recorded in 74% of kidneys with satisfactory urography films available. Renal growth less than ⫺0·5 change in SDS was seen in 13 (27%) of 49 kidneys in the medical and 11 (29%) of the 44 in the surgical group, and was confined to those with type D or extensive type B scarring in both groups, irrespective of whether VUR persisted or not.
∆=change; GFR=mean mL/min per 1·73m2 (SD); ∆GFR%=mean (SE); *=mean (95% CIs).
Table 3: Outcome for glomerular filtration rate (GFR) at 4 and 10 years’ follow-up
1331
For personal use. Only reproduce with permission from The Lancet Publishing Group.
GFR at 4 years follow-up (mL/min per 1·73 m2)
ARTICLES
also a boy, with an antenatal diagnosis of urological abnormality and bilateral renal dysplasia. He entered the study aged 2 years with a GFR of 27 mL/min per 1·73 m2 and reached end-stage renal failure aged 10 years. Patient c (figure 2) was a girl who entered the study aged 6 years with a GFR of 66 mL/min per 1·73 m2. She lost renal function 4 years later in association with labile hypertension and proteinuria after 2 years without clinic attendance and developed end-stage renal failure aged 12. Patient d (figure 2) had a normal intravenous urography aged 3 years when investigated for a urinary tract infection, but on referral aged 8 years, after further infections, he was shown to have bilateral VUR and renal scarring with a GFR of 28 mL/min per 1·73 m2 and was randomised into the surgical limb of the trial. He reached end-stage renal failure aged 17 years.
150 Medical Surgical
100
50
b,d
a
c
Discussion
0 0
50 100 GFR (mL/min per 1·73 m2 BSA)
150
Figure 2: GFR at enrolment and at 4 years Line=line of identity.
medical and 22 from the surgical group, median age 15·5 years, range 10·7–24·9). Of the remaining four, one had emigrated, and three declined or were unavailable for further investigation. No differences were seen between the groups for absolute values or percentage change in GFR (table 3). Two children from the medical and two from the surgical group (three boys and one girl) developed end-stage renal failure by 10 years after enrolment (relative risk 0·88, 95% CI 0·14–5·73, patients a–d in figures 2 and 3). All four had GFRs at entry below the median value, which fell more than 30% within the first 4 years (figure 2). None had proteinuria reported at enrolment, but all did by 4 years after enrolment. These children did not have documented urinary tract infections during this period, although all had had recurrent infections before entry. One of these patients was a boy (a, figure 2) who developed heavy proteinuria at 9 years after enrolment with normal sized kidneys on intravenous urography and ultrasound at age 11 years. At 13, he had end-stage renal failure, suggesting a glomerular contribution to the deterioration of function, but renal histology was not available. Another patient (b, figure 2) was GFR
>79
40–79
20–39 <20 a b
ESRF 0
c d
4 10 yr 0 4 10 yr Medical Surgical Figure 3: GFR groups at enrolment, 4, and 10 years The GFR groups are: >79, 40–79, 20–39, <20, and end-stage renal failure. In most instances the movement across the interface at 80 indicates only small changes in GFR. ESRF=end-stage renal failure.
1332
We have shown that there are no differences in the outcome of children with serious VUR using medical or surgical management strategies. Our study was done to compare surgical correction with medical management of children with severe non-obstructive bilateral VUR, including grade V, bilateral renal scarring, and a history of urinary tract infection. Children with this severity of disease are rare. Although they were recruited from referrals to two major paediatric nephro-urological units with wide catchment areas, recruitment took 5 years. The primary endpoint of the trial was the change in GFR at enrolment compared with that at 4 years after enrolment, expressed as a percentage change in GFR at entry. We chose this approach rather than the absolute change in GFR because we believe that it results in a better indication of the progress of the disorder. For example, an absolute change in GFR of 20 mL/min per 1·73 m2 represents a 20% decrease in a child who started with a GFR of 100, but a 50% decrease in one who started with a GFR of 40. There was a difference of 7·1% in the change in GFR4% between the two groups, but the confidence limits included zero. If there was a true difference of 7% then it would have required about 90 children in each group to have a 90% power of finding a significant difference. Such a finding might not be of clinical importance by itself, but would point to a different rate of progression between the two groups, which might become more obvious with longer follow-up. The failure to detect a difference between the medically and surgically managed patients might be due to small trial size, insufficient length of follow-up, or broad heterogeneity of the patient groups. The trial effect and continuity of care might also be important because of the fact that inclusion of a child in the study promoted better management, including control of factors such as urinary tract infection or hypertension, which under ordinary clinical care might selectively disadvantage one group.21 The coarse, irregular, and asymmetric nephropathy associated with VUR is an important cause of chronic renal failure in young people, although its frequency in children with end-stage renal failure varies depending on the diagnostic criteria applied and the population studied. Nephropathy is thought to be mainly a manifestation of acquired pyelonephritic scarring, a consequence of the facilitated access of bacteria into the renal parenchyma in urinary tract infections.1,22–24 However, the importance of renal dysplasia with a genetic background in the pathogenesis of nephropathy is increasingly recognised—ie, maldevelopment of the kidney in association with a congenital anomaly of the uretero-vesical junction.25–27 Congenital and acquired nephropathies are, however, not mutually exclusive and can coexist. 11 of the children in this study had a family history of VUR, but 16 were known to
THE LANCET • Vol 357 • April 28, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
have had a radiologically normal kidney (eight had two normal kidneys) before entering the study with bilateral scarring.20 Surgical correction of the VUR in children has been practised for at least 40 years.28 Since we designed our protocol, 17 years ago, there had been a move away from surgical correction of VUR in cases such as those in our trial because of awareness of the contribution of renal dysplasia to nephropathy.23 However, there has not been any controlled evidence of benefit to renal function from the procedure. In part this is because VUR is a heterogeneous condition with only a few patients slowly progressing to endstage renal failure. Although a number of randomised studies of medical and surgical management of children with severe VUR have been done, none have been confined to those in whom the condition was bilateral and associated with bilateral renal scarring. Astley and colleagues4,5 did a study of about 100 children with VUR and included unilateral cases, and showed no difference in outcome between medical and surgical treatment of patients at 2 years’, or at 5 years’ follow-up. The International Study of Reflux in Children6–9 is a randomised trial of medical or surgical management, confined to those with VUR grades III and IV. Although 49% of the 306 children enrolled in the European group had renal scarring at entry (33% in both kidneys), a GFR less than 70 mL/min per 1·73 m2 was an exclusion criterion. There were no differences in renal growth, function, or scarring at 5 years, or renal function at 10 years. The main factor that determines outcome in studies such as ours seems to be the extent of renal parenchymal reduction and the degree of functional impairment at the start. In our study population, factors such as failure of renal growth associated with dysplasia, hypertensive nephropathy, or remnant kidney nephropathy with proteinuria and occasionally focal glomerulosclerosis might occur whether VUR is present or not.29 These factors contribute to the heterogeneity of the four patients who developed end-stage renal failure. This possibility does not exclude a deleterious effect of VUR, especially with urinary tract infection, at an earlier stage of the disease but makes it difficult to detect. A study with a larger, longer follow-up of a more homogeneous group of patients than in our study would be necessary.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 19
Contributors J M Smellie, T M Barratt, C Chantler, and P G Ransley conceived and designed the study. J M Smellie was responsible for the conduct of the trial and the clinical management of the patients. T M Barratt, J M Smellie, N P Prescod, and A S Woolf analysed and interpreted the data and drafted the article. N P Prescod organised the randomisation procedure, collected the data and maintained the records. P G Ransley and colleagues did the surgery. I Gordon and colleagues were responsible for the imaging. T M Barratt, C Chantler, and A S Woolf participated in follow-up of the children after 4 years. S Logan provided statistical advice. C J Reid and S L Cohen continued follow-up of the patients at Guy’s and University College London Hospitals, respectively. All contributors approved the final version of the manuscript.
20
21
22 23
24
Acknowledgments We thank our consultant colleagues for referring patients for entry into the trial, junior medical staff for assistance in the inpatient and outpatient management of the children and members of the nursing staff for help in the clinics. We are especially grateful to the children and their parents for their co-operation. J M Smellie and NPP were supported by the National Kidney Research Fund, the Children Nationwide Medical Research Fund, and the Kidney Research Aid Fund.
References 1 2
Smellie JM, Edwards D, Hunter N, Normand ICS, Prescod N. Vesicoureteric reflux and renal scarring. Kidney Int 1975; 8: S65–72. Smellie JM, Prescod NP, Shaw PJ, Risdon RA, Bryant TN. Childhood reflux and urinary infection: a follow-up of 10–41 years in 226 adults.
THE LANCET • Vol 357 • April 28, 2001
25
26
27
28 29
Pediatr Nephrol 1998; 12: 727–36. Lewis M. Report of the Paediatric Renal Registry 1999. In: Ansell D, Feest T. The Second Annual Report of the UK Renal Registry. Bristol: Renal Association 1999; 175–87. Astley R, Clark RC, Corkery JJ, et al. Prospective trial of operative versus nonoperative treatment of severe vesicoureteric reflux in children: two years’ observation in 96 children. BMJ 1983; 287: 171–74. Astley R, Clark RC, Corkery JJ, et al. Prospective trial of operative versus nonoperative treatment of severe vesicoureteric reflux in children: five years’ observation. BMJ 1987; 295: 237–41. Olbing H, Claësson I, Ebel K-D, et al. Renal scars and parenchymal thinning in children with vesicoureteral reflux: a 5-year report of the International Reflux Study in Children (European branch). J Urol 1992; 148: 1653–56. Smellie JM, Tamminen-Möbius T, Olbing H, et al. The International Reflux Study in Children: five-year study of medical or surgical treatment in children with severe reflux: radiological renal findings. Pediatr Nephrol 1992; 6: 223–30. Piepsz A, Tamminen-Möbius T, Reiners C, et al. Five-year study of medical or surgical treatment in children with severe vesico-ureteral reflux: dimercaptosuccinic acid findings. Eur J Pediatr 1998; 157: 753–58. Olbing H, Hirche H, Koskimies O, et al. Renal growth over 10 years in a prospective study of medical or surgical treatment in children with severe vesicoureteral reflux: 10-year prospective study of medical and surgical treatment. Radiology 2000; 216: 731–37. Lebowitz RL, Olbing H, Parkkulainen KV, Smellie JM, Tamminen-Möbius TE. International Reflux Study in Children: international system of radiographic grading of vesicoureteric reflux. Pediatr Radiol 1985; 15: 105–59. Aperia A, Broberger O, Theodenius K, Zetterström R. Development of renal control of salt and fluid homeostasis during the first year of life. Acta Padiatr Scand 1975; 64: 393–98. Cohen SJ. The Cohen or transtrigonal method of ureteroneocystostomy. In: O’Donnell B, Koff SA, eds. Pediatric Urology, 3rd edn. Oxford: Butterworth-Heinemann, 1997, 466–69. Tanner J, Whitehouse R, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965—I. Arch Dis Child 1966; 41: 454–71. Tanner J, Whitehouse R, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965—II. Arch Dis Child 1966; 41: 613–35. Horan MJ, Kalkner B, Kimm SYS, et al. Second Task Force on Blood Pressure Control in Children: report of the Second Task Force on Blood Pressure Control in Children—1987. Pediatrics 1987; 79: 1–29. Hodson CJ. Reflux nephropathy: scoring the damage. In: Hodson CJ. Kincaid-Smith P, eds. Reflux Nephropathy. New York: Masson Publishing, 1979: 29–47. Chantler C, Barratt TM. Estimation of glomerular filtration rate from the plasma clearance of 51-chromium edetic acid. Arch Dis Child 1972; 47: 613–17. Dubois D, Dubois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 1916; 17: 863–71. Smellie JM, Rigden SPA. Pitfalls in the investigation of children with urinary tract infection. Arch Dis Child 1995; 72: 251–55. Smellie JM, Poulton A, Prescod NP. Retrospective study of children with renal scarring associated with reflux and urinary infection. BMJ 1994; 308: 1193–96. Bergstrom J, Alvestrand A, Bucht H, Gutienez A. Progression of chronic renal failure in man is retarded with more frequent follow-ups and blood pressure control. Clin Nephrol 1986; 25: 1–6. Ransley PG, Risdon RA. Reflux and renal scarring. Br J Radiol 1978; 51 (suppl 14): 1–35. Wennerström M, Hanson S, Jodal U, Stockland E. Primary and acquired renal scarring in boys and girls with urinary tract infection. J Pediatr 2000; 136: 30–34. Smellie JM, Ransley PG, Normand ICS, Prescod N, Edwards D. Development of new renal scars: a collaborative study. BMJ 1985; 290: 1957–60. Hinchliffe SA, Chan YF, Jones H, Chan N, Kreezy A, van Velzen D. Renal hypoplasia and postnatally acquired cortical loss in children with vesicoureteral reflux. Pediatr Nephrol 1992; 6: 439–44. Risdon RA, Yeung CK, Ransley PG. Reflux nephropathy in children submitted to unilateral nephrectomy: a clinicopathology study. Clin Nephrol 1993; 40: 308–14. Feather SA, Malcolm SA, Woolf AS, et al. Primary, non-syndromic vesicoureteric reflux and its nephropathy is genetically heterogeneous, with a locus on chromosome 1. Am J Hum Genet 2000; 66: 1420–25. Politano VA, Leadbetter WF. Operative technique for correction of vesico-ureteric reflux. J Urol 1958; 79: 932–47. Fogo AB, Kon V. Pathophysiology of progressive renal disease. In: Barratt TM, Avner ED, Harmon WE, eds. Pediatric nephrology, 4th edn. Baltimore: Lippincott Williams & Wilkins, 1999; 1183–96.
1333
For personal use. Only reproduce with permission from The Lancet Publishing Group.
Medical versus surgical treatment in children with severe bilateral vesicoureteric reflux and bilateral nephropathy: a randomised trial Jean M Smellie, T Martin Barratt, Cyril Chantler, Isky Gordon, Nina P Prescod, Philip G Ransley, Adrian S Woolf
Summary Background Nephropathy associated with vesicoureteric reflux (VUR) and urinary tract infection can result in end-stage renal failure, hypertension, or both. Whether long-term VUR contributes to these outcomes is unknown. We compared, in a randomised trial, medical with surgical management of children with bilateral severe VUR and bilateral nephropathy. Methods We stratified by age and glomerular filtration rate (GFR) 25 boys and 27 girls aged 1–12 years and randomly assigned them to medical or surgical management. At enrolment and 4 years’ follow-up we estimated GFR from the plasma clearance of 51Cr-labelled edetic acid (EDTA), and did intravenous urography. We also did a metastable 99mTc-labelled dimercaptosuccinic acid (DMSA) assay and contrast cystography. The change in GFR at 4 years, expressed as a percentage change between enrolment and 4 years, was available for 26 of 27 patients in the medical and 24 of 25 in the surgical group. We assessed GFR in 48 patients 10 years after enrolment. Findings Mean GFR at enrolment was 72·4 mL/min per 1·73 m2 (SD 24·1) in the medical and 71·7 mL/min per 1·73 m2 (22·6) in the surgical group. The mean percentage change in GFR at 4 years was ⫺2·4% (SE 4·5) versus 4·7% (5·0) in the medical and surgical groups, respectively. The difference in change in GFR at 4 years between the two groups was not significant (7·1%, 95% CI ⫺6·4% to 20·6%). Interpretation Our data do not lend support to the view that the outcome for renal function is improved by surgical correction of VUR in children with bilateral disease. Lancet 2001; 357: 1329–33 See Commentary page 1309
Introduction Vesicoureteric reflux (VUR) can be associated with other abnormalities of the bladder or lower urinary tract. 1–2% of apparently healthy children have this disorder, and it is present in 30–40% of children with urinary tract infection, of whom about 30% have radiological evidence of renal scarring. Such scarring is a coarse irregular nephropathy resulting from congenital dysplasia, acquired postinfectious damage, or both.1 This nephropathy is a major cause of severe hypertension in children and young adults. It occasionally progresses to chronic renal failure,2 which accounts for about 25% of UK children with end-stage renal failure, requiring regular dialysis or transplantation.3 Several factors might lead to this progression, including inadequate renal growth, further scarring, associated hypertension, and secondary glomerulosclerosis with proteinuria. Common practice has been to correct VUR surgically, on the assumption that this disorder predisposes to urinary tract infection and facilitates the spread of infection to the renal parenchyma. However, findings from comparative trials of medical and surgical management of children with severe VUR have shown no significant difference in renal growth or acquisition of new scars for 10 years.4–9 The optimum management of children with bilateral severe VUR and bilateral nephropathy, the group thought to have the highest risk of developing end-stage renal failure,2 is not known. We, therefore, undertook a randomised trial of medical management versus surgical correction in 53 such children, with a glomerular filtration rate (GFR) at enrolment down to 20 mL/min per 1·73 m2 body surface area. We aimed to assess whether the expected decline in renal function could be prevented by surgical correction of VUR. In this study we report the clinical, renal function, and imaging outcomes at 4 years, and the clinical outcome and GFR at 10 years.
Methods
Great Ormond Street Hospital for Children NHS Trust and Institute of Child Health, University College London, London WC1N 1EH, UK (J M Smellie FRCP, Prof T M Barratt FRCP, I Gordon FRCP, N P Prescod BSc, P G Ransley FRCS, Prof A S Woolf FRCPCH); and King’s College, Guy’s and St Thomas’ Hospitals’ Medical and Dental Schools (J M Smellie, Prof C Chantler FRCP, N P Prescod) Correspondence to: Prof T Martin Barratt (e-mail: [email protected])
THE LANCET • Vol 357 • April 28, 2001
Patients Children were eligible for enrolment if they were attending the nephrology or urology clinics at Great Ormond Street or Guy’s Hospitals, London, UK, between Nov 1, 1985, and Dec 31, 1989, and had bilateral severe primary VUR (International Reflux Study in Children grades III–V),10 bilateral renal scarring diagnosed on intravenous urography with corresponding defects in metastable 99mTc-labelled dimercaptosuccinic acid (DMSA) uptake, and a history of bacteriologically confirmed urinary tract infection. The criteria for enrolment, were: age from 1 to 12 years; bilateral severe VUR diagnosed by contrast micturating cystourethrography, GFR of 20 mL/min per 1·73 m2 or more, no current hypertension (or hypertension controlled by antihypertensive therapy for more than 6 months), and no current urinary tract infection. Children younger than 1 year were excluded from the trial because of the confounding effect of the normal increase in GFR at that age11 and of the high surgical complication rate. We excluded children with bladder outflow obstruction, overt neuropathic bladder, previous surgery to the uretero-vesical junction, a duplex system or one kidney, or one kidney contributing less than 10% to overall renal function.
1329
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Study design The children were taken from a large group of unknown size, who might have been suitable for the trial, and were referred by consultant nephrological and urological staff to an investigator (JMS) for review. 53 children fulfilled the criteria and parental written informed consent for the trial was obtained for every child. We stratified children by age (1–5 or 6–12 years) and GFR (20–39, 40–79, or ⭓80 mL/min per 1·73 m2). Envelopes for the six categories were prepared containing equal numbers of cards for each treatment arm. An investigator (NPP) took a concealed card from the appropriate envelope to randomly assign children to a group. Children in the medical management group took antibacterial prophylaxis with co-trimoxazole (sulfamethoxazole 5–10 mg per kg daily and trimethoprim 1–2 mg per kg daily), trimethoprim (1–2 mg per kg daily) alone, or nitrofurantoin (1–2 mg per kg daily). These children had double micturition at bedtime, and we carefully noted bladder and bowel habits. We treated those with bacteriologically confirmed urinary tract infection with appropriate antibacterial drugs. An investigator (JMS) supervised the children in special clinics every 3 months for the first 4 years and at longer intervals thereafter. Surgical correction of VUR was done as soon as possible after randomisation and completion of the preliminary investigations (median 0·19 years, range 0·01–0·79). Individual surgeons used their preferred technique, usually a Cohen ureteric advancement.12 Antibacterial prophylaxis was stopped for patients in the surgical group after 6 months if VUR was shown to have been corrected by micturating cystourethrography or indirect radionuclide cystogram. The trial protocol was approved by the ethics committees of both hospitals. Patient assessment Before the children were randomly assigned to groups we measured height, weight, and systolic blood pressure, with standard deviation scores (SDS) derived from published data.13–15 We also did urine microscopy and culture, and dipstick test for protein. We measured GFR and did contrast micturating cystourethrography, a renal 99mTcDMSA scan, and limited intravenous urography (one to two x-ray films only). These measurements were used as enrolment values if they had been done less than 6 months before randomisation. We did the DMSA scan, limited intravenous urography, cystogram, and GFR again about 4 years after initial measurements (GFR was measured between 3·7 and 4·9 years). We graded VUR between 0 and V, in accordance with International Reflux Study in Children Criteria.10 For patients in the surgical group we did a postoperative 99mTcdiethylenetriamine penta-acetic acid (DTPA) scan, with micturating cystourethrography or indirect radionuclide cystogram. These were only repeated if we saw VUR or evidence of uretero-vesical junction obstruction (in the latter years of follow-up we sometimes did an indirect radionuclide cystogram instead of micturating 99m 99m cystourethrography with Tc-DTPA or Tcmercaptoacetyltriglycine [MAG3]). We measured renal length with intravenous urography, related this measurement to height, and expressed the relation as SDS.16 We calculated renal growth as the difference between SDS at 4 years and enrolment. Growth of ⫺0·5 or more was satisfactory. Renal scarring on intravenous urography and 99mTc-DMSA image defects were independently assessed at entry and after completion of the trial by at least two investigators. Radiological scarring was typed from A to D: type A indicated 1 or 2 scars, type B
1330
more extensive scarring with some normal parenchyma, type C an irregularly thinned rim of parenchyma that surrounded deformed renal papillae and calyces, and type D a small deformed kidney.1 Renal function We estimated GFR from the plasma clearance of 51 Cr-edetic acid (EDTA) by single exponential analysis of blood samples at 2 and 4 h after injection17 and corrected it to 1·73 m2 of body surface area with the Dubois formula.18 The coefficient of variation of replicate estimates of GFR by this method in the same individual is 3·9%.17 The difference in each child between the GFR at 4 years (GFR4) and the initial GFR (GFR0) was expressed as a percentage of GFR0: ∆GFR4%=100⫻(GFR4⫺GFR0)/GFR0 We measured GFR about 10 years after entry (GFR10) in 44 children and calculated the change from GFR0 in the same way as at 4 years (∆GFR10%). In the four children who had reached end-stage renal failure, GFR was not measured but was deemed as being 5 mL/min per 1·73 m2 for statistical purposes. Departures from protocol (figure 1) We withdrew one child in the medical group because she was judged on review after randomisation, to have only unilateral renal damage, and was lost to follow-up after 2 years. Two children in the surgical group proved to have unilateral VUR: one was detected on subsequent review of micturating cystourethrography, and the other (who had previously had bilateral severe VUR) on cystography during surgery. These two children had only unilateral ureteric reimplantation and no reflux was seen on postoperative cystography. Both had bilateral nephropathy and were retained in the trial on the basis of intention to treat. We excluded one child in the surgical group from the 4-year analysis because the estimate of GFR at 4 years of 174 mL/min per 1·73 m2 was deemed erroneous because of a high volume of distribution of the 51Cr-EDTA with flanking values of 102 and 113. GFR at 4 years was not available for another child in the medical group because she had emigrated. Both were available for follow-up at 10 years. Statistical analysis The primary endpoint was a comparison of mean percentage change in GFR at 4 years in the medical and surgical groups. We calculated the difference between group means and 95% CI. Other endpoints were somatic growth, blood pressure, recurrence of urinary tract infections, renal growth, and additional renal scarring. We calculated between-group differences on other continuous variables and the 95% CI. For dichotomous variables, relative risk and 95% CI were calculated with EPI INFO version 6.04. Observed differences and relative risks were not regarded as significant if they fell within the 95% CI.
Results Entry data 52 children were followed up: 28 were randomly assigned to medical management (one child was withdrawn) and 25 to surgery (figure 1). There was a family history of VUR in 11 children. The antecedent histories of the patients have been published.19,20 16 children had a radiologically normal kidney (two normal kidneys in eight) before enrolment with bilateral disease. Two children, one in each group, had hypertension controlled by drugs. At enrolment there were no differences between the two groups at entry for sex, age, SDS for systolic blood pressure and height, or GFR (table 1). Table 2 shows VUR grade, scar type, and kidney length SDS in the 104 kidneys and ureters. A closely similar
THE LANCET • Vol 357 • April 28, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
Medical (n=54)
Surgical (n=50)
Enrolment
4 years’ follow-up
Enrolment
4 years’ follow-up
VUR grade 0 I II III IV V
0 0 0 8 29 17
11 9 6 13 11 4
1 0 2 15 19 13
46 1 0 1 2 0
Renal scar type A B C D
16 25 8 5
16 21 8 9
14 26 6 4
14 24 6 6
Renal length SDS >⫺1·0 ⫺1·0 to⫺2·0 <⫺2·0 Missing data
38 9 3 4
30 13 6 5
32 10 6 2
32 8 6 4
Children aged 1–12 years with bilateral VUR and nephropathy and history of UTI Excluded if other urological abnormality GFR <20 mL/min per 1·73 m2 or uncontrolled hypertension 53 randomised
28 assigned to medical management
25 assigned to surgical management
1 withdrawn 1 not available for primary endpoint at 4 years
1 not available for primary endpoint at 4 years
26 analysed for primary endpoint
24 analysed for primary endpoint
Renal growth ∆SDS ⭓⫺0·5 (normal) <⫺0·5 to ⫺1·0 (slow) <⫺1·0 none (slow) Missing data
34 8 5 5
33 6 5 6
∆=change. SDS=standard deviation score.
Table 2: Data on kidneys and ureters for all patients 26 followed-up at 10 years
22 followed-up at 10 years
Figure 1: Trial profile
distribution between the treatment groups was noted. Outcome at 4 years One or more bacteriologically confirmed urinary tract infections developed in 11 children in the medical and in six in the surgical group (relative risk 1·69, 95% CI 0·74–3·86). One child in the medical and two in the surgical group developed hypertension (systolic blood pressure >2 SDS) that required treatment. The change in SDS mean height was 0·25 (SE 0·14) in the medical and 0·47 (0·17) in the surgical group (difference 0·22, 95% CI ⫺0·22 to 0·66). VUR was successfully corrected in 47 of the 50 ureters in the surgical group with no complications apart from transient dilatation of the ureter in one child. There was minimum or no VUR (grade 0–I) at 4 years in 20 of the 54 ureters in the medical group (table 2). 26 children in the medical and 24 in the surgical group reached the primary endpoint of measurement of GFR at 4 years. Neither the absolute values nor the percentage change were significantly different at 4 years (table 3). Recalculation of the data, including the one GFR measurement deemed to be erroneous, did not affect the results. At entry all kidneys were scarred to a varied extent (table 2). After 4 years no new focal scars were seen, but
Medical
Surgical
Difference*
27 72·4 (24·1)
25 71·7 (22·6)
·· ··
⫺0·13 (⫺1·80 to 2·06) ⫺0·30 (⫺3·80 to 1·30)
4 years Number GFR ∆GFR
26 70·2 (26·3) ⫺2·4% (4·5)
24 73·7 (24·9) 4·7% (5·0)
·· 3·5 (⫺11·0 to 17·9) 7·1% (⫺6·4 to 20·6)
73 (28–130) 3 12 10
10 years Number GFR ∆GFR
26 68·3 (29·8) ⫺6·3% (5·9)
22 74·1 (35·6) 2·6% (7·6)
·· 5·8 (⫺11·0 to 22·6) 8·9% (⫺10·3 to 28·2)
Medical
Surgical
Number (boys)
27 (13)
25 (12)
Age (years) 1–5 6–12
4·7 (1·2–12·2) 15 12
4·4 (1·0–11·7) 13 12
Systolic blood pressure (SDS) Height (SDS)
0·04 (⫺1·03 to 2·07) ⫺0·10 (⫺3·10 to 1·20)
GFR (mL/min per 1·73 m2) 20–39 40–79 >79
76 (24–108) 3 12 12
Values are median (range), unless otherwise indicated. GFR=glomerular filtration rate. SDS=standard deviation score.
THE LANCET • Vol 357 • April 28, 2001
Outcome at 10 years After 4 years the children were followed up without protocol constraints, with some children cared for at other centres. Three patients in the medical group had their urteters reimplanted but their results were included in the analysis on the basis of intention to treat. Two further patients in each group developed hypertension that required treatment with drugs. Four children reached end-stage renal failure, and GFR results were obtained in 44 other children between 9·2 and 13·0 years after enrolment (26 from the
Entry Number GFR
Characteristics
Table 1: Characteristics at enrolment
there was progressive thinning of the parenchyma and contraction of scarred areas in 15 kidneys (seven patients from the medical and eight from the surgical group). This contraction was sufficient to change the scarring type from B or C to C or D in four kidneys from the medical and two from the surgical group. Renal length measurements were available in 25 of the 27 children in the medical and 22 of the 23 in the surgical group. Data were missing for three children because intravenous urography was poor quality, had been omitted for clinical reasons, or the child was unavailable for measurement. In both groups renal growth of ⫺0·5 change in SDS or more was recorded in 74% of kidneys with satisfactory urography films available. Renal growth less than ⫺0·5 change in SDS was seen in 13 (27%) of 49 kidneys in the medical and 11 (29%) of the 44 in the surgical group, and was confined to those with type D or extensive type B scarring in both groups, irrespective of whether VUR persisted or not.
∆=change; GFR=mean mL/min per 1·73m2 (SD); ∆GFR%=mean (SE); *=mean (95% CIs).
Table 3: Outcome for glomerular filtration rate (GFR) at 4 and 10 years’ follow-up
1331
For personal use. Only reproduce with permission from The Lancet Publishing Group.
GFR at 4 years follow-up (mL/min per 1·73 m2)
ARTICLES
also a boy, with an antenatal diagnosis of urological abnormality and bilateral renal dysplasia. He entered the study aged 2 years with a GFR of 27 mL/min per 1·73 m2 and reached end-stage renal failure aged 10 years. Patient c (figure 2) was a girl who entered the study aged 6 years with a GFR of 66 mL/min per 1·73 m2. She lost renal function 4 years later in association with labile hypertension and proteinuria after 2 years without clinic attendance and developed end-stage renal failure aged 12. Patient d (figure 2) had a normal intravenous urography aged 3 years when investigated for a urinary tract infection, but on referral aged 8 years, after further infections, he was shown to have bilateral VUR and renal scarring with a GFR of 28 mL/min per 1·73 m2 and was randomised into the surgical limb of the trial. He reached end-stage renal failure aged 17 years.
150 Medical Surgical
100
50
b,d
a
c
Discussion
0 0
50 100 GFR (mL/min per 1·73 m2 BSA)
150
Figure 2: GFR at enrolment and at 4 years Line=line of identity.
medical and 22 from the surgical group, median age 15·5 years, range 10·7–24·9). Of the remaining four, one had emigrated, and three declined or were unavailable for further investigation. No differences were seen between the groups for absolute values or percentage change in GFR (table 3). Two children from the medical and two from the surgical group (three boys and one girl) developed end-stage renal failure by 10 years after enrolment (relative risk 0·88, 95% CI 0·14–5·73, patients a–d in figures 2 and 3). All four had GFRs at entry below the median value, which fell more than 30% within the first 4 years (figure 2). None had proteinuria reported at enrolment, but all did by 4 years after enrolment. These children did not have documented urinary tract infections during this period, although all had had recurrent infections before entry. One of these patients was a boy (a, figure 2) who developed heavy proteinuria at 9 years after enrolment with normal sized kidneys on intravenous urography and ultrasound at age 11 years. At 13, he had end-stage renal failure, suggesting a glomerular contribution to the deterioration of function, but renal histology was not available. Another patient (b, figure 2) was GFR
>79
40–79
20–39 <20 a b
ESRF 0
c d
4 10 yr 0 4 10 yr Medical Surgical Figure 3: GFR groups at enrolment, 4, and 10 years The GFR groups are: >79, 40–79, 20–39, <20, and end-stage renal failure. In most instances the movement across the interface at 80 indicates only small changes in GFR. ESRF=end-stage renal failure.
1332
We have shown that there are no differences in the outcome of children with serious VUR using medical or surgical management strategies. Our study was done to compare surgical correction with medical management of children with severe non-obstructive bilateral VUR, including grade V, bilateral renal scarring, and a history of urinary tract infection. Children with this severity of disease are rare. Although they were recruited from referrals to two major paediatric nephro-urological units with wide catchment areas, recruitment took 5 years. The primary endpoint of the trial was the change in GFR at enrolment compared with that at 4 years after enrolment, expressed as a percentage change in GFR at entry. We chose this approach rather than the absolute change in GFR because we believe that it results in a better indication of the progress of the disorder. For example, an absolute change in GFR of 20 mL/min per 1·73 m2 represents a 20% decrease in a child who started with a GFR of 100, but a 50% decrease in one who started with a GFR of 40. There was a difference of 7·1% in the change in GFR4% between the two groups, but the confidence limits included zero. If there was a true difference of 7% then it would have required about 90 children in each group to have a 90% power of finding a significant difference. Such a finding might not be of clinical importance by itself, but would point to a different rate of progression between the two groups, which might become more obvious with longer follow-up. The failure to detect a difference between the medically and surgically managed patients might be due to small trial size, insufficient length of follow-up, or broad heterogeneity of the patient groups. The trial effect and continuity of care might also be important because of the fact that inclusion of a child in the study promoted better management, including control of factors such as urinary tract infection or hypertension, which under ordinary clinical care might selectively disadvantage one group.21 The coarse, irregular, and asymmetric nephropathy associated with VUR is an important cause of chronic renal failure in young people, although its frequency in children with end-stage renal failure varies depending on the diagnostic criteria applied and the population studied. Nephropathy is thought to be mainly a manifestation of acquired pyelonephritic scarring, a consequence of the facilitated access of bacteria into the renal parenchyma in urinary tract infections.1,22–24 However, the importance of renal dysplasia with a genetic background in the pathogenesis of nephropathy is increasingly recognised—ie, maldevelopment of the kidney in association with a congenital anomaly of the uretero-vesical junction.25–27 Congenital and acquired nephropathies are, however, not mutually exclusive and can coexist. 11 of the children in this study had a family history of VUR, but 16 were known to
THE LANCET • Vol 357 • April 28, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
ARTICLES
have had a radiologically normal kidney (eight had two normal kidneys) before entering the study with bilateral scarring.20 Surgical correction of the VUR in children has been practised for at least 40 years.28 Since we designed our protocol, 17 years ago, there had been a move away from surgical correction of VUR in cases such as those in our trial because of awareness of the contribution of renal dysplasia to nephropathy.23 However, there has not been any controlled evidence of benefit to renal function from the procedure. In part this is because VUR is a heterogeneous condition with only a few patients slowly progressing to endstage renal failure. Although a number of randomised studies of medical and surgical management of children with severe VUR have been done, none have been confined to those in whom the condition was bilateral and associated with bilateral renal scarring. Astley and colleagues4,5 did a study of about 100 children with VUR and included unilateral cases, and showed no difference in outcome between medical and surgical treatment of patients at 2 years’, or at 5 years’ follow-up. The International Study of Reflux in Children6–9 is a randomised trial of medical or surgical management, confined to those with VUR grades III and IV. Although 49% of the 306 children enrolled in the European group had renal scarring at entry (33% in both kidneys), a GFR less than 70 mL/min per 1·73 m2 was an exclusion criterion. There were no differences in renal growth, function, or scarring at 5 years, or renal function at 10 years. The main factor that determines outcome in studies such as ours seems to be the extent of renal parenchymal reduction and the degree of functional impairment at the start. In our study population, factors such as failure of renal growth associated with dysplasia, hypertensive nephropathy, or remnant kidney nephropathy with proteinuria and occasionally focal glomerulosclerosis might occur whether VUR is present or not.29 These factors contribute to the heterogeneity of the four patients who developed end-stage renal failure. This possibility does not exclude a deleterious effect of VUR, especially with urinary tract infection, at an earlier stage of the disease but makes it difficult to detect. A study with a larger, longer follow-up of a more homogeneous group of patients than in our study would be necessary.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 19
Contributors J M Smellie, T M Barratt, C Chantler, and P G Ransley conceived and designed the study. J M Smellie was responsible for the conduct of the trial and the clinical management of the patients. T M Barratt, J M Smellie, N P Prescod, and A S Woolf analysed and interpreted the data and drafted the article. N P Prescod organised the randomisation procedure, collected the data and maintained the records. P G Ransley and colleagues did the surgery. I Gordon and colleagues were responsible for the imaging. T M Barratt, C Chantler, and A S Woolf participated in follow-up of the children after 4 years. S Logan provided statistical advice. C J Reid and S L Cohen continued follow-up of the patients at Guy’s and University College London Hospitals, respectively. All contributors approved the final version of the manuscript.
20
21
22 23
24
Acknowledgments We thank our consultant colleagues for referring patients for entry into the trial, junior medical staff for assistance in the inpatient and outpatient management of the children and members of the nursing staff for help in the clinics. We are especially grateful to the children and their parents for their co-operation. J M Smellie and NPP were supported by the National Kidney Research Fund, the Children Nationwide Medical Research Fund, and the Kidney Research Aid Fund.
References 1 2
Smellie JM, Edwards D, Hunter N, Normand ICS, Prescod N. Vesicoureteric reflux and renal scarring. Kidney Int 1975; 8: S65–72. Smellie JM, Prescod NP, Shaw PJ, Risdon RA, Bryant TN. Childhood reflux and urinary infection: a follow-up of 10–41 years in 226 adults.
THE LANCET • Vol 357 • April 28, 2001
25
26
27
28 29
Pediatr Nephrol 1998; 12: 727–36. Lewis M. Report of the Paediatric Renal Registry 1999. In: Ansell D, Feest T. The Second Annual Report of the UK Renal Registry. Bristol: Renal Association 1999; 175–87. Astley R, Clark RC, Corkery JJ, et al. Prospective trial of operative versus nonoperative treatment of severe vesicoureteric reflux in children: two years’ observation in 96 children. BMJ 1983; 287: 171–74. Astley R, Clark RC, Corkery JJ, et al. Prospective trial of operative versus nonoperative treatment of severe vesicoureteric reflux in children: five years’ observation. BMJ 1987; 295: 237–41. Olbing H, Claësson I, Ebel K-D, et al. Renal scars and parenchymal thinning in children with vesicoureteral reflux: a 5-year report of the International Reflux Study in Children (European branch). J Urol 1992; 148: 1653–56. Smellie JM, Tamminen-Möbius T, Olbing H, et al. The International Reflux Study in Children: five-year study of medical or surgical treatment in children with severe reflux: radiological renal findings. Pediatr Nephrol 1992; 6: 223–30. Piepsz A, Tamminen-Möbius T, Reiners C, et al. Five-year study of medical or surgical treatment in children with severe vesico-ureteral reflux: dimercaptosuccinic acid findings. Eur J Pediatr 1998; 157: 753–58. Olbing H, Hirche H, Koskimies O, et al. Renal growth over 10 years in a prospective study of medical or surgical treatment in children with severe vesicoureteral reflux: 10-year prospective study of medical and surgical treatment. Radiology 2000; 216: 731–37. Lebowitz RL, Olbing H, Parkkulainen KV, Smellie JM, Tamminen-Möbius TE. International Reflux Study in Children: international system of radiographic grading of vesicoureteric reflux. Pediatr Radiol 1985; 15: 105–59. Aperia A, Broberger O, Theodenius K, Zetterström R. Development of renal control of salt and fluid homeostasis during the first year of life. Acta Padiatr Scand 1975; 64: 393–98. Cohen SJ. The Cohen or transtrigonal method of ureteroneocystostomy. In: O’Donnell B, Koff SA, eds. Pediatric Urology, 3rd edn. Oxford: Butterworth-Heinemann, 1997, 466–69. Tanner J, Whitehouse R, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965—I. Arch Dis Child 1966; 41: 454–71. Tanner J, Whitehouse R, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965—II. Arch Dis Child 1966; 41: 613–35. Horan MJ, Kalkner B, Kimm SYS, et al. Second Task Force on Blood Pressure Control in Children: report of the Second Task Force on Blood Pressure Control in Children—1987. Pediatrics 1987; 79: 1–29. Hodson CJ. Reflux nephropathy: scoring the damage. In: Hodson CJ. Kincaid-Smith P, eds. Reflux Nephropathy. New York: Masson Publishing, 1979: 29–47. Chantler C, Barratt TM. Estimation of glomerular filtration rate from the plasma clearance of 51-chromium edetic acid. Arch Dis Child 1972; 47: 613–17. Dubois D, Dubois EF. A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med 1916; 17: 863–71. Smellie JM, Rigden SPA. Pitfalls in the investigation of children with urinary tract infection. Arch Dis Child 1995; 72: 251–55. Smellie JM, Poulton A, Prescod NP. Retrospective study of children with renal scarring associated with reflux and urinary infection. BMJ 1994; 308: 1193–96. Bergstrom J, Alvestrand A, Bucht H, Gutienez A. Progression of chronic renal failure in man is retarded with more frequent follow-ups and blood pressure control. Clin Nephrol 1986; 25: 1–6. Ransley PG, Risdon RA. Reflux and renal scarring. Br J Radiol 1978; 51 (suppl 14): 1–35. Wennerström M, Hanson S, Jodal U, Stockland E. Primary and acquired renal scarring in boys and girls with urinary tract infection. J Pediatr 2000; 136: 30–34. Smellie JM, Ransley PG, Normand ICS, Prescod N, Edwards D. Development of new renal scars: a collaborative study. BMJ 1985; 290: 1957–60. Hinchliffe SA, Chan YF, Jones H, Chan N, Kreezy A, van Velzen D. Renal hypoplasia and postnatally acquired cortical loss in children with vesicoureteral reflux. Pediatr Nephrol 1992; 6: 439–44. Risdon RA, Yeung CK, Ransley PG. Reflux nephropathy in children submitted to unilateral nephrectomy: a clinicopathology study. Clin Nephrol 1993; 40: 308–14. Feather SA, Malcolm SA, Woolf AS, et al. Primary, non-syndromic vesicoureteric reflux and its nephropathy is genetically heterogeneous, with a locus on chromosome 1. Am J Hum Genet 2000; 66: 1420–25. Politano VA, Leadbetter WF. Operative technique for correction of vesico-ureteric reflux. J Urol 1958; 79: 932–47. Fogo AB, Kon V. Pathophysiology of progressive renal disease. In: Barratt TM, Avner ED, Harmon WE, eds. Pediatric nephrology, 4th edn. Baltimore: Lippincott Williams & Wilkins, 1999; 1183–96.
1333
For personal use. Only reproduce with permission from The Lancet Publishing Group.