To re-evaluate the assumption that enterocystoplasty in children has a detrimental effect on linear growth (which is almost exclusively based upon a chance finding in a retrospective study 10 years ago) in a larger cohort and with a longer follow-up.
PATIENTS AND METHODS
The original 12 children who had impaired linear growth in a previous study 10 years earlier were re-measured. A larger cohort was identified from the 242 children and adolescents who had undergone enterocystoplasty between 1982 and 1997. Patients with conditions involving organ systems apart from the urinary tract, and those with myelomeningocele, malignant diseases, reduced glomerular filtration rate and incomplete notes were excluded. In the definitive study cohort (123; mean age at operation 8.6 years; mean age at investigation 16.8 years) enterocystoplasty had been undertaken using colon in 70, ileum in 37, a combination of both in 11, ileocaecal segments in three and stomach in two patients.
Of the original 12 patients, six had regained or surpassed their preoperative position on their growth charts. In all patients with a known target centile range the final height was within their genetic growth potential. In the cohort of 123 patients, 1215 height and weight measurements had been recorded. The distribution of percentile positions before and after enterocystoplasty showed a normal configuration, with 83% and 80% of patients growing within two standard deviations of the 50th percentile. After surgery, 85% either remained on the same or reached a higher centile. Nineteen (15.5%) were in a lower position, with a similar tendency in the weight centile. A clinically relevant growth disorder was recognized in four patients with a complete endocrinological evaluation; in none of these was enterocystoplasty thought to be a causal factor.
It is very unlikely that the loss of the preoperative percentile position on the growth curve in 15% of children after enterocystoplasty is a consequence of the surgery. Rather it is a non-specific phenomenon that has to be considered in any clinical population of the same size and age distribution after the same length of time.
In 1992, Wagstaff et al. reported a particularly worrying complication of enterocystoplasty in patients operated in the present authors' unit. They found delayed linear growth in 12 of 60 children whose urine was stored in pouches formed in whole or in part from bowel. Similar observations were reported by Mundy and Nurse  in three of six children who had undergone colocystoplasty, but in none of 10 who had undergone ileocystoplasty. In the same year, Koch et al. assessed patients with myelomeningocele managed by clean intermittent catheterization and an ileal conduit, concluding that linear growth was adversely affected by urinary diversion. In a case-controlled study, Gros et al. reported that growth in patients born with exstrophy was lower in those who had intestinal augmentation than in those who had successful primary closure of the bladder.
All of these studies, including ours, had some methodological flaws. As a diagnosis of abnormal growth by definition requires long-term monitoring , it seemed reasonable to re-assess the issue 10 years later in a larger cohort and with a longer follow-up.
PATIENTS AND METHODS
Permission was obtained from the local ethics committee for the study. The records of 242 patients who had undergone enterocystoplasty in childhood or adolescence at three institutions between 1982 and 1997 were retrieved and reviewed, including those of the earlier study . Patients with complex conditions involving organ systems apart from the urinary tract, and those with myelomeningocele, malignant diseases, a GFR of < 55 mL/min  and incomplete notes were excluded (119, Table 1). In the definitive study cohort (123, 90 males) enterocystoplasty had been undertaken using colon in 70, ileum in 37, a combination of both in 11, ileocaecal segments in three and stomach in two patients; 55 patients had a simultaneous or secondary Mitrofanoff procedure. The mean (range) age at operation was 8.6 (0.04–16.3) years, with 65% of children being < 10 years old. The underlying conditions are also detailed in Table 1.
Table 1. Exclusion criteria and underlying conditions
Number of patients
Myelomeningocele, sacral agenesis
VATER syndrome, anorectal anomalies,
GFR < 55 mL/min
Malignant disease (radio/chemotherapy)
Neurogenic bladder dysfunction
Ureterocele, ectopic ureter
Posterior urethral valves
All patients were followed prospectively according to a standardized protocol published previously . Height was measured prospectively since 1992 using standard techniques . Measurements were plotted on British Standard Growth Curves (Boys/Girls Birth-20 years 4-in-1 Decimal Growth Charts; Child Growth Foundation, 1996/1) [9,10].
As all raw data from the original study  were still available, the 12 children with an assumed growth delay could be identified and re-evaluated. To determine their individual genetic growth potential, mid-parental height and target centile range (the mid-parental height ± 8.5 cm in women, and ± 10 cm in men) were calculated where possible.
The patients underwent a total of 1045 surgical interventions (cystoscopy under anaesthesia excluded; mean 8.4, range 1–24), and in 72% more than five procedures. Data were calculated before and after the definitive enterocystoplasty. The mean (range) age at investigation was 16.8 (4.5–26.3) years with a mean follow-up of 8.3 (0.5–15.2) years after enterocystoplasty; 69% of patients had a follow-up of> 5 years. The mean number of height and weight measurements was four before and six after surgery, with 97 patients (79%) having> six measurements.
The distribution of percentile positions before (Fig. 1a) and after enterocystoplasty (Fig. 1b) showed a normal configuration, with 83% and 80% of patients growing within two sd of the 50th percentile. After surgery, 85% either remained on the same (53) or reached a higher centile (51). Nineteen patients (15.5%) over the whole range of the distribution curve (Table 2) showed a decrease in position, but 14 of these stayed above the ninth percentile. Similar tendencies in the weight centile were seen in 17 patients. Where X-rays of the hand had been taken, the bone age was appropriate.
Table 2. Distribution of percentile positions (height) in 19 children with a lower percentile position after enterocystoplasty
In 105 patients (85%) the serum creatinine level was normal at the last follow-up; 18 had a mild increase, but with a GFR stable over time. On at least one of the last three follow-up examinations 36 patients (29%) had a definite or borderline hyperchloraemic metabolic acidosis, and symptomatic UTIs were reported in 43 (35%). In 52 patients (42%) at least one bladder stone had to be removed. Five reservoir ruptures occurred in four patients, all requiring surgical management. There was no excess of cystoplasty-related complications in the group with lower growth percentile positions.
A clinically relevant growth disorder was recognized in four patients (3.3%), of whom one was too tall. A complete endocrinological evaluation in a growth clinic resulted in the diagnosis of hypogonadotrophic hypogonadism in one and constitutional delay of growth and puberty in two patients. In none of the patients was enterocystoplasty thought to be a causal factor. Growth-related problems were not addressed by patients, families or medical staff in any of the other cases.
The re-evaluation of the 12 patients from the original study  is detailed in Table 3. Ten patients had reached their final height, which was ≥ 25th percentile in nine. When compared with their percentile position in the original study, there was no change in two patients. Of the 10 patients who had improved their position, four had surpassed and two regained their preoperative percentile. In the seven patients where the mid-parental height and target centile range could be determined, the final height was within their genetic growth potential. In two patients a constitutional delay of growth and puberty was diagnosed. Most children had more than seven surgical interventions, including three with other forms of urinary diversion before enterocystoplasty. Two of the patients had had life-threatening complications, with prolonged illness after spontaneous perforation of their cystoplasty. Mild ectasia of the collecting system was frequent, although in all but one patient with no evidence of obstruction on renal scintigraphy. One patient with the VATER syndrome had hypertrophy of his left body half; he was not included in the definitive cohort of 123 patients.
Table 3. The details of the 12 original patients who were re-assessed
Chronic pyelonephritis; rupture of augmented bladder
Parenchymal scars left kidney
Anal atresia (colostomy); CRF, single kidney
Colon conduit since first days of life
Chronic pyelonephritis; small kidneys with PE
Bilateral renal scars; colon conduit since first days of life
Lower function of right kidney
Bilateral renal scars
Spontaneous rupture of augmented bladder
Chronic pyelonephritis; bilateral PE
Bilateral renal scars; Gersuny bladder since age 2 years
Linear growth results from a complex process of integrated physiological mechanisms and may be disturbed by several of these factors, acting either independently or together. Although there is no doubt that failure of normal growth can result from a wide variety of organic diseases, with a crude overall prevalence of 1 in 250–500 children measured , it may be extremely difficult to establish such a causal connection, particularly in a retrospective study. Apart from the well described pitfalls and imprecision in measuring height and calculating growth velocity , there is no exact definition of failure of normal growth in an individual because such a definition varies according to clinical and epidemiological needs. However, there seems to be a consensus that the shorter a child's stature, the more likely it is that he or she is failing to grow normally . While growth velocity is liable to misinterpretation, a diagnosis of abnormal growth requires long-term monitoring and is best seen as series of height measurements crossing the centiles on the height chart .
The assumption that the incorporation of bowel segments into the lower urinary tract of children may result in growth impairment has been based on four retrospective studies with numerous confounding factors [1–4]. Discussing the shortcomings of their study Wagstaff et al. noted that the effects of enterocystoplasty on growth were unexpected, a chance finding, and in the context of a retrospective study, may have been artefactual.
Mundy and Nurse  assessed calcium balance, growth and skeletal mineralization in 16 children who had undergone augmentation cystoplasty. While 10 children with an ileocystoplasty showed no change in growth pattern, three of six children with a colocystoplasty showed an average 20% reduction in growth potential. The authors commented that given the retrospective nature of their study and the few patients, the findings must be interpreted with caution . Moreover, it is unclear how growth potential and rate of growth were defined and determined in their patients. A further follow-up of the 16 children has never been published.
Of 93 children with myelomeningocele studied by Koch et al. only 47 underwent morphometric analysis (19 patients with an ileal conduit), and selection criteria for this subdivision were not given. Statistical differences between the groups were found in biacromial span percentile and elbow-hand length percentile, but not in height and weight. The significance of this outside this extremely complex population is unknown.
In the case-controlled study of Gros et al. delayed growth, as defined by a decrease in percentile height, was apparent in 14 of 17 children with exstrophy who had required a bladder augmentation. Unfortunately, 33 patients had to be excluded as they did not fulfil the inclusion criteria (at least one height measurement before and after operation). Even in the group with successful primary closure of an exstrophic bladder, a third of patients decreased in position on the growth chart, albeit by small percentiles.
In an extensive search of publications reviewing case series, with several thousand patients who had undergone classical or modified ureterosigmoidostomy, enterocystoplasty or continent diversion , growth was addressed in only three further articles. While Clementson-Kockum et al. reported unaffected growth in their elaborate prospective study of 15 children with bladder exstrophy or neurogenic bladder after bladder augmentation or continent urinary diversion, the Mansoura group reported conflicting results. In their first analysis of children with bladder exstrophy after urinary diversion to a low-pressure rectal reservoir in 1996, ‘growth-rate patterns’ were within normal limits using growth charts constructed for a Swiss population . Five years later the same group noted decreased linear growth (below the third percentile) in 57% after the same procedure and a comparable mean follow-up . However, in both studies the observations were based on one measurement of supine height at the last routine examination.
After the report by Wagstaff et al. in 1991, the ensuing awareness of potential growth impairment resulted in regular height and weight monitoring in the routine follow-up of our patients, eliminating some of the drawbacks inherent in a retrospective analysis. The transformation of height to its centile was thought to be appropriate as a reflection of growth, as most patients were within the normal range, where the calculation of growth velocity does not add significantly to the analysis . An average of 10 measurements per patient with a mean follow-up of 8.3 years matches the basic requirements for diagnosing abnormal growth. While the children's developmental stages were not recorded routinely at the respective times of height measurement, this information was available in all patients with a clinically relevant growth disorder.
To reduce confounding factors we excluded all patients with diseases and treatments (e.g. chemotherapy) which have a potential influence on growth, including chronic renal failure . Patients with myelomeningocele had been excluded for several reasons. First, significantly short stature is part of this abnormality with many contributing factors; second, the final height in these patients depends on the level of their neurological lesion, requiring further stratification, and third, morphometric measures such as arm span are not interchangeable with height in this population . The attempt to define the effect of enterocystoplasty on growth in such a complex setting was considered futile.
The major advantage of the present study is the sample size and the very many measurements, allowing the construction of percentile distribution curves, which closely resemble a normal Gaussian configuration before and after enterocystoplasty. The vast majority of children and adolescents were growing within the normal limits and either remained on their preoperative percentile or reached a higher position. To ascribe the latter to enterocystoplasty may be tempting, but is impossible to substantiate, especially as the patients had undergone an average of 8.4 operations.
Assessing the 19 children with a lower percentile position after surgery, most of the changes occurred above the ninth centile, and in eight of the 19 above the 25th centile. This inevitably raises the critical question of the clinical relevance of these phenomena, that are clearly not recognizable other than on the growth chart. In the four patients with a clinically obvious growth disorder a meticulous endocrinological evaluation in a growth clinic excluded enterocystoplasty as a cause.
This re-evaluation of these patients highlights the complexity of the group and the wide range of confounding factors, rendering it impossible to establish valid causal connections. However, it is notable that even under complicated circumstances the decrease in percentile position was reversible in most patients.
Placing these findings in perspective, the relative importance of growth failure from the patient's view must be considered. A recent study showed that growth is a major concern for 30% of parents and 28% of children with short stature caused by chronic renal failure . These proportions may be even less in bladder exstrophy, where urinary continence, genital appearance, fertility and sexual function are competing issues .
In conclusion, it is very unlikely that the loss of the preoperative percentile position on the growth curve in 15% of children after enterocystoplasty is a consequence of the surgery. Rather it is a non-specific phenomenon that must be considered in any clinical population of the same size and age distribution after the same length of time. This does not mean that monitoring growth should be abandoned altogether in the follow-up of these children, especially as concomitant renal impairment may limit the ability to compensate for the metabolic consequences of enterocystoplasty over time.
Supported by grants from Deutsche Forschungsgemeinschaft (GE 973/1–1), Bad Godesberg, Germany, and the Incontinence Research Trust, London, England.