Valverde A, Hay J, Fingerhut A, Boudet M, Petroni R, Pouliquen X, Msika S, Flamant Y, for the French Association for Surgical Research. Senna vs Polyethylene Glycol for Mechanical Preparation the Evening Before Elective Colonic or Rectal ResectionA Multicenter Controlled Trial. Arch Surg. 1999;134(5):514–519. doi:10.1001/archsurg.134.5.514
Senna is more efficient than polyethylene glycol as mechanical preparation before elective colorectal surgery.
Prospective, randomized, single-blind study.
Multicenter study (18 centers).
Five hundred twenty-three consecutive patients with colonic or rectal carcinoma or sigmoid diverticular disease, undergoing elective colonic or rectal resection followed by immediate anastomosis.
Two hundred sixty-two patients were randomly allotted to receive senna (1 package diluted in a glass of water) and 261 to receive polyethylene glycol (2 packages diluted in 2-3 L of water), administered the evening before surgery. All patients received 5% povidone iodine antiseptic enemas (2 L) the evening and the morning before surgery. Ceftriaxone sodium and metronidazole were given intravenously at anesthetic induction.
Main Outcome Measures
Degree of colonic and rectal cleanliness.
Colonic cleanliness was better (P=.006), fecal matter in the colonic lumen was less fluid (P=.001), and the risk for moderate or large intraoperative fecal soiling was lower (P=.11) with senna. Overall, clinical tolerance did not differ significantly between groups, but 20 patients receiving polyethylene glycol (vs 16 with senna) had to interrupt their preparation, and 15 patients (vs 8 with senna) complained of abdominal distension. Senna, however, was better tolerated (P=.03) in the presence of stenosis. There was no statistically significant difference found in the number of patients with postoperative infective complications (14.7% vs 17.7%) or anastomotic leakage (5.3% vs 5.7%) with senna and polyethylene glycol, respectively.
Mechanical preparation before colonic or rectal resection with senna is better and easier than with polyethylene glycol and should be proposed in patients undergoing colonic or rectal resection, especially patients with stenosis.
SEVERAL CONTROLLED studies have shown beyond any doubt that abdominal infective complications after colonic and rectal surgery were decreased by systemic antibiotic prophylaxis.1 Antiseptic enemas such as povidone iodine also have been shown to be bacteriologically2 and clinically3,4 efficient. Mechanical preparation plays a role as well, ensuring adequate colonic and rectal cleanliness.5,6 Although it has never been shown that endoluminal cleanliness decreased the rate of postoperative complications,7- 12 most authors agree that mechanical cleansing is preferable before performing colonic resection,7 and it is used by 30% to 51% of North American colorectal surgeons.13- 15 Several mechanical cleansing preparations are available, and these can be divided into volume-induced agents, such as mannitol or polyethylene glycol, and secretory and stimulant agents (laxatives), such as sodium phosphate or senna, sometimes used in association with various types of enemas (tap water, saline solution, or povidone iodine).3,4,16 Polyethylene glycol is used by many endoscopists because the fluid contents can easily be aspirated.17 Controlled studies in elective colorectal surgery have shown that cleanliness was better with polyethylene glycol than with water enema alone17 and better with senna compared with mannitol.16 However, to our knowledge, there are no prospective, controlled studies comparing polyethylene glycol with senna, particularly the evening before surgery. We therefore undertook this prospective, randomized multicenter study to compare both types of mechanical colonic preparations as regards colonic cleanliness in elective colonic or rectal resection.
From February 1, 1992, to October 1, 1996 (56 months), 548 consecutive patients (279 men and 269 women; mean age, 64±12 years; age range, 32-93 years) were eligible for this prospective trial. Although all 18 centers (2 university hospitals, 14 teaching hospitals, and 2 private clinics) participating in this study did not start at the same time, they all finished on October 1, 1996, which explains that a median of 27 patients (range, 11-82) were entered per center and that a median of 19 (range, 11-46) patients were entered per year and per center. Twenty-five surgeons performed or supervised all operations.
All patients undergoing elective surgery for carcinoma of the colon or the proximal or middle rectum or for sigmoid diverticular disease were eligible for this study. All patients underwent resection, with palliative or curative intent in the case of carcinoma, followed by immediate anastomosis.
Ileocolonic, colocolonic, colorectal or ileorectal, or coloanal anastomoses were performed, sometimes protected by a diverting stoma, and sometimes covered by omentoplasty. The degree of stenosis was quantified by the surgeon clinically, after opening the resected intestinal specimen, as narrowing of the lumen of less than one third, one third to two thirds, and more than two thirds of normal caliber, and endoscopically as narrowing of the lumen precluding the passage of an adult-size colonoscope. The degree of stenosis was not a reason for noninclusion, any more than the presence of organ (heart, pulmonary, kidney, hepatic, or other) compromise. Patients with specific or ulcerative colitis, benign tumor, or familial polyposis without carcinoma; patients who did not undergo resection or immediate anastomosis (eg, the Hartmann procedure or abdominoperineal resection); and patients who underwent emergency resection (for obstruction or peritonitis), reversal of the Hartmann procedure, or simple closure of colostomy were not considered for the study.
Patients received senna solution (X Prep Sarget) (one120-mg package of flavored powder diluted in a glass of water or 2 packages for obese patients) or polyethylene glycol (ColoPeg Nicholas SA) (two 59-g packages of flavored powder diluted in 2-3 L of water) on the evening before surgery. All patients received 2 L of 5% povidone iodine enema on the evening before and on the morning before (at least 2 hours) surgery.4 Single-dose ceftriaxone sodium (1 g) and metronidazole (1 g), diluted in 125 mg saline solution infused for 15 minutes, were administered to all patients intravenously at anesthetic induction.4 Patients with previously recognized allergy to 1 of these or related drugs were not included.
The main outcome measure was the degree of colonic and rectal cleanliness as judged by the operating surgeon in the upstream and downstream intestinal segments and defined according to Hollender et al18 as follows: 0 indicates no fecal matter; plus sign, small amounts of fecal matter, not bothersome to the surgeon; and 2 plus signs, fecal matter bothersome to the surgeon. A 0 and/or a plus sign in the upstream and/or downstream segments were considered to be satisfactory cleanliness. The quantity and consistency of fecal matter were assessed in the proximal and distal segments through the opening made in the intestine for manual anastomosis, to introduce the mechanical stapling devices for side-to-end or end-to-end anastomoses, or after confection of the purse string in circular mechanical anastomoses. In the double-stapling technique, the endoluminal contents were assessed when the anvil was inserted through the anus. The surgeon was unaware of the type of colonic preparation administered. Secondary outcome measures included (1) consistency of fecal matter defined as solid, soft, or fluid; (2) the rate and magnitude of intraoperative fecal soiling, defined as nil, minimal, moderate, or large; (3) tolerance of the preparation as attested by the absence of abdominal pain, distension, malaise, vomiting, need to discontinue the preparation, or other complications; and (4) the rate of abdominal infective complications (including wound abscess or disruption, local deep abscess or generalized peritonitis, blood-borne infection, clinical or radiological anastomotic leakage [all patients received a routine diatrizoate sodium enema between days 8 and 10]), repeated operations, or death occurring during the hospital stay or the 30 days following hospital discharge.19 All patients who died in the hospital underwent an autopsy.
Patients were divided into 2 strata, ie, those with carcinoma and those with sigmoid diverticular disease. At the latest on the evening before surgery, patients were randomly allotted to their colonic preparation by unfolding the previously stapled upper corner of a questionnaire in preference to the envelope method,20 under which "senna" or "polyethylene glycol," determined by random number tables, was hidden. Random assignment was balanced every 6 patients by center and by stratum.
Based on the 14% difference between the previous 66% rate for colonic cleanliness after senna preparation16 and the previous mean 80% rate after polyethylene glycol preparation,21,22 the number of patients required was calculated to be 226 per group, ie, 452 patients in all, with an α and β risk equal to .05, in a 2-tailed test.23
Preoperative and postoperative patient demographics and secondary outcome measures were compared within both groups using the χ2 test for categorical values and the Student t test for continuous variables. This study was approved by the ethical committee of the coordinating center. The center effect was analyzed.
Twenty-five patients were withdrawn from the study after random allotment (12 in the senna group and 13 in the polyethylene glycol group) because of benign tumor (n=7), absence of resection or anastomosis (n=11), or a random allotment error (n=7). Five hundred twenty-three patients remained for final analysis.
Both groups of patients were comparable as regards sex, age, disease, factors affecting healing, and risk factors (Table 1).
Senna was significantly more efficient in providing intestinal cleanliness in the upstream (P=.04) and in both intestinal segments taken together (P=.006) (Table 2). Conversely, the difference in the downstream segment was not statistically significant (Table 2).
Senna significantly decreased the rate of fluid matter in the upstream and downstream (P=.001) or both intestinal segments (P =.001), whereas there was no statistically significant difference found in the rate of fecal soiling between groups (Table 3). When spillage did occur, however, soiling was more often nil or minimal with senna than moderate or large (P =.11) (Table 3).
Overall clinical tolerance was similar between groups, but fewer patients receiving senna had to discontinue their preparation or complained of abdominal distension (16 and 8 vs 20 and 15 with polyethylene glycol, respectively) (Table 4). Neither difference, however, was statistically significant.
Stenosis was found in 59.7% of our patients. Comparing the patients with or without stenosis (Table 5), stenosis was significantly associated with less cleanliness (P=.001), smaller soiling (P =.001), and poorer tolerance (P=.06) compared with patients without stenosis. The results in favor of senna were similar, irrespective of the presence or absence of stenosis (Table 2 and Table 3). Compared with polyethylene glycol, however, senna was tolerated significantly better (P=.03) in the presence of stenosis.
There was no significant difference found between senna compared with polyethylene glycol as regards the rate of each postoperative infective complication, anastomotic leakages, repeated operations, or deaths, as well as the rate of patients with at least 1 infective complication (Table 6).
No center effect was found concerning patient demographics or results.
Senna preparation was significantly associated with more effective intestinal cleanliness (Table 2), and, when fecal matter persisted, it was less fluid (Table 3), decreasing the risk for moderate or large soiling when intraoperative spillage occurred (Table 3). The tolerance and the rate of postoperative complications did not differ significantly between groups (Table 4 andTable 6). However, senna was better tolerated in case of stenosis.
The 69.5% satisfactory cleanliness rate that we found for senna (Table 2) was comparable with results (66%) of a previous study.16
The cleanliness provided using polyethylene glycol has been reported to be subjectively good, very good, or excellent in 72%,22 90%,21 and 100%24 of patients, rates that were higher than those found in our study. Our 57.8% upstream-downstream cleanliness rate with polyethylene glycol, however, was close to the 61% (33/54) cleanliness rate found by Wolters et al.12 Possible explanations for thesediscrepancies include use of the semiquantitative grading system of Hollender et al18 in our study (Table 3), less subjective than that used by others,12,22,24 and the high rate of stenosis (59.5%)(Table 4) observed in our series, which significantly (P=.001) limited intestinal cleanliness (55.1% [172/312] in patients with stenosis vs 75.8% in patients without stenosis) (Table 5). Only 1 study25 on polyethylene glycol reported a cleanliness rate of 58% with stenosis vs 68% without stenosis, especially in the left compared with the right colon, where feces are more fluid. In other reports,21,22,24 the rate and degree of stenosis were not evaluated. Moreover, stenosis is probably seen more often in the surgical population than in the overall colonoscopic population.21,22 The quantity of fluid retained in the upstream segment after preparation was directly proportional to the amount of cleansing fluid administered,12 and, notably in the case of stenosis, explained why upstream fluid endoluminal contents after 1 glass of senna were significantly (P=.001) (Table 2) less than after 2 to 3 L of polyethylene glycol. Conversely, cleanliness in the downstream segment, similar in both groups, was undoubtedly due to the amounts (2 L) and efficacy of associated enemas. The 100% cleanliness rate observed by Beck et al24 was obtained with polyethylene glycol administered until clean effluent was evacuated, followed by a bisacodyl capsule in a small group of 30 patients without significant stenosis. In the same study,24 senna was administered 48 hours before surgery, whereas it has been shown that the maximal action of senna takes place 6 to 12 hours after administration.26 Sodium phosphate also is more efficient when administered twice, with the latter dose given the morning before surgery.27
A further advantage of senna is that when the residual fecal contents of the colon are less fluid (Table 3 andTable 5), the risk for spillage into the operative field is lower,7 potentially reducing the threat of anastomotic leakage.11 This is why, in agreement with several authors,5,7,28 we believe that endoluminal fluid contents should be avoided in surgery. This contrasts with the expectations of colonoscopists who, conversely, easily can aspirate fluid contents as opposed to solid or soft contents.17
As in other types of mechanical preparations, clinical tolerance depends on the agent itself, as well as on the quantity of fluid administered and the use of associated preparations (enemas) or medications (metoclopramide hydrochloride).24 In our series, tolerance of polyethylene glycol was comparable with that reported in the literature for traditional 3-day preparations, including enemas,17 mannitol preparations,18,29,30 and polyethylene glycol when used alone.21,22,24 Our overall rate of patient discomfort (21.0%) (Table 4) was very close to those found in the literature, ie, 20.7%,14 24%,22 and 25%.24 Our rate of vomiting was the same as that found by Solla and Rothenberger14 (2.7%), but lower than those found by others, ie, 6%,21 10%,24 and 11%.22 In contrast, Wolters et al12 reported no vomiting. The nausea rate in their series, however, was high, 33% of 54 patients, but the quantity of fluid administered was often abundant (1.3-7 L) until clean fluid was evacuated. In conclusion, satisfactory tolerance of polyethylene glycol seems to depend on the relatively moderate quantity of the fluid ingested in our study (2-3 L).
In the literature, senna tolerance has been judged to be excellent because there were no major complaints recorded.3,4 Our study shows that clinical tolerance of senna and polyethylene glycol was similar (Table 4). However, in other randomized studies,31,32 sodium phosphate has been shown to be significantly better tolerated than polyethylene glycol for colonoscopy preparation. In our study, it is difficult to ascertain and to evaluate the tolerance of senna or polyethylene glycol alone, because the few, minor discomforts observed may in fact be observed or enhanced due to associated povidone iodine enemas.3,4 Enemas were used not only for their mechanical action,13- 15 but also for their local antiseptic properties.2- 4 In our series, senna was tolerated better than polyethylene glycol, especially in the presence of stenosis (Table 5), which confirms that the quantity of fluid ingested plays a role in tolerance. Last, when studying tolerance, one must take into account the number of patients for whom it is necessary to discontinue the preparation. This occurred in 6.9% of patients in our series (Table 4), with no significant difference between senna and polyethylene glycol, comparing favorably with the 0% to 25% of cases in the survey conducted by Beck and Fazio,13 but the exact number of patients receiving laxatives or polyethylene glycol (43% of the overall population) was not known. On the other hand, 27% of patients receiving polyethylene glycol for colonoscopy33 and 63% of patients undergoing resection34 had to discontinue their colonic preparation, but this was with 4 L of fluids.
We did not study the cost-effectiveness of both preparations. Compared with oral sodium phosphate,32 senna is easy to administer and relatively less expensive and may be used as an outpatient bowel preparation,21,35 thus reducing costs and preoperative hospital stay.
In our study, there was no significant difference found between groups of patients as concerns the rate of postoperative infective complications (Table 6). To the best of our knowledge, no clinical study to date has shown that mechanical preparation alone, excepting povidone enemas,3,4 can reduce postoperative abdominal infective complications. Oral mechanical preparation in itself did not decrease the bacterial load or their type.10,11
Cleanliness of the intestines and the consistency of residual fecal matter are important factors to consider, especially in laparoscopic colonic resection, because of the difficulty in ensuring adequate clamping of the colonic segments in this technique.
As with 90 mL of sodium phosphate,34 senna can be considered the standard for elective colonic or rectal resection. Further progress can be expected when senna vs sodium phosphate, various combinations of oral and systemic antibiotics, and other antiseptic enemas (povidone vs hypochlorite)36 are compared.
Reprints: Jean-Marie Hay, MD, Associations de Recherche en Chirurgie, 8 Avenue des Peupliers 92270 Bois-Colombes, France.
Eaubonne: Patrice Baillet, MD. Aulnay-sous-Bois: André Elhadad, MD, Didier Brassier, MD, Elias Habib, MD. Thonon les Bains: Christian Dilin, MD. Poissy: Abe Fingerhut, MD, FRCS. Dinan: Jacques Francin, MD. Romorantin: Henri Hennet, MD. Créteil: Michel Julien, MD, Pierre-Louis Fagniez, MD, Nelly Rotman, MD, Marie-Jeanne Boudet, MD. Chatellerault: Marc Kalfon, MD. Corbeil: Gérard Kohlmann, MD. Pau: Yves Laborde, MD. Paris: Hughes Levard, MD. Meulan: Simon Msika, MD. Besançon: Edouard Pelissier, MD. Senlis: Patrick Peyramaure, MD. Vernon: Claude Rey, MD. Nice: Jean-Louis Sicard, MD. Montmorency: Yves Soulier, MD. Colombes: Jean-Marie Hay, MD, Yves Flamant, MD, Guy Zeitoun, MD.