A Randomized Controlled Trial of 0.5% Ferric Hyaluronate Gel (Intergel) in the Prevention of Adhesions Following Abdominal Surgery

Abstract

Introduction:

Intestinal adhesion following abdominal surgery is a significant sequela to abdominal surgery. Intergel is a hyaluronate-based gel that reduces the incidence of postoperative adhesions when added to the peritoneal cavity before closure in gynecologic surgery. This is a randomized controlled trial evaluating the efficacy and safety of Intergel in colorectal resections. Although the study aimed to recruit 200 patients based on power analysis, recruitment was suspended because of the high morbidity in the treatment group.

Methods:

A total of 32 patients were randomized to either Intergel treatment (treatment group) or no treatment (control group) following open abdominal surgery. Primary endpoints included the incidence of adhesive obstruction, the need for subsequent adhesiolysis, and the incidence of wound and anastomotic complications. A secondary endpoint involved quality-of-life assessment.

Results:

Seventeen patients were randomized to the treatment group and 15 to the control group. All patients, except 1 in the treatment group, underwent resection and anastomosis of the colon or rectum for benign or malignant disease. A significant difference was observed in the number of patients with postoperative morbidities between the 2 groups (65% treatment group versus 27% control group, P = 0.031). There was a high rate of anastomotic dehiscence (5 treatment group versus 1 control group, P = 0.178) and prolonged postoperative ileus (10 treatment group versus 2 control group, P = 0.011) observed in treatment group. One case of peritonitis occurred in the treatment group in the presence of an intact anastomosis. Wound complications were more common in treatment group but failed to reach statistical significance.

Conclusion:

The use of Intergel in abdominal surgery where the gastrointestinal tract is opened leads to unacceptably high rates of postoperative complications.

The use of ferric hyaluronate gel in the prevention of adhesions following abdominal surgery involving intestinal resection and anastomosis leads to a high incidence of postoperative complications.

Peritoneal adhesion is an inevitable consequence of abdominal surgery and a cause of significant postoperative morbidity. They predispose to recurrent episodes of abdominal pain and intestinal obstruction that frequently require hospitalization.^1,2 This adds to the financial burden placed on the healthcare system and adversely affects quality of life.^3,4

Much interest has been focused on reducing the incidence of postoperative adhesions. Simple approaches include the use of unpowdered gloves, generous irrigation of the peritoneum with distilled water or normal saline after surgery, and the addition of anti-inflammatory agents into the wash. Such measures have not been very effective.⁵ A number of therapeutic agents have shown potential promise, including antiadhesive membranes (INTERCEED Absorbable Adhesion Barrier, Ethicon, Somerville, NJ; Preclude Surgical Membrane, W.L. Gore Co.; Seprafilm Surgical Membrane, Genzyme, Cambridge, MA) that may be laid on localized areas of the peritoneum; or viscous gels (Hyskon, Pharmacia, Piscataway, NJ; Sepracoat, Genzyme) instilled into the peritoneal cavity. Although gel formulations have the theoretical advantage of reducing adhesion formation throughout the peritoneal cavity, the use of some preparations has been associated with undesirable side effects because of their oncotic nature resulting in ascites.⁶ In addition, the gravitational pooling of gels may limit their effectiveness following certain types of abdominal surgery.⁵

A new antiadhesive gel preparation, Intergel (Lifecore Biomedical Inc., Chaska, MN), has been introduced that uses a viscous formulation of the naturally occurring polysaccharide, hyaluronic acid (HA). Cross-linkage of the carboxylate groups on the HA molecules by chelation with ferric (Fe³⁺) ions results in a viscous lubricating gel preparation. Instillation into the peritoneal cavity following abdominal surgery minimizes tissue apposition during the critical period of fibrin formation and mesothelial regeneration. Lymphatic drainage appears to be the major route of elimination from the peritoneal cavity, with an estimated half-life of 51 hours.⁷ HA-based gels have been shown to reduce adhesion formation in numerous animal studies.^8–13 In humans, a randomized controlled trial involving the Intergel formulation has shown promising results following pelvic gynecologic surgery, achieving a 59% reduction in adhesion formation.¹⁴

A randomized controlled trial was started at our institution to evaluate the role of Intergel as an antiadhesive agent following open abdominal surgery. The trial was suspended after recruitment of 32 patients because of the occurrence of a significant number of adverse events in the treatment arm.

PATIENTS AND METHODS

Protocol

Recruitment was started on September 3, 2001 and suspended on October 22, 2001. This trial was run in conjunction with the Clinical Trials and Epidemiology Research Unit, Ministry of Health, Singapore (CTERU). It aimed to recruit 700 patients to either treatment with Intergel (treatment group) or no treatment (control group) over a period of 2 years (Fig. 1). There were 2 parts to this trial. The first was a randomized controlled trial evaluating the use of Intergel in reducing the rate of intestinal obstruction due to adhesions following abdominal colorectal surgery. The second part was the addition of laparoscopy to evaluate abdominal adhesions to the abdominal wall surgical site at the time of ileostomy reversal in patients who had had ileostomy at initial surgery.

FIGURE 1. Schematic representation of randomized controlled trial of Integel as an antiadhesive agent following abdominal surgery.

All patients over 21 years of age, capable of giving written informed consent, and undergoing laparotomy were recruited. Exclusion criteria included pregnancy, the presence of frank peritoneal infection, evidence of hepatic or renal impairment, congestive cardiac failure, a history of severe drug allergy or autoimmune disease, and prior intra-abdominal adhesive complications. Patients with hemachromatosis were excluded due to the Fe³⁺-load associated with the treatment solution. Placement of an intraperitoneal drain at the end of surgery precluded randomization into the trial. Patients who had a defunctioning ileostomy are to undergo laparoscopic assessment for intra-abdominal adhesions at the time of ileostomy closure. Written informed consent was obtained from all patients prior to entry into the trial.

Interventions

At laparotomy, the extent and severity of preexisting adhesions were recorded using a standardized scoring system: 0, no adhesion; 1, filmy adhesions; 2, intermediate adhesions; 3, dense adhesions. This was assessed according to 12 sites viz. peritoneum in the midline, right upper quadrant, right lower quadrant, left upper quadrant, left lower quadrant, omentum, among small intestines, colon right of midline, colon left of midline, rectosigmoid portion of colon, pelvic area, and anastomotic site, giving a maximum score of 36. Operative details were documented with particular attention to the site and type of anastomosis, the incidence of technical complications, excessive blood loss, and possible peritoneal contamination by enteric contents. Randomization was performed on completion of surgery and prior to closure of the abdominal wound by telephone calls to CTERU on completion of laparotomy before closure. This was performed independent of the clinical participants by a computerized randomization system at CTERU. Patients received instillation of either 300 mL of Intergel solution, prewarmed to 37°C, into the peritoneal cavity (treatment group) or prewarmed distilled water without Intergel instillation (control group). Care was taken to avoid spillage of the gel onto the abdominal wound in the treatment group. Abdominal wounds were closed by a conventional mass closure technique.

Outcome Measures

The primary endpoints included the incidence of postoperative adhesive obstruction, the need for subsequent adhesiolysis, and the rates of wound and anastomotic complications. A secondary endpoint involved quality-of-life assessment (Fig. 2).

FIGURE 2. Quality-of-life assessments.

Preoperative investigations included a full history and physical examination, measurement of routine blood profiles (full blood count, urea and electrolytes, and liver function tests), electrocardiography, and a chest radiograph. Patients were supplied with a standardized patient information sheet. Baseline quality-of-life assessment was recorded using the Gastrointestinal Quality of Life Index.¹⁵ Postoperatively, patients were closely monitored, with particular attention given to the length of postoperative ileus (assessed by the paucity of bowel sounds in the absence of colicky pain), time to commencement of normal diet, time to first bowel motion, the incidence of wound and anastomotic complications, and other postoperative complications and adverse events. Patients were followed up on 30 days postoperation and at 3-month intervals for 3 years to detect symptoms related to adhesive obstruction. Quality-of-life assessment was scheduled for 6 and 12 months postoperatively. Data on any patients requiring readmission for intestinal obstruction were to be recorded prospectively, including an assessment of degree and severity of adhesions in those requiring laparotomy. The clinical assessors were not present at the time of the surgery and are blinded to the intervention groups (NYY). Postoperative intestinal obstruction was diagnosed and scored when there is at least 3 of the following parameters (one point for each parameter): colicky pain, vomiting, abdominal distension as reported by the patient on interview (measured as the increase in greater than 20% in abdominal girth at the level of umbilicus compared with preoperatively), lack of flatus/feces for more than 24 hours, and/or dilated small bowel air-fluid levels on abdominal x-ray. Clinical monitoring and data collection were performed by Clinical Trials and Epidemiology Unit, Ministry of Health, Singapore. The protocol was approved by the local hospital ethics committee.

Sample Size Estimation, Assignment, and Blinding

Power calculation was based on a 14%, 3-year incidence of adhesive obstruction following abdominal surgery.¹⁶ With a relative reduction of 50% in the incidence of adhesive obstruction in the treatment group as compared with the control group,¹⁴ approximately 300 patients would be required per treatment arm for a two-sided test size of 5% and a power of 80%. A patient attrition rate of 10% was allowed for, giving a preliminary recruitment target of 700 patients. Patients were entered into the trial using randomized blocks of 4 and 6 on a 1:1 treatment allocation. The treatment assignment was generated using a computer program. All analysis was run by SPSS v10 (Chicago, IL) and is to be done on an intention-to-treat basis.

RESULTS

Thirty-two patients undergoing elective surgery were entered into the trial, with 17 randomized to treatment group and 15 to control group. Two ineligible patients were recruited, 1 in each of the 2 groups. One had placement of intra-abdominal drain and 1 had a history of congestive cardiac failure. Their data have been included in the respective groups and analyzed on an intention-to-treat basis. Patient demographics, preoperative diagnoses, rates of comorbidity, and previous abdominal operations are detailed in Table 1. There was no statistical difference between the 2 groups in terms of sex distribution, age, diagnoses, or previous abdominal operations. Patients in the treatment group had a higher rate of preexisting comorbidities than those in the control group, but this failed to reach statistical significance (9 in treatment group versus 4 in control group, P = 0.131). There is no difference in the median total score of severity and extent of preoperative adhesions in the 2 groups (P = 0.065, Mann-Whitney U test).

TABLE 1. Patient Demographics, Diagnoses, Comorbidities, and Previous Abdominal Operations in Patients Randomized to Intergel Treatment (Treatment Group) or No Treatment (Control Group)

All patients underwent laparotomy, with 31 of 32 having resection of either the colon or rectum. One male patient in treatment group had a mesenteric gastrointestinal tumor that was completely excised without recourse to bowel resection. This was the only patient in which an enteric anastomosis was not performed. All patients received perioperative antibiotics and antiembolic prophylaxis as per unit policy. The operations performed are shown in Table 2. Reconstruction after low anterior resection was by either colopouch-anal or straight coloanal anastomosis, depending on the surgeon's discretion. Low anterior resection was combined with defunctioning ileostomy in two thirds and one half of the treatment group and the control group, respectively. In no case was there any excessive blood loss or significant contamination of the peritoneal cavity during bowel resection or anastomosis. One patient in treatment group sustained a small capsular tear to the spleen that was treated with diathermy and the application of hemostatic gauze. All anastomoses were constructed by conventional stapling methods, apart from one coloanal anastomosis in control group that was hand-sewn. There were no patients who had undergone laparoscopic assessment before the trial was suspended.

TABLE 2. Operative Procedures Performed in Patients Randomized to Intergel Treatment (Treatment Group) or No Treatment (Control Group)

Postoperatively, a significant difference in overall morbidity was observed in the number of patients with at least one mortality between the 2 groups. This has resulted in suspension of the trial (65% treatment group versus 27% control group, P = 0.031). Details of the postoperative complications are given in Table 3. Of particular concern was the high rate of anastomotic dehiscence in the treatment group as compared with the control group although this failed to reach statistical significance because of the small numbers involved (5 treatment group versus 1 control group, P = 0.178). Anastomotic dehiscence in treatment group involved 2 colopouch-anal, 1 straight coloanal, 1 colorectal, and 1 ileorectal anastomoses. The 2 colopouch-anal anastomoses had been defunctioned by a covering ileostomy. Three of the 5 dehiscences in the treatment group required further laparotomy for peritoneal lavage with formation of a defunctioning stoma if not constructed at initial operation. One patient with dehiscence of the colopouch-anal anastomosis and who had been defunctioned required examination under anesthesia with drainage of a pelvic abscess. The remaining clinical anastomotic dehiscence (not verified at relaparotomy) occurred in an elderly female patient following a high anterior resection. No further intervention was undertaken at the relatives' request because of her age and physical condition, and she subsequently died. In total, 3 patients had relaparotomy in the treatment group for anastomotic problems. The single anastomotic dehiscence in control group involved a hand-sewn coloanal anastomosis with defunctioning ileostomy.

TABLE 3. Thirty-Day Postoperative Morbidity and Mortality

One patient in the treatment group developed peritonitis on postoperative day 2 after an extended right hemicolectomy and was found to have purulent infected material throughout the peritoneal cavity at relaparotomy. The anastomosis was intact, and there had been no spillage of enteric contents at initial operation.

Microbiologic analysis of infected peritoneal contents was available on 5 of the 6 anastomotic dehiscences. They showed growth of Escherichia coli and Enterbacter spp in 1 case each, and mixed enteric flora in the other three.

A significant difference was observed in the number of patients with prolonged postoperative ileus (>4 days) between the 2 groups (10 of 17 in the treatment group versus 2 of 15 in the control group, P = 0.011). Five of the 10 in treatment group were associated with the above-mentioned anastomotic dehiscences or purulent peritonitis. The remaining 5 patients showed unexplained abdominal distension, nausea, and vomiting that eventually resolved with nonoperative measures. In 1 case, ileus was associated with fluctuating pyrexia that responded to intravenous antibiotics.

An increase in the number of wound infections was noted in treatment group, but this was not significant statistically (6 in the treatment group versus 3 in the control group, P = 0.444). Three partial wound dehiscences occurred in the treatment group and were all associated with coexisting anastomotic failure or purulent peritonitis. One case of superficial wound dehiscence requiring secondary suture occurred in the control group. Other postoperative complications included 1 case each of pulmonary embolism, enteric fistulation, cardiac failure, and bronchopneumonia in treatment group and 1 case of myocardial infarction in the control group.

There were 2 fatalities: 1 in the treatment group (5.9%) involving the elderly patient with anastomotic dehiscence described above who developed pulmonary embolism after relaparotomy and 1 in the control group (6.7%) following a massive myocardial infarct secondary to a postoperative bleed on postoperative day 2. The rest of the patients made full recovery and were discharged home.

Follow-up

Four patients including 1 ineligible patient were not followed up, 1 was lost to follow-up, and 2 patients died before any follow-up was done. Twenty-eight patients had a 3-month follow-up, 15 patients had a 6-month follow-up, 20 patients had a 9-month follow-up, and only 10 patients had 1-year follow-up. The incidence of adhesive complications is shown in Table 4. There was significant difference in the incidence of abdominal distension in the treatment group during the first 3 months of follow-up (P = 0.018, Fisher exact test). One patient from treatment group developed intestinal obstruction due to adhesion during the follow-up period. This was treated successfully with intravenous fluids and nasogastric decompression without the need for relaparotomy.

TABLE 4. Adhesive Complications During Scheduled Follow-up

Quality-of-Life Assessments

Baseline quality-of-life assessments were available for 29 patients, 6-month quality-of-life assessments were available for 16 patients, and 1-year quality-of-life assessments were available for 11 patients. There was 1 patient who declined quality-of-life assessments. Only 3 patients had 3 assessments (baseline, 6-month, and 1-year) completed. Repeat measurement technique was used to evaluate the effects of treatment and time, while the variance and covariance matrix was modeled as ante-dependence (ANTE). It showed that neither the effect of treatment nor the effect of time was significant. (P = 0.36 and P = 0.058, respectively).

DISCUSSION

Intra-abdominal adhesions remain a significant problem following abdominal surgery. Some 14% of patients will develop adhesive intestinal obstruction within 2 years of surgery, with 2.6% requiring adhesiolysis.¹⁶ After a first episode of adhesive obstruction, 53% of patients will go on to develop a second episode, and 83% of these will have chronic symptoms.¹⁷ An urgent need therefore exists to develop effective strategies to reduce adhesion formation.

HA-based gels have shown promise as antiadhesive agents in numerous experimental studies. One large multicenter randomized trial has shown Intergel to be effective in adhesion reduction following pelvic gynecologic surgery.¹⁴ No serious or potentially serious adverse events were reported in this trial, and no difference was noted in the incidence of complications between the treatment and control groups.

Gynecologic procedures are generally of a “clean” nature and do not usually involve opening of the gastrointestinal or urinary tracts. Gastrointestinal surgery, on the other hand, invariably involves opening of the bowel and the construction of an anastomosis. Such operations will release bacteria-laden enteric contents into the peritoneal cavity and initiate intra-abdominal sepsis. Clinically, this is rarely a problem in most instances. However, the effect of the gel instillates on the incidence of intra-abdominal sepsis following bowel resection is unknown in the human situation. Infection potentiation studies performed in rat models of peritonitis have shown conflicting results. Johns and diZerega used a rat model of bacterial peritonitis to test various solutions of Ringer's lactate, dextran, and Intergel for infection potentiation.⁷ No significant increase in mortality was observed between Intergel-treated and control animals. In contrast, high-dose, but not low-dose, dextran solutions significantly increased mortality. Interestingly, both Intergel and dextran instillation significantly reduced the incidence of abscess formation, although the authors give no explanation for this. In direct contrast to these findings, Tzianabos et al found a significant increase in the mortality rate following bacterial challenge in rats receiving gels composed of hyaluronate/carboxymethylcellulose or ferric ion cross-linked HA.¹⁸ The authors point out that iron-containing solutions can markedly increase the virulence of many different bacterial species and that certain virulent strains of human pathogens have evolved mechanisms to scavenge free iron from their environment.^19–21 Biodegradable antiadhesion membranes were not associated with an increased mortality in this study but have been implicated in potentiating peritonitis in other animal²² and human situations.²³

The effect of antiadhesive gels on anastomotic healing has not been previously tested in humans. In rats, HA-based membranes reduce postoperative adhesions without compromising the strength of ventral hernia repair.²⁴ In rabbits, repair of intentionally formed enterotomies has been used to test the effect of HA membranes on bowel healing.²⁵ Although the membranes failed to reduce the incidence of adhesions related to the enterotomies, they were not associated with an increased enterotomy leak rate.

Because of the lack of information regarding the potential for adverse effects with Intergel treatment in humans, it was with caution that we embarked on the current clinical trial. It became evident after the accrual of only 32 patients that there was an unacceptably high morbidity rate in the Intergel-treated group. The occurrence of 5 of 17 anastomotic dehiscences in treatment group was a cause for considerable concern. In these patients, a characteristic pattern of events was observed. Typically, the patient would develop signs and symptoms of a prolonged ileus, with abdominal distension, nausea, and vomiting, which would persist for several days until overt peritonitis became apparent late in the postoperative course. It is not known whether Intergel interfered with anastomotic healing, predisposing to dehiscence and secondary peritonitis, or if Intergel associated infection was the initiating event that subsequently led to disruption of the anastomoses. In the case where infected peritonitis was found in the presence of an intact anastomosis, the latter etiology must be suspected. The single dehiscence in the control group was apparent on the classic fifth postoperative day, whereas 4 dehiscences in the treatment group were detected on days 6, 7, 12, and 14, and one on water soluble contrast enema at day 21 prior to reversal of defunctioning ileostomy. Nonetheless, it may be possible that the gel interfered with anastomotic healing and hence predisposed to anastomotic dehiscence as it is probably by the same effect that adhesions are lysed or prevented from forming.

The high rate of prolonged ileus in the treatment group is a further cause for concern. In the light of the above discussion, it is tempting to postulate that in some of these cases a subclinical peritonitis had been initiated, which either resolved spontaneously or responded to antibiotic therapy before anastomotic integrity was compromised.

The results of this prematurely suspended trial indicate that Intergel treatment, in the context of bowel resection, is associated with an unacceptably high morbidity rate. Its use predisposed to the development of late postoperative peritonitis and anastomotic dehiscence. Based on these findings, the use of Intergel, in situations where the gastrointestinal tract has been opened, is not endorsed. Further animal studies are needed to investigate the potential mechanisms behind these adverse events.

ACKNOWLEDGMENTS

The authors thank Johnson & Johnson Pharmaceuticals Plc, which donated the Intergel preparation used in the study.