World J Urol (1998) 16: 242±250

Ó Springer-Verlag 1998

Eric A. Kurzrock á Laurence S. Baskin á Barry A. Kogan

Gastrocystoplasty: is there a consensus?

Abstract The problems encountered with ileal and colocystoplasty have led to the use of the stomach for bladder augmentation, termed gastrocystoplasty. The advantages of gastrocystoplasty over intestinal segment augmentation include reduced chloride reabsorption, decreased mucus production, decreased urinary infection in the presence of acid urine, extremely low incidence of stones, and avoidance of complications from short bowel syndrome. The gastric patch provides comparable improvements in bladder volume, pressure, and continence. The thick muscular wall of the stomach facilitates ureteric reimplantation as compared with the small intestine, but the rate of stenosis and re¯ux may not be superior. The disadvantages of the gastric patch include complications of severe systemic alkalosis, which is usually manifest in dehydrated, renal compromised patients, and the hematuria-dysuria syndrome (HDS), which is more prevalent in patients with renal insuciency, normal pelvic sensation, and urinary incontinence. The postoperative complication rate of gastrocystoplasty is comparable with that of other augmentation procedures and similarly warrants proper selection and close follow-up of patients. In this report we review the literature and present the results, including a discussion of the technique and the pathophysiology of its complications.

dynamics, augmentation cystoplasty becomes a reliable alternative to ensure low-pressure storage of urine for upper-tract preservation and continence. The use of ileum and colon, i.e., intestinocystoplasty, for augmentation has led to problems such as hyperchloremic metabolic acidosis, poor emptying secondary to mucus, urinary infection, diarrhea, and bladder stones [6, 37, 38]. In contrast to the intestine, the stomach wall acts as a barrier to ammonium and chloride absorption; it excretes acid, produces less mucus, and provides relatively better microbial resistance [25, 35, 45]. Gastrocystoplasty seems to be particularly advantageous for patients with renal insuciency and acidosis and children with cloacal exstrophy and de®cient bowel [1]. Despite the bene®cial physiology of the stomach, systemic and local complications from the aciduria have arisen. This leads to the question as to the ideal candidate for gastrocystoplasty. We reviewed the literature on gastrocystoplasty in children. Manuscripts reporting on complete bladder substitution in adults were excluded. Some of the pediatric series reviewed included bladder substitutions, but in all reports these were a small fraction of the cases cited. Table 1 illustrates some of the more salient points of the subject; these and other details gained from the review are discussed herein [1, 3, 4, 7, 12, 16, 19, 47, 48]. In addition, we present the relevant anatomical and technical aspects of gastrocystoplasty.

Since the introduction of clean, intermittent catheterization as a practical and safe method of bladder emptying [28], urinary diversion has lost favor to bladder optimization for children with bladder dysfunction. When medical management fails to improve bladder

History

E. A. Kurzrock á L. S. Baskin (&) á B. A. Kogan Pediatric Urology, Department of Urology, U-575, University of California, San Francisco School of Medicine, San Francisco, CA 94143-0738, USA

Sinaiko [49] was the ®rst author to describe the use of stomach for urinary diversion in 1956. Creating a urinary pouch from stomach fundus and corpus (body) in dogs, he demonstrated the ability of the transplanted stomach wall to secrete acid in response to meals and also found the acidic urine to be bacteriostatic [49]. Two decades later, Leong and Ong cited promising results using the antrum of the stomach in dogs [33] and then humans [34]. Mitchell and Piser [38] were the ®rst

243

Fig. 1 The wedge-shaped segment of stomach is based upon the right gastroepiploic artery. Division of short branches from the proximal artery to the wedge allows adequate mobilization

Anatomy and physiology

Fig. 2 The gastric patch is mobilized through transverse colon and ileocecal mesenteries. After it has been sutured to the bladder the mesenteric traps are closed and omentum is used to cover the suture lines on the bladder

The stomach receives a rich arterial supply from the right and left gastric and right gastroepiploic branches of the hepatic artery, and from the left gastroepiploic and short gastric branches of the splenic artery [21]. The majority of acid production takes place in the parietal cells of the body of the stomach, although parietal cells can also be found in the pyloric antrum. The parietal cells in the gastric glands are stimulated by a hormone,

investigators to describe gastrocystoplasty in children. They introduced technical modi®cations that incorporated more body and less antrum into the patch [38]. Table 1 Review of pediatric gastrocystoplasty First author, city, year

Pts. (n)

Mean Mean f/u NGB age (months) pts. (years)

Adams [1], Indiana, 1988 Dykes [12], London, 1992

13 8

6.4 10

13 21

Atala [3], Boston, 1993

20

10

Gosalbez [19], Atlanta, 1993 Ganesan [16] and Nguyen [41], Seattle, 1992 and 1993 Bogaert [7], S. F., 1994

30

9

PUV pts.

Exstrophy pts.

Other pts.

Renal Preop Postop Preop insuciency volume volume end-®ll pts. (cm3) (cm3) pressure (cmH2O)

5 0

2 5

5 1

1 2

6 (46%) 8 (100%)

122 106

275 324

80 54

<36

13

1

2

4

N/A

160

352

45

14

21

1

8

0

2 (7%)

130

340

67

[41]

73

14.1

26.5

24

8

28

13

10 (18%)

N/A

N/A

N/A

28

9

29

23

1

3

1

18 (64%)

N/A

N/A

N/A

Sheldon [48], Cincinnati, 1995 Benedetto [4], Marseille, 1997

23

6.1

36

5

3

7

8

6 (26%)

78

304

N/A

20

10.6

42

0

1

19

0

N/A

N/A

N/A

N/A

Total or weighted mean

215

11

<27

73 (34%)

29 (14%)

50/159 (31%)

121

323

62

91 (42%)

22 (10%)

244

gastrin, which is produced by the argentan G-cells found mainly in the pyloric region and duodenum [5]. Gastrin release is stimulated by gastric distention, vagal stimulation, or antral alkalization [10, 14]. The parietal cells can also be stimulated by histamine via paracrine action and acetylcholine via neurocrine action [14].

Technique As preparation, the child should be placed on a liquid diet for 2 days, then undergo mechanical bowel cleansing and be given culture-speci®c antibiotics. After induction of anesthesia the child is given a histamine-2 (H2) receptor-blocking agent intravenously. If there is any possibility of latex allergy a histamine-1 blocker should also be given. Via a midline transperitoneal incision from the xiphoid to the pubis, both the stomach and the bladder are exposed. To prevent yeast or anaerobic infection, before being opened the stomach is irrigated with betadine solution followed by multiple ¯ushes of sterile water via the nasogastric tube. The stomach is dissected ®rst so as to limit contamination of the abdomen by infected urine. Prolonged leakage of acidic secretions from the stomach are prevented by the H2-blocker. After inspection of the vascular arcade, either the left or the right gastroepiploic artery is selected for the augmentation. In most patients the right gastroepiploic artery is used because it is generally of larger caliber and easier to dissect to its origin at the gastroduodenal artery. More importantly use of the right gastroepiploic artery makes it highly unlikely that any antrum will be used in the augmentation.

As much omentum as possible is mobilized on the vascular pedicle for later use to cover the augmentation. With particular attention being paid to avoid the antrum, a 12- to 15-cm length of stomach body is then delineated along the greater curvature (Fig. 1). To avoid the inclusion of antrum in the gastric segment an imaginary vertical line is drawn from the angularis incisura. This line also marks the point where the distance between the right gastroepiploic artery and the stomach diminishes distally. Although there is no absolute anatomical landmark, designing the gastric segment proximal to this line will likely include only the body of the stomach. Congo red mapping of the gastric mucosa may also identify the di€erence in the pH of the mucosa between the body and the antrum of the stomach, but this di€erence in pH will be present only in children who do not receive an H2blocker before the surgery. For development of the pedicle based upon the right gastroepiploic artery, the branches between the artery and the greater curvature of the stomach are ligated and divided up to the right side of the delineated patch (Fig. 1). On the left side of the segment the left gastroepiploic artery is dissected and ligated after inspection to con®rm adequqte ¯ow from the right. A triangular wedge of stomach is delineated with bowel clamps. Care is taken to preserve the lesser curvature and the vagus nerve. The segment to be used for the augmentation is divided from the rest of the stomach using electrocautery. The stomach is then closed using a continuous suture of chromic catgut on the inner layer, supported by interrupted silk overlapping sutures. The gastric segment is then mobilized through both the transverse colon and the ileocecal mesenteries (Fig. 2). The

Table 1 (Contd.) Postop MildSevere end-®ll moderate HDS pts. pressure HDS pts. (cmH2O)

Fasting Asymptomatic gastrin bacteruria pts. >150 pg/ml, pts.

Serum chloride change (mEq/l)

Serum bicarb. change (mEq/l)

Symptomatic metabolic alkalosis pts.

Major complications and reoperations

40 25

0 1

0 1

0/13 (0) 0/6 (0)

4 N/A

)3 1

4 4.5

1 (8%) 1 (13%)

39

4

0

N/A

N/A

N/A

N/A

1 (5%)

27

0

0

N/A

N/A

)2

1.5

2 (7%)

N/A

18

3

2/14 (14%)

N/A

N/A

N/A

9 (12%)

<20

1

0

0/14 (0)

5

)1.5

2.5

2 (7%)

N/A

2

0

N/A

N/A

N/A

N/A

0 (0)

N/A

8

2

N/A

0

N/A

N/A

0 (0)

<29

34

6

2/47

9/61

)2

+3

16 (7%)

Redo reimplant (2) Reduction of gastric patch for HDS (1995 manuscript) Repair of bladder perforation; postop gastric bleed Postop subphrenic abscess; postop gastric hemorrhage; late ileal patch due to alkalosis GI bleed, urinary extravasation, SBO (7), bladder perforation, augment revision (3) Small-bowel gastric-pouch ®stula; partial SBO Intraop latex reaction; VP shunt malfunction SBO (2); anastomotic dehiscence; redo reimplant (2)

(17%)

(3%)

(4%)

(15%)

Pts., patients; NGB, neurogenic bladder; PUV, posterior urethral valve; HDS, hematuria dysuria syndrome; S. F., San Francisco; N/A, not available

245

mesenteric defects should be closed to prevent internal herniation of bowel. Next the bladder is opened from the bladder neck anteriorly to the trigone posteriorly. Adjunctive procedures, for example, ureteric reimplantation, placement of an arti®cial sphincter, or creation of an apendicovesicostomy, are performed at this time. The gastric patch is sewn in place between the bladder halves using a single-layer running chromic suture. Omentum is wrapped around the bladder to cover the augmentation and the suture line. Nasogastric suction should be maintained until the ileus resolves. The bladder is drained for 3 weeks with a large-caliber (generally 20- to 24-F) suprapubic catheter, after which intermittent catheterization is performed. Patients are advised to consume small, frequent meals in the early postoperative period. Acid secretion is suppressed using H2-blocker medication for 3 months.

Vesical function and urodynamics The stomach does not have peristaltic waves similar to those of the tubular intestine; its contractions are designed to churn and agitate the contents for digestion while propulsing them forward. Since detubularization is not necessary to decrease contractions, the diamondshaped gastric patch is easily adaptable to the bladder opening and a€ords the creation of a more spherical reservoir than that a€orded by the intestine [3]. This allows a decrease in redundancy and folds, which may facilitate volitional bladder emptying in valve patients [24, 48] or lessen problems in emptying with a catheter in neurogenic patients. Because the gastric patch is not detubularized, bowel-bowel suture lines can be avoided, theoretically optimizing the vascularity of the patch [3]. The thicker muscle, better vascularity, and more spherical shape of the graft should provide for a stronger augmentation and, possibly, lower the incidence of rupture. The ultimate goal of augmentation is to enable storage of suitable volumes of urine at a low pressure to ensure upper-tract preservation and promote continence. Studies that reported pre- and postoperative urodynamics showed average increases in bladder volume of greater than 200 cm3 (150%). Mean end-®ll pressures were decreased to less than 30 cmH2O (Table 1) L. Atala et al. [3] noted complete resolution of preoperative uninhibited contractions in six of ten patients, with three having persistent contractions of above 30 cm-H2O postoperatively. We noted uninhibited contractions in four patients, three of whom had resolution with treatment of urinary infection [7]. In total, persistence of uninhibited contractions of greater than 30 cm were encountered in 6 of 78 (8%) patients tested [3, 7, 19]. This compares favorably with the rates of 10%, 10%, and 33% reported in ileal, sigmoid, and ileocecal patches, respectively [38]. Distinct from uninhibited contractions, Atala et al. [3] and Gosalbez et al. [19]

found low-amplitude, rhythmic contractions of the gastric segment of 20±30 cmH2O in up to half of the patients studied. Excluding patients with end-stage renal disease, no study has reported worsening renal function or hydronephrosis secondary to the augmentation. Continence was signi®cantly improved in all studies. Since almost half of the patients underwent Mitrofano€ diversion or bladder outlet procedures, improvements in continence cannot be attributed to the augmentation alone, nor can one make judgements about the superiority of gastrocystoplasty versus ileocystoplasty relative to continence. It is safe to say that if either bowel provides a low-pressure, capacious reservoir, then the promotion of continence should be equal.

Metabolic properties Augmentation of the bladder with intestine carries a signi®cant risk for hyperchloremic, hypokalemic, metabolic acidosis, especially in children with impaired renal function [37, 38, 43]. Mitchell and Piser [38] reviewed 129 cases of intestinocystoplasty and found signi®cant changes in chloride in all patients. The renal failure patients' mean postoperative serum chloride and bicarbonate values were 113 and 16 mEq/l, respectively [38]. Mild chronic acidosis and consequent calcium loss from intestinal diversion has been related to osteomalacia, rickets, decreased growth, and increased orthopedic morbidity. Loss of distal ileum may also result in fat malabsorption and decreased bile-salt absorption and absorption of fat-soluble vitamins [37]. These problems are especially concerning in children with cloacal exstrophy, who are at high risk for short-bowel syndrome. In contrast to the intestine, the stomach mucosa acts as a barrier to chloride and ammonium absorption and actually excretes chloride [45]. Evaluation of dogs after gastrocystoplasty [25] and follow-up of children after gastrocystoplasty have shown persistent but small decreases in serum chloride and increases in serum carbon dioxide (Table 1). Because of these unique properties, gastrocystoplasty has been advocated in patients with compromised renal function and acidosis, since the protection against hyperchloremic acidosis is particularly advantageous in this group and may even be therapeutic [1]. Sheldon et al. [48] reported on ®ve azotemic patients who were capable of discontinuing bicarbonate supplementation after gastrocystoplasty. Although the aciduria and chloride loss of the gastric mucosa is advantageous in most patients, some are at risk for a syndrome of hypochloremic, hypokalemic, metabolic alkalosis. Our review of the series that cited this problem found that 16 (7%) patients developed this syndrome severely enough to warrant hospitalization (Table 1). In most cases the alkalosis was preceded by a dehydrated state, usually from viral enteritis. Renal insuciency appears to be a signi®cant risk factor. Some cases were exacerbated by diuretics or bicarbonate

246

therapy. There is some controversy as to whether dehydration is necessary to elicit this syndrome and what role hypergastrinemia plays in its etiology. Adams et al. [1] postulated that obligate ¯uid and salt losses secondary to the urinary concentrating defect present in patients with renal insuciency compounded by acid excretion from the gastric patch may lead to dehydration and metabolic alkalosis when adequate ¯uid intake cannot be maintained. On the other hand, McDougal [37] believes that the decreased ability for bicarbonate excretion by the severely impaired kidney is primarily responsible for the development of metabolic alkalosis in these patients. Of the alkalotic patients reported in the literature (but not necessarily in Table 1) who had their fasting gastrin level evaluated during the episode, seven of seven had evidence of hypergastrinemia [19, 26, 46]. Presumably, the elevated level of serum gastrin stimulated acid (chloride) loss from the bladder despite the marked hypochloremia. This was supported by the elevated fractional excretion of chloride observed in these patients. Before prematurely attributing the syndrome to hypergastrinemia, one must be aware that some of these patients had only mild elevations in gastrin and the majority were on H2-blockers, which are known to cause hypergastrinemia [15, 31]. Plawker et al. [46] presented two patients who developed severe metabolic alkalosis without laboratory evidence of dehydration. One patient had a history of renal insuciency and both had evidence of hypergastrinemia. Gosalbez et al. [20] described the ®rst two patients with normal renal function (based on serum creatinine values) to develop metabolic alkalosis preceded by gastrointestinal ¯uid loss (vomiting and diarrhea). Since both of their patients presented with an elevated fractional excretion of chloride despite profound hypochloremia, these authors postulated that the continued secretion of hydrochloric acid from the gastric patch rather than an inability of the renal tubule to conserve chloride was the causative factor [20]. More likely, their patients did have some de®ciency in urinary concentrating ability despite normal creatinine values, since both had a history of re¯ux and urinary diversion. A multifactorial cause of this problem is supported by the pathophysiology. Dehydration can produce a cascade of events that are further potentiated by renal insuciency and gastrin-stimulated chloride loss from the gastric patch. Intravascular volume depletion leads to a decreased glomerular ®ltration rate and increased distal tubule sodium resorption in exchange for potassium and hydrogen ions. In addition, hypovolemia results in renin release and increased aldosterone production, further enhancing distal sodium resorption and potassium loss [20]. Potassium depletion, when severe, can sustain alkalosis initiated by acid loss by increasing the rate of renal tubular bicarbonate reabsorption. Thus, if the kidney is incapable of responding to the dehydration, a pathologic cycle be-

gins that can be further aggravated by the excessive loss of chloride through the augmentation. Treatment of metabolic alkalosis has included intravenous replacement of sodium chloride, potassium chloride, and H2-blockers. Omeprazole, which blocks the hydrogen-potassium-adenosine triphosphatase pump in the parietal cell, has proved to be e€ective in unresponsive cases. Omeprazole should be used with caution in children; its dosing should be limited to short courses of 2±4 weeks. There are concerns regarding its long-term use, and it should be given under the direction of a gastroenterologist if needed beyond 4 weeks [26]. Gosalbez et al. [20] and Ganesan et al. [16] reported on one patient each with symptomatic alkalosis who was unresponsive to medical therapy and required gastric patch replacement with ileum.

Hematuria dysuria syndrome and acid secretion The hematuria dysuria syndrome (HDS) is characterized by red or co€ee-brown urine, suprapubic pain, skin irritation, and/or burning upon urination. Our review of the literature found the incidence of mild to moderate symptoms to be 17% and that of severe symptoms to be 3%. Only 2 of 40 (5%) patients did not obtain relief from medication. In the severest forms of the syndrome, ulcers can form in the bladder. Ironically, the relationship of this syndrome to aciduria is sporadic and illde®ned. The largest study on this syndrome was reported by Nguyen et al. [41], who followed 57 patients for a mean of 23 months. In all, 27 (36%) had symptoms consistent with HDS. The most common symptoms were co€eebrown or bright red hematuria (16 patients, 76%), bladder or suprapubic pain (15 patients, 71%), skin irritation (8 patients, 38%), and dysuria (5 patients, 23%). The severity of symptoms was ranked as mild in 8 patients (38%), moderate in 10 (48%), and severe in 3 (14%). Only 5 patients were on continuous medication to control symptoms. All together, 6 of 10 patients (60%) with renal insuciency in this series reported symptoms as compared with 15 of 47 individuals (31%) with normal renal function (P ˆ 0.148). Of the 12 incontinent patients, 8 (66%) had symptoms of HDS as compared with 13 of 45 continent patients (28%). In our review, 31 of the 40 (78%) reported patients with HDS came from the series of Nguyen et al. [41] and Benedetto and Monfort [4], which had the highest proportion of nonneurogenic patients. On the other hand, the series of Bogaert et al. [7] and Gosalbez et al. [19] with the lowest incidence of HDS (2%) had a large proportion of neurogenic patients. There appears to be a strong relationship between HDS and patients who are incontinent, void per urethra, and/or have normal pelvic sensation. In the severest form of the syndrome, ulceration of the native bladder can occur. Reinberg et al. [47] reported on a child with chronic renal failure and a

247

defunctionalized bladder (no urine) who had a spontaneous perforation due to a peptic ulcer. An adult patient with radiation cystitis formed an ulcer in the native bladder after gastrocystoplasty [39]. It is noteworthy that this patient was on oral bicarbonate, which stimulates gastrin secretion and consequently increases acid production by the patch [51]. Surprisingly, there is not a strong relationship between aciduria and HDS. In the study of Nguyen et al. [41] the pre- and postprandial urinary pH of symptomatic patients was evaluated at home and found to be 6.1 and 6.0, respectively. Only 5 patients had a urinary pH of less than or equal to 3, and only 2 of the 14 patients tested had elevated gastrin levels [41]. However, urinary pH has not been shown necessarily to re¯ect the actual bladder mucosal pH (using an intravesical probe) [12, 46]. In summary, mild to moderate symptoms of HDS are reported in 17% of patients and severe forms, in 3%. Symptoms can be controlled with medication in the vast majority. Patients with incontinence, renal insuciency, and normal sensation are at higher risk of having symptoms. Anuria and radiation may predispose to ulceration. The relationship of HDS and aciduria is unde®ned, possibly indicating other causative factors such as stomach enzymes.

Hypergastrinemia As mentioned previously, gastrin release is stimulated by gastric distention, vagal stimulation, or antral alkalization [10, 14]. Theoretically, antral mucosa in the bladder could secrete gastrin in response to bladder distention. Since the antrum is not large, creation of a neobladder or cystoplasty from antrum would likely include parietal cells from the adjacent gastric body as well as antral parietal cells. Thus, in e€ect, the antral gastrocystoplasty will self-stimulate acid secretion in response to bladder distention. On the other hand, a patch of pure stomach body, excluding G-cells, should not respond to bladder distention and will secrete acid only in response to gastrin released from the native stomach after a meal. To verify this we extensively evaluated 13 children who had gastrocystoplasty with stomach body and special attention to avoid the inclusion of antrum. We demonstrated that the gastric body patch continues to be regulated in the same way as the native stomach body [8]. Normally, after a meal there is an ``alkaline tide'' in the urine. After gastrocystoplasty, these children's fasting level of serum gastrin was normal (mean 78 ng/l) and their fasting urinary pH was neutral (mean 7.8) with no titratable acid. After meals, serum gastrin levels (mean 253 ng/l) increased and urinary pH (mean 4.8) decreased markedly. As the gastric patch no longer had vagal innervation, it is likely that the gastrin increased after meals induced the acid secretion by the patch similar to the case for native stomach mucosa. In the same patients

we demonstrated that bladder catheterization and distention did not result in urinary acid secretion or gastrin elevation. It has been suggested that hypergastrinemia may be due to loss of the inhibitory e€ect of gastric acid, which has been diverted from the antral mucosa, and may lead to hyperplasia of the remaining gastrin cells. Plawker et al. [46] found no evidence of hyperplasia in two patients with hypergastrinemia as evaluated by upper endoscopy and biopsy. To evaluate the e€ect of antral mucosa placement in the urinary tract, Lim et al. [36] retrospectively studied 13 adults after antral neobladders or gastrocystoplasty and found no signi®cant di€erence between preoperative and postoperative fasting serum gastrin levels. However, the postoperative, postprandial gastrin response was signi®cantly lower and the acid output from the native stomach was lower. This is consistent with general surgery studies showing a decreased acid output and gastrin response to a food stimulus after antrectomy [2]. In the case of antral gastrocystoplasty, this suggests that the antral bladder patch might contribute to the maintenance of basal serum gastrin through bladder distention but does not respond to a food stimulus in the stomach. Unfortunately, these authors did not study the e€ect of bladder distention on gastrin release or acid secretion. Earlier studies in dogs have shown con¯icting results. Ti€any et al. [50] reported on six dogs; all four dogs with antral patches developed hypergastrinemia and two developed ulcers. On the other hand, neither of two dogs with fundic patches developed ulcers or hypergastrinemia [50]. These results were in con¯ict with prior dog studies showing no hypergastrinemia after antral gastrocystoplasty [29, 32, 33]. Lau et al. [29] found that urine in contact with antral mucosa suppressed gastrin secretion in dogs. In summary, stomach body mucosa placed in the bladder continues to be regulated in the same manner as native stomach. Gastrin released by meals will stimulate acid production in the bladder patch. Bladder distention does not a€ect the patch. On the other hand, the physiology of antral patches is controversial, except that antrectomy does cause gastrin elevation. Since both types of patches might incorporate gastrin-producing cells and acid-producing cells in varying amounts, their comparison is imperfect.

Mucus, infection, and stones The glands of the stomach also produce mucus [5]. Animal models of gastrocystoplasty [27, 35] and multiple clinical series [1, 3, 7, 40] have noted lower mucus production following gastric as compared with intestinal augmentation. The decreased mucus production minimizes problems of bladder emptying by spontaneous voiding or catheterization. Less residual urine, decreased mucus, and the acidic urine have been proposed as mechanisms to explain the lower incidence of bacilluria

248

and symptomatic infections in gastric versus intestinal augmentation [1, 3, 7, 27, 35, 40]. In our review, 3 series studied the rate of infection and reported 9 of 61 patients (15%) with asymptomatic bacteruria [1, 4, 7]. After intestinocystoplasty the rate of bacilluria approaches 100% [38]. Enteric augmentations tend to lower the urinary urea concentration and increase the pH. In contrast to intestinal patches, gastric patches decrease the urinary pH and induce a smaller decrease in urea concentration [30]. This is associated with a lower incidence of bacterial colonization as compared with ileocystoplasty in animal models [30, 35] and with a lower rate of bacteruria in humans [1, 4, 7]. Bacteruria and mucus have been implicated in the formation of stones after urinary diversions with intestine [37, 44]. The presence of alkaline urine secondary to infection with urea-splitting organisms and hypercalciuria secondary to chronic metabolic acidosis are purported risk factors [37]. After ileocystoplasty the rate of formation of calculi is reported to be as high as 52% [6, 44]. In contrast to intestinocystoplasty, our review of gastrocystoplasty series (Table 1) found only one report of stone formation after augmentation. Garzotto and Walker [17] described a uric acid stone forming in a child who had a colostomy, acidic urine, and elevated uric acid and calcium excretion. As most of the stones reported in ileocystoplasty patients formed within 4 years, longer follow-up will probably validate these excellent short-term results of gastrocystoplasty.

Ureteral reimplantation In contrast to the small intestine, many authors cite the uniformly thick muscle of the stomach to be advantageous for antire¯ux implantation of ureters or e€erent limbs when adequate bladder wall is not available [1, 19, 48]. In our review of cystoplasties and pouches requiring ureteral-gastric anastamosis we found 6 of 24 implants to be su€ering from re¯ux or stenosis [1, 19, 40]. Despite its ease, whether the stomach wall o€ers any advantage over the intestine for reimplantation remains to be decided.

Cancer Development of adenocarcinoma and, rarely, transitional-cell carcinoma is well documented after ureterosigmoidostomy [23]. Tumors have been reported following other diversions, including ileocystoplasty, with manifestations being observed up to 50 years after surgery [13, 22]. It is evident that for tumors to develop, the urothelium must be adjacent to the intestinal mucosa, and contact with feces substantially increases the risk [37]. Ngan et al. [40] performed cystoscopy and biopsy on eight adult patients after gastric neobladders with a median follow-up of 10 years. They found atrophy in four cases, intestinal metaplasia in one case, and a mild

to moderate degree of polymorphonuclear and plasma cell in®ltration in all biopsies. No dysplasia, squamous metaplasia, or papilloma formation was evident [40]. We performed cystoscopy and biopsy of stomach and bladder mucosa after gastrocystoplasty in ®ve patients and found no evidence of dysplasia or hyperplasia. Mild in¯ammation was found in two bladders [7, 8]. Buson et al. [9] studied mucosal changes in rats after augmentation with ileum, colon, or stomach. They found super®cial transitional metaplasia in all types of bowel, although papillary hyperplasia was found only in gastric and sigmoid patches [9]. In summary, all types of bowel placed in chronic contact with urine pose a risk of carcinogenesis. With short-term follow-up, no tumor has been reported after gastrocystoplasty. This by no means demonstrates a lower risk of carcinogenesis and will require routine surveillance similar to that needed after intestinocystoplasty.

GI side effects The GI complications of gastrocystoplasty are similar to those of other urinary diversion procedures, including rare ®stulae (1), anastomotic dehiscence (1), hemorrhage (3), and small-bowel obstruction (10). Inappropriate resection of the stomach can lead to chronic, early satiety; injury to the vagal nerve; or malfunction of the gastroesophageal mechanism. Neither we nor other investigators of large series have found problems with gastric ulcers or re¯ux [1, 3, 7, 48]. Gold et al. [18] reported on ®ve patients referred to the gastroenterology service with weight loss (maximum 5±10%) at 4±6 months after surgery. Two patients had dumping syndrome, delayed gastric emptying, and esophagitis. All patients were successfully treated with medical therapy [18]. Sheldon et al. [48] found early transient weight loss in 3 of 6 azotemic patients versus 2 of 15 nonazotemic patients. Plawker et al. [46] demonstrated signi®cant gastroesophageal re¯ux and distal esophagitis in two patients who presented with metabolic alkalosis. Transient intolerance of food is common [11]. As early satiety is to be expected immediately after wedge resection, we advise all our patients to consume small, frequent meals in the early postoperative period. As compared with ileal resection, gastrocystoplasty has some nutritional advantages as discussed previously and is useful in patients at risk for short-bowel syndrome. Careful attention to detail and close follow-up should prevent most GI complications.

Autoaugmentation gastrocystoplasty Since the majority of complications of augmentation are related to the transplanted mucosa, some centers are attempting to preserve the native urothelium, and

249

replace the excised detrusor with demucosalized intestine or stomach. Autoaugmentation gastrocystoplasty in humans has had promising preliminary results, including uptake of urothelium and the absence of complications secondary to aciduria [11, 42]. Urodynamic improvements have not been as impressive [11], and complications from nonadherence of the bladder mucosa to the muscular graft have been noted [11, 42]. The stomach has been advocated for this technique because demucosalization is less challenging in that organ than in the intestine [11, 42]. This technique has many theoretical advantages, including decreased metabolic complications, stone formation, and carcinogenesis, but there are also concerns of potentially increased carcinogenesis. Long-term results are anxiously awaited.

Conclusion The advantages of gastrocystoplasty over intestinal segment augmentation include reduced chloride reabsorption, decreased mucus production, decreased urinary infection in the presence of acidic urine, extremely low incidence of stones, and avoidance of complications from short-bowel syndrome. The gastric patch provides comparable improvements in bladder volume, pressure, and continence. The thick muscular wall of the stomach facilitates ureteric reimplantation as compared with the small intestine, but the rate of stenosis and re¯ux may not be superior. Severe systemic alkalosis is reported in 7% of patients and is usually manifest in dehydrated, renal compromised patients. The relationship of this syndrome to hypergastrinemia is evident but not uniform; a multifactorial cause is probable. This syndrome may be avoidable via special instructions for high-risk patients. The hematuriadysuria syndrome presents in about 20% of patients and is more prevalent in patients with renal insuciency, normal pelvic sensation, and urinary incontinence. Over 90% of patients with the syndrome have symptoms that are transient in nature or controlled by medication. Short-term follow-up has not shown an increase in carcinogenesis following gastrocystoplasty as compared with intestinal augmentation, but similarly close surveillance is required in both cases. The postoperative complication rate is comparable with that of other augmentation procedures. The preliminary results of autoaugmentation gastrocystoplasty are promising. Gastrocystoplasty is a useful procedure that provides a lower rate of infection and stone formation but can cause two unique syndromes of metabolic alkalosis and hematuria-dysuria, which can be avoided and medically treated with proper patient selection and follow-up.

2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

References 1. Adams MC, Mitchell ME, Rink RC (1988) Gastrocystoplasty: an alternative solution to the problem of urological recon-

26.

struction in the severely compromised patient. J Urol 140: 1152±1156 Ashby DB, Himal HS (1977) E€ect of antrectomy and vagisecretion on gastric acid output and gastrin secretion. Am J Gastroenterol 68: 556±559 Atala A, Bauer SB, Hendren WH, Retik AB (1993) The e€ect of gastric augmentation on bladder function. J Urol 149: 1099±1102 Benedetto VD, Monfort G (1997) Stomach versus sigmoid colon in children undergoing major reconstruction of the lower urinary tract. Pediatr Surg Int 12: 393±396 Bloom W, Fawcett D (1975) The esophagus and stomach. A textbook of histology. Saunders, Philadelphia, pp 643±657 Blyth B, Ewalt DH, Duckett JW (1992) Lithogenic properties of enterocystoplasty. J Urol 148: 575±577 Bogaert GA, Mevorach RA, Kogan BA (1994) Urodynamic and clinical follow-up of 28 children after gastrocystoplasty. Br J Urol 74: 469±475 Bogaert GA, Mevorach RA, Kim J, Kogan BA (1995) The physiology of gastrocystoplasty: once a stomach, always a stomach. J Urol 153: 1977±1980 Buson H, Diaz DC, Manivel JC, Jessurun J, Dayanc M, Gonzalez R (1993) The development of tumors in experimental gastroenterocystoplasty. J Urol 150: 730±733 Davison JS (1989) Gastric secretion. Wright, London Dewan PA, Stefanek W (1994) Autoaugmentation gastrocystoplasty: early clinical results. Br J Urol 74: 460±464 Dykes EH, Ransley PG (1992) Gastrocystoplasty in children [see comments]. Br J Urol 69: 91±95 Egbert BM, Kraft JK, Perkash I (1980) Undi€erentiated sarcoma arising in an augmented ileocystoplasty patch. J Urol 123: 272±274 Feldman M (1989) Gastric secretion in health and disease. In: Sleisenger MH, Fordtran JS (eds) Gastrointestinal disease, pathophysiology, diagnosis, management. Saunders, Philadelphia, pp 713±734 Fraser AG, Sawyerr AM, Hudson M, Smith M, Pounder RE (1994) E€ects of ranitidine 150 mg four times a day on 24-hour intragastric acidity and 24-hour plasma gastrin concentration. Dig Dis Sci 39: 91±96 Ganesan GS, Mitchell ME, Nguyen DH (1992) Bladder reconstruction using stomach: 73 patients and 6 years later. J Urol 147: 253A Garzotto MG, Walker RD (1995) Uric acid stone and gastric bladder augmentation. J Urol 153: 1976±1979 Gold BD, Bhoopalam PS, Reifen RM, Harvey E, Marcon MA (1992) Gastrointestinal complications of gastrocystoplasty. Arch Dis Child 67: 1272±1276 Gosalbez R Jr, Woodard JR, Broecker BH, Parrott TS, Massad C (1993) The use of stomach in pediatric urinary reconstruction. J Urol 150: 438±440 Gosalbez R Jr, Woodard JR, Broecker BH, Warshaw B (1993) Metabolic complications of the use of stomach for urinary reconstruction. J Urol 150: 710±712 Gray H (1985) The digestive system. In: Clemente CD (ed) Anatomy of the human body. Lea and Febiger, Philadelphia, pp 1463±1469 Hasegawa S, Ohshima S, Kinukawa T, Matsuura O, Takeuchi Y, Hattori R, Murakami S (1989) Adenocarcinoma of the bladder 29 years after ileocystoplasty. Hinyokika Kiyo 35: 671±674 Husmann DA, Spence HM (1990) Current status of tumor of the bowel following ureterosigmoidostomy: a review. J Urol 144: 607±611 Kajbafzadeh AM, Quinn FM, Du€y PG, Ransley PG (1995) Augmentation cystoplasty in boys with posterior urethral valves. J Urol 154: 874±877 Kennedy HA, Adams MC, Mitchell ME, Rink RC, Piser JA, McNulty A (1988) Chronic renel failure and bladder augmentation: stomach versus sigmoid colon in the canine model. J Urol 140: 1138±1143 Kinahan TJ, Khoury AE, McLorie GA, Churchill BM (1992) Omeprazole in post-gastrocystoplasty metabolic alkalosis and aciduria. J Urol 147: 435±437

250 27. Kulb TB, Rink RC, Mitchell ME (1986) Gastrocystoplasty in azotemic canines. North Central Section, American Urological Association, Palm Springs, California 28. Lapides J, Diokno AC, Silber SJ, Lowe BS (1972) Clean, intermittent self-catheterization in the treatment of urinary tract disease. J Urol 107: 458±467 29. Lau JL, Richter HMd, Fowler JE Jr, Bhatti R, Ray V (1990) Suppression of canine antral gastrin secretion by urine. J Urol 143: 402±406 30. Lauvetz RW, Monda JM, Kramer SA, Husmann DA (1995) Urinary pH and urea concentration correlate to the bacterial colonization rate in gastric, colonic, ileal and myoperitoneal bladder augmentation. J Urol 154: 899±902 31. Lazzaroni M, Sangaletti O, Bianchi Porro G (1992) Gastric acid secretion and plasma gastrin during short-term treatment with omeprazole and ranitidine in duodenal ulcer patients. Hepatogastroenterology 39: 366±370 32. Leong CH (1978) Use of the stomach for bladder replacement and urinary diversion. Ann R Coll Surg Engl 60: 283±288 33. Leong CH, Ong GB (1972) Gastrocystoplasty in dogs. Aust N Z J Surg 41: 272±279 34. Leong CH, Ong GB (1975) Proceedings: gastrocystoplasty. Br J Urol 47: 236±240 35. Lewis AG, Gardner B, Gilbert A, Jarvis PD, Wacksman J, Bove KE, Sheldon CA (1995) Relative microbial resistance of gastric, ileal and cecal bladder augmentation in the rat. J Urol 154: 1895±1899 36. Lim STK, Lam SK, Lee NW, Wong J, Ong GB (1983) E€ects of gastrocystoplasty on serum gastrin levels and gastric acid secretion. Br J Surg 70: 275±277 37. McDougal WS (1992) Metabolic complications of urinary intestinal diversion. J Urol 147: 1199±1206 38. Mitchell ME, Piser JA (1987) Intestinocystoplasty and total bladder replacement in children and young adults: follow-up in 129 cases. J Urol 138: 579±583 39. Muraishi O, Ogawa A, Tsuruta T, Kato H (1994) Bladder peptic ulcer after gastrocystoplasty for radiation cystitis. J Urol 152: 473±474

40. Ngan JH, Lau JL, Lim ST, Chan KW, Tam PC, Li LMK (1993) Long-term results of antral gastrocystoplasty. J Urol 149: 731±734 41. Nguyen DH, Bain MA, Salmonson KL, Ganesan GS, Burns MW, Mitchell ME (1993) The syndrome of dysuria and hematuria in pediatric urinary reconstruction with stomach [see comments]. J Urol 150: 707±709 42. Nguyen DH, Mitchell ME, Horowitz M, Bagli DJ, Carr MC (1996) Demucosalized augmentation gastrocystoplasty with bladder autoaugmentation in pediatric patients. J Urol 156: 206±209 43. Nurse DE, Mundy AR (1989) Metabolic complications of cystoplasty. Br J Urol 63: 165±170 44. Palmer LS, Franco I, Kogan SJ, Reda E, Gill B, Levitt SB (1993) Urolithiasis in children following augmentation cystoplasty. J Urol 150: 726±729 45. Piser JA, Mitchell ME, Kulb TB, Rink RC, Kennedy HA, McNulty A (1987) Gastrocystoplasty and colocystoplasty in canines: the metabolic consequences of acute saline and acid loading. J Urol 138: 1009±1013 46. Plawker MW, Rabinowitz SS, Etwaru DJ, Glassberg KI (1995) Hypergastrinemia, dysuria-hematuria and metabolic alkalosis: complications associated with gastrocystoplasty. J Urol 154: 546±549 47. Reinberg Y, Manivel JC, Froemming C, Gonzalez R (1992) Perforation of the gastric segment of an augmented bladder secondary to peptic ulcer disease. J Urol 148: 369±371 48. Sheldon CA, Gilbert A, Wacksman J, Lewis AG (1995) Gastrocystoplasty: technical and metabolic characteristics of the most versatile childhood bladder augmentation modality. J Pediatr Surg 30: 283±288 49. Sinaiko E (1956) Arti®cial bladder from segment of stomach and study of e€ect of urine on gastric secretion. Surg Gynecol Obstet 102: 433±437 50. Ti€any P, Vaughan JEJ, Marion D, Amberson J (1986) Hypergastrinemia following antral gastrocystoplasty. J Urol 136: 692±697 51. Wolfe MM, Soll AH (1988) The physiology of gastric acid secretion. N Engl J Med 319: 1707±1715