Covering stoma for elective anterior resection of the rectum: An outmoded operation?

Covering Stoma for Elective Anterior Resection of the

Rectum: An Outmoded Operation?

L. Peter Ffeiding, MB, FRCS, Waterbury, Conriecticut

Sarah Stewart-Brown, MRCP, London, England

Rosemary Hlttinger, London, England

Linda Blesovsky, BA, London, England

The prevention of septic complications after large bowel surgery is a major preoccupation of the dedi- cated colorectal surgeon. Anastomotic dehiscence is a potent cause of postoperative intraabdominal and pelvic sepsis, and a number of technical axioms are usually discussed in this context: satisfactory bowel preparation, good blood supply to the bowel ends, and total avoidance of anastomotic tension. Fur- thermore, it is often stated that if a surgeon finds technical difficulty when attempting to fashion the anastomosis, a covering-protective colostomy should be placed proximal to the anastomosis because the stoma may not prevent an anastomotic disruption but is likely to mitigate the disastrous consequences of progressive pelvic sepsis if an anastomotic leak occurs in the presence of fecal stream continuity. It is generally accepted that anastomotic complications occur more frequently after rectal excision with an anastomoses below the peritoneal reflection than for intraperitoneal anastomoses [I]; however, in the literature there has been little definition concerning the precise criteria for the use of such a proximal stoma after elective colorectal resection. Further- more, common experience has shown that some surgeons use a proximal vent frequently whereas others rarely do so.

The purpose of this analysis was to study this subject using the data base of the Large Bowel Can- cer Project in Britain which is a prospective investigation, started in 1976, to record the clinical events and outcome of colorectal cancer management.

From the Department of Surgery, St. Mary’s Hospital, London, England. Supported In part by Downs Surgical Ltd., the Jessie Wlllment Bequest, North-West Themes Regional Health Authority, and the Cancer Research Camoakn. London. Endand.

f&u&s for rep~int;should be eddressed to L. Peter Fielding. St. Mary’s Hospital. 56 Franklin Street, Waterbury. Connecticut 06702

Presented at the 64th Annual Meeting of the New England Surgical So- cfety. Bretton Woods, New Hampshire, September 29-October 2, 1983.

524

Patients and Methods

The Large Bowel Cancer Project started in May 1976, and currently 94 surgeons in 23 hospitals are taking part. The records of all patients with a diagnosis of colorectal cancer were collated by specially trained researchers who traveled to the hospitals concerned to review the patients clinical chart’s, operating room records, and histopathology documents. The information was stored in a computer in Imperial College, London in a structured data base format (Info1 System) which was then transferred to the main frame University of London computer for data analysis carried out using the program suite called “Statistical Package for the Social Sciences” [2].

The records of 4,500 patients with adenocarcinoma of the large bowel have been studied of which 2,057 patients were found to have had an elective anastomosis excluding those who presented with obstruction, bowel perforation, or those requiring elective reconstruction after a staged operation. The results described are those relating to the treatment of the presenting condition and include all operations required. Postoperative mortality is defined as death occurring while the patient was undergoing treatment in the hospital. Clinically significant anastomotic breakdown was accepted when a frank fecal fistula was present, an anastomotic breakdown was seen at laparotomy or postmortem in association with peritonitis, when a patient exhibited clinical features of anastomotic leak that was then confirmed by sigmoidoscopy, rectal examination, or radiologic investigation, and when certain types of intraabdominal abscesses were seen after a very careful clinical review. These criteria were adhered to strictly, and thus the incidence of anastomotic leakage may be somewhat underestimated.

After review of the clinical findings and hi&pathology reports, the majority of patients were classifiable as having curative or palliative procedures, depending on excision of all macroscopic tumor. In a small number of patients there was some doubt about the degree to which all tumor had been excised; this small subgroup was termed the not known curative group. Statistical analysis was carried out using the chi-square test unless otherwise stated [3].

The Amerkan Journal of Surgery

Stomas in Elective Anterior Resection

TABLE I Overall Reeutts Anaetomtk Leek Rate and Mortality According to Presence or Abeence al Stoma

Total Leak Patients n %

Total Died Patients

n % Died ( % )

All anastomoses (n = 2,057)

No stoma Stoma

lntraperitoneal anastomoses (n = 1,407)

No stoma stoma

Extraperitoneal anastomoses (n = 623)

No stoma Stoma

1,731 126 71 33 26.2 19 326 56 17.8 5 8.6 1.5

1,299 67 5.2’ 23 34.3 1.8 108 11 10.2 1 91 1

407 65 15.9+ 10 15.4 2.4 214 50 23.4 4 8 1.9

l p <0.05.

+ p <0.02.

Results Overall figures: Of the total 2,057 patients

studied, the majority (1,731 patients, 84.2 percent) did not have a covering stoma but the remainder (326 patients, 15.8 percent) had some type of bowel decompression by transverse or sigmoid colostomy (239 patients) or cecostomy or loop ileostomy (87 patients). The overa mortality in these groups was similar (with stoma 7 percent, without stoma 6.1 percent). The mean length of hospital stay in patients who left the hospital alive was different (with stoma 26 days, without stoma 15 days; p <0.05, Student’s t test for group data). The anastomotic leak rate when a stoma was in place was higher than when it wasn’t (17.8 percent versus 7.1 percent; p <0.05), and the mortality in each leak group expressed as a percentage of those who had an anastomotic dehiscence was also different (with stoma 8.6 percent versus without stoma 26.2 percent, p < 0.05). However, when this mortality was expressed as a percentage of the total number of patients in each group no differences were found (with stoma 1.5 percent versus without stoma 1.9 percent). A similar analysis was carried out which further subdivided the patients into those with an intraperitoneal anastomosis and those with an extraperitoneal anastomosis (anterior resection of the rectum) (Table I). Some of the statistical differences were lost because of the reduced number of patients in each subgroup. There was no precise information concerning the specific reasons for establishing a stoma in some of these patients. There was occasional commentary on the difficulty of constructing an anastomosis, and in 50 patients, there was a specific statement concerning poor bowel preparation with residual fecal loading. Of these 50 patients with stated fecal loading, 19 (1 percent) were in the no stoma group, whereas the remaining 31 patients (10 percent) were in the stoma group. This

VolullW 147, April 1984

difference was statistically significant at the 5 percent level. Furthermore, for those patients in whom fecal loading was recorded, the anastomotic leak rate was 18 percent compared with only 9 percent in the remaining patients (p CO.05). Thus, fecal loading seemed to have provided one incentive to construct a covering colostomy, and this feature was associated with a higher anastomotic leak rate. There were few comments concerning other technical aspects of anastomotic construction, namely blood supply to the bowel ends, anastomotic tension, and mobiliza- tion of the splenic flexure.

Results according to surgeon policy: The percentage of patients who received a covering stoma varied from surgeon to surgeon (Figure 1). To at- tempt an analysis of surgical policy, the results for those surgeons who undertook stomas infrequently (low stoma group, less than 5 percent of patients received stomas) were compared with those of surgeons who performed the procedure most frequently (high stoma group, greater than 20 percent of patients received stomas). A detailed comparison of patients who presented to these two groups of surgeons and those patients in the remainder of the data base are shown in Table II according to age, site of tumor, tumor mobility, curative tumor resection, and clinically evident anastomotic leakage rate. The data showed no significant differences between the two groups of surgeons and demonstrated that the two subpopulations were clinically very well matched (we again emphasize that only elective management was included in this study). However, the length of hospital stay for patients who left the hospital alive was different. Low stoma group versus high stoma group, median length of stay was 16 days and 20 days, (25 percent difference) for the low stoma group and high stoma group, respectively, and the mean length of stay was 18 days and 23 days for the respective

525

Fielding et al

No. of Surgeons (n-41)

% covered by stoma

Figure 1. Frequency distribution of number of surgeons against their covering stoma rate. Oniy surgeons with 20 or more anas- ~h~studyhawebeenlnckrdsdlnordertod~o~the percentags fi@a@s with smaii numbers of patients per surgeon.

groups (28 percent difference). These substantial differences are attributal to the hospitalization time required to close the stoma and manage the higher rate of anastomotic leakage in the high stoma group.

It is generally agreed that the anastomoses that are at greatest risk for disruption are those below the peritoneal reflection, and the percentage of such anastomoses for the two groups of surgeons was simii (30.5 percent and 32 percent for the low stoma group and high stoma group, respectively). The overall results according to surgeon policy group (Table III) clearly demonstrated that those surgeons using stomas very frequently have significantly worse results than those surgeons using stomas infrequently: overall anastomotic leakage rate 20 percent and 7.8 percent, respectively, p <0.05; overall mortality 8.4 percent and 3.6 percent, respectively, 0.1 p <0.05.

The difference in the numbers of patients in whom a stoma was fashioned as a protective cover to these extraperitoneal anastomoses was surprisingly large (low stoma group 4.3 percent, high stoma group 63.2 percent) (Figure 2). In those patients who had a stoma, the overall anastomotic leak rate was similar in both groups (low stoma group 16.7 percent, high stoma group 22.4 percent), but the mortality was higher in the latter group (low stoma group 0, high stoma group 7.5 percent). Furthermore, it is important to note that in those patients who did not have a stoma, both the leak rate and mortality were greater for those patients being treated by surgeons in the high stoma group than those treated by surgeons in the low stoma group. The leak rates were 15.8 percent and 7.4 percent for the low and high stoma groups, respectively, and the respective mortalities 10.5 percent and 3.7 percent. The type of stomas used in association with these extraperitoneal anastomoses in the low stoma group was transverse or sigmoid

TABLE II Patient Groups According to Surgeon’s Policy (values expressed as percentages)

Low High Stoma Stoma Remainder Group Group Group

Data (n = 462) (n = 446) (n = 1,111)

Age (yr) 7-40 1.1 1.7 2.3 40-49 5 6 7.4 50-59 16.9 15.9 14.4 60-69 26.1 32.2 31.6 70-79 36.7 36.2 32 60+ 10.2 6.1 12.2

Site of tumor l Right colon 36.2 32.7 32.7 Splenic flexure 3 2.3 3.3 Left colon 49.9 46.9 49.2 Rectum 9.7 16.1 13.4 Others 1.1 2.7 1.4

Tumor mobility Mobile 67.5 69 66.9 Not mobile 32.5 31 33.1

Cure rate+ Curative 67.3 71.9 74.5 NKC 11.4 9.4 6.2 Palliative 21.3 16.7 19.4

Length of stay (d) Median 16 20 19 Mean 16 23 22 Range 5-79 4-135 4-213

l Sites of tumor not available for seven patients in the remainder group.

+ Cure rate not available in 15 patients in the low stoma group.

NKC = not known curative (see “Patients and Methods” for ex- planation).

colostomies in all patients (4.3 percent), and in the high stoma group, transverse or loop sigmoid colostomy was used in 51.6 percent and cecostomy or loop ileostomy in 11.6 percent, for a total of 63.2 percent. Patients in both groups with intraperitoneal anastomoses had much lower covering stoma rates. They also had similar rates of anastomotic dehiscence and mortality: low stoma group 4.8 percent and 6.7 percent, respectively; high stoma group 4.5 percent and 4.3 percent, respectively. Only 2 of the 312 patients in the low stoma group had a stoma (both with transverse colostomy). Forty-five of 326 patients in the high stoma group (13.8 percent) had a covering stoma, of which 28 patients had transverse or loop sigmoid colostomy, and the remaining 17 patients had a cecostomy or loop ileostomy. It should be noted that the studies started in 1976 and concluded at the time when anastomotic stapling devices were being used more regularly. However, only 21 patients were treated with these instruments and therefore this small subgroup has not been studied further.

528 The American Journal of Surgery


TABLE III Resulfe for ExtraperItoneal Anaefomoses (valwe lndkated are percentages)

Patients Patients With Without Overall

Qoup stomas stomas Results*

Low stoma (n = 141)

% of patients 4.3 95.7 . . . Leak rate 16.7 7.4 7.6 Mortality 0 3.7 3.6

High stoma (n = 155)

% of patients 63.2 36.6 . . . Leak rate 22.4 15.6 20 -w 7.5 10.5 6.4

l Comparisons of low versus high stoma groups (p <0.05).

Comments

Over the last 25 years there have been many pub- lished reports regarding the complications and mortality of temporary colostomy associated with colorectal resection [4,5-101. However, the indications for surgery in these reports were most frequently some type of emergency presentation requiring urgent operation (bowel obstruction, perforation, or trauma). Much less has been written about the specific indications for carrying out a protective-covering stoma at the time of elective colorectal resection. Maingot [11] has stated that a covering colostomy should be carried out “if (i) there is any tension on the suture line of the anastomosis; (ii) the blood supply to the bowel ends engaged in the anastomosis is doubtful; (iii) there is any marked soiling of the operation field, or (iv) the bowel preparation has failed.” However, he provided no evidence to validate these recommendations. More recently Goligher et al [12], in a controlled trial of anastomotic technique in anterior resection of the rectum, used a transverse colostomy in most of his patients. In a later report [13] he stated that “perhaps the main value of (temporary transverse) colostomy at the conclusion of an anterior resection is not so much to prevent dehiscence as to facilitate its management if it should occur.” However, Goligher has acknowl- edged that his practice is different from that of others by referring to a personal communication with Beahrs of the Mayo Clinic who has used covering stomas only rarely.

Surprisingly we have found only one report that has set out, in a controlled study, to ascertain the benefit or harm that a temporary colostomy may contribute in elective low anterior resection [14]. In that report, an end-to-end stapling device was used to fashion the anastomosis. The study demonstrated that stomas are not required on a routine basis and are of questionable value even for the small number

Stoma

No Stoma

Low Stoma Group n=141

Leak Rate 16.7% Mortality 0.0%

High Stoma Group n= 155

63.2% 36.8 %

Leak Rate 22.4% 158% Mortality 75% 10.5%

of patients in whom a “vulnerable” anastomosis is identified. The results in the present study demon- strate that even for hand-sewn anastomoses, clinical practice varies widely among surgeons with a range of covering stomas for extraperitoneal anastomosis being less than 5 percent to more than 75 percent, depending on the individual surgeon’s decision. The elective nature of the surgery and the very consid- erable similarity of the patient-related and tumor- related factors in the population do not explain the observed surgeon-related differences in clinical practice and outcome.

We conclude that much of the surgeon-based variation derives from a spectrum of clinical opinion about the indications and thresholds for the use of a covering stoma. It is also possible that we have identified a true difference in the occurrence of vulnerable anastomoses caused by technical difficulty as experienced by some surgeons and based on a variation in technical skills. There have been some recent reports of techniques to improve the construction and ease of closure of transverse colostomies [15,16] and a renewed suggestion that tube cecostomy may be of value under these circumstances [17]. It has also been suggested that by reducing the sodium content in the diet, transverse colostomy function may be improved [IS]. Despite these technical and management suggestions, a more attractive objective for the future is to reduce the number of patients who undergo a concomitant covering stoma.

volume 147, April 1984 527

Fielding et al

There have been strong opinions voiced supporting this concept [8,19], and this paper provides some objective information which should help place these ideas on a firmer footing. It is important that we not make over-simplistic recommendations based on these figures. It could be argued that results might become worse if surgeons in the high stoma group simply change their practice and use protective stomas less frequently. However, such a suggestion would not explain the higher anastomotic leak rates seen in the high stoma group patients who did not undergo a covering stoma, in whom the surgeon must have considered the anastomosis to be sound. Fur- thermore, it is also possible that the results observed can be explained on the basis that the stoma itself may contribute to the higher leak rate and thus mortality, but it would be wrong to assume that this is the mechanism for the worse results in the high stoma group patients. Whatever the theoretical result of a sudden change in practice, a more cautious ap- proach should be taken. We suggest that all surgeons become aware of their own clinical and subclinical (radiologic) anastomotic leak rates by investigating them 10 days after resection by water-soluble contrast enema. If and when a leak rate, particularly for anastomoses without a covering stoma, is brought to an acceptable level, say 5 percent of patients, then a gradual reduction in the usage of covering stomas would be expected to be associated with an improved overall result. If this view were taken, particular at- tention would need to be paid to the technical details of carrying out a postoperative water-soluble contrast enema. The traditional barium enema rectal tube should not be used. Instead, a simple 10 F. Jakes or Foley catheter without balloon filling should be in- serted into the lower rectum and, under fluoroscopy, a minimum volume of contrast material (approxi- mately 40 ml) should be instilled into the rectum to bath the anastomosis to determine the presence or absence of total anastomotic integrity.

The factors leading to a low anastomotic leak rate are well known and are the antithesis of the items just mentioned which have been suggested as the indications for a covering stoma. We may infer from the high rates of anastomotic leakage, particularly in the high stoma group, that these well-established criteria are not being fulfilled for a substantial segment of the population currently undergoing colonic resection. Clearly this is an unfortunate situation which en- courages those who support the proposition of colorectal subspecialization. Because much of the cost of hospitalization is attributable to the length of hospital stay, there would be a substantial cost savings per treatment (20 to 25 percent) if the outcome for those surgeons in the low stoma groups could be transferred to their colleagues who frequently use covering stomas. Furthermore, in electively treated patients, the concept that a defunctioning stoma is only rarely required in conjunction with an extrap-

eritoneal anastomosis, should become the accepted norm of all who undertake colorectal resection.

Summary

A prospective multicenter study of the management of large bowel cancer recorded the results in 4,500 patients in whom 2,056 have had an elective colorectal anastomoses. Of these patients, 15.8 percent had a synchronous covering stoma to protect the anastomoses. Although the anastomotic leak rate was high in patients with a stoma, no overall differences were observed in mortality between those patients who had a covering stoma and those patients who did not (7 percent and 6.1 percent, respectively). How- ever, when surgical policies were analyzed, clinically large and statistically significant differences were found. Some surgeons frequently used a covering stoma for low anterior resection whereas others only rarely did so. The differences in anastomotic leak and mortality were 20 percent and 7.8 percent, and 8.4 percent and 3.6 percent, respectively. We conclude that all surgeons should know their own clinical and radiologic anastomotic leak rate. If and when this figure becomes low (less than 5 percent), the covering stomas will become necessary except for the very rare and difficult case.

Acknowledgment: We gratefully acknowledge the co- operation of the surgeons who took part in the study and their associated colleagues: D. W. Blair, A. I. Davidson, J. Engesett, P. R. Jones, J. Kyle, N. G. Matheson, S. S. Miller, and G. Smith, Aberdeen Royal Infirmary, Aberdeen; N. Baker and 0. D. Morris, Ashford General Hospital, Ash- ford, Middlesex; A. A. Gunn and D. A. D. Macleod, Ban- gour General Hospital, Bronxbourne, West Lothian; W. A. Anderson, K. Callum, P. Goodall, G. Harrison, S. G. Hollander, and D. R. Thomas, Derby Royal Infirmary, Derby; J. W. P. Bradley, T. J. C. Cooke, J. V. Piper, J. E. L. Sales, and C. G. Scorer, Hillingdon Hospital, Hillingdon, Middlesex; J. Chamberlain, A. H. Petty and C. W. Ven- ables, and R. G. Wilson, Newcastle General Hospital, Newcastle upon Tyne; H. B. Devlin and A. Peel, North Tees General Hospital, Stockton-on-Tees; A. G. Cox, A. Elton, A. E. Kark, J. Lewis, and D. Pinto, Northwick Park Hospital, Harrow, Middlesex; A. Davidson, Perth Royal Infirmary, Perth; G. L. Bohn, D. Goodwin, C. S. Kirkham, C. Latto, M. M. Ross, and G. Rothie, Royal Berkshire Hospital and Battle Hospital, Reading, Berkshire; R, H. Grace, Royal Hospital, Wolverhampton; A. Chant, Sir James Fraser, S. Karran, and J. Webster, Royal Southants Hospital, Southhampton; P. H. Dickinson, L. V. Fleming, F. D. Hindmarsh, I. D. A. Johnston, I. F. McNeil, and R. Taylor, Royal Victoria Infirmary, Newcastle Upon Tyne; A. S. Chilvers, J. M. Edwards, B. P. Flannery, A. G. Nash, and B. W. Wells, St. Helier Hospital, Carshalton, Surrey; P. R. Hawley, H. E. Lockhart-Mummery, C. V. Mann, Sir Alan Parks, J. Thomson, and I. Todd, St. Mary’s Hospital, London ECI; H. A. F. Dudley, H. H. Eastcott, G. Glazer, and J. R. Kenyon, St. Mary’s Hospitals, London; A. V. Pollock, Scarborough Hospital, Scarborough, Yorkshire; S. Desai, M. Horwich, C. Maddox, and A. S. Oscier, Scun- thorpe General Hospital, Scunthorpe, Lincolnshire; J.


Jenkins, F. P. McGinn, T. Rowntree, and C. J. Smart, Southampton General Hospital, Southampton; P. Cassell, R. Ramsay, and E. J. Williams, Wexham Park Hospital, Slough, Buckinghamsire; D. Cairns, J. L. Grogono, B. Higgs, P. H. Lord, Wycombe and Amersham General Hospitals, High Wycombe and Amersham, Buckingham- shire. We are also most grateful to Professor H. A. F. Dudley for his help and encouragement in the establish- ment and continued viability of the Large Bowel Cancer Project.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

10.

17.

18.

19.

References

Fielding LP, Stewart-Brown S, Blesovsky L, Kearney G. Anas- tomotic integrity after operations for large-bowel cancer: a multkentre study. Br Med J 1980;281:41 l-4.

Nie NH, Bent DH, Hull CH. SPBS: statistical package for the social sciences. New York: McGraw-Hill, 1970:104-l 1.

siegais.~staustkxforthebehevioualsckMces. lnternetional student edition. New York: McGraw-Hill Koga- kualion, 1956: 104.

Bozzetti F, Nava M, Bufalino R, et al. Early local complications following colostomy closure in cancer patients. Dis Colon Rectum 1963;28:25-9.

Chapius P, Killingback M. Colostomy closure: technique and morbkfity. Aust’NZ J Surg 1979;49:363-8.

Varenell J, Pembetton LB. Risk factors in colostomy closure. Burgery 1981;89:683-6.

Abrams 9, Alsikafi F, Waterman N. Colostomy: a new lode at morbklity and mortality. Am J Surg 1979;45:462-4.

Winkier M, Volpe P. Loop transverse colostomy. Dis Colon Rectum 1982;25:321-8.

Freund HR, Raniel J, Muggia-Bulam M. Factors affecting the morbidity of colostomy closure. Dis Colon Rectum 1982; 25:712-5

Garber H, Morris D. Eisenstat T, Coker D, Annous M. Factors influencing the morbidity of colostomy closure. Dis Colon Rectum 1982;25:464-70.

Meingot R. Abdominal operations. 5th ed. New York: Appleton Century Crofts, 1969: 1704.

Golie JC, Sirnpkins KC, Lintott DJ. A controlled comparison of one-two layer techniques of suture for high and low colorectal anastomosis Br J Surg 1977;64:609-14.

Goligher JC. Surgery of the anus, rectum, and colon. 4th ed. London: Bailliere Tindall, 1$80:563, 605.

Graffner H, Fredtund P, Olson S, et al. Protective colostomy in low anterior resection of the rectum using the EEA stapling instrument. A randomized study Dis Colon Rectum 1983; 26:87-SO.

Eng K, Locelio A. Simplified complimentary transverse colostomy for low colcfectal anastomosis. SW Gynecol Dbstet 1981;153:735.

Browning GGP, Parks AG. A method and results of loop co- kxtomy. Dis Colon Rectum 1983;28:223-6.

Wolff LH, Wolff WA, Wolff LH Jr. A m-evaluation of tube-cecostomy. Surg Gynecol Obstet 151:257-S.

Wright HK. Improving transverse colostomy functioning. Am J Surg 137~475-7.

Thompson JS, Hodgson PE. Colostomy-are there alternatives? Nebr hkd J 68:47-50

Elton Cahow (Woodbridge, CT): It’s my impression from the literature and our own experience at Yale that it’s unusual for surgeons, in the United States at least, to use covering proximal stomas during resection of the abdominal colon. If they do, it is usually in anticipation of an anastomotic leak. It would seem more reasonable in these

vellum 147, April 1984


cases, rather than to construct an anastomosis and bring out a proximal stoma, to bring out the transected proximal end of the bowel as an end colostomy and do the definitive anastomosis at a subsequent operation. Two operations are always necessary. But if a leak occurs even when protected by a stoma, the patient may require more operations for correction of intestinal obstruction, to drain abscesses, and to correct stenosis of the anastomosis. I think the protected proximal stoma should be considered primarily, as it was pointed out today, when one encounters a technically difficult low anterior anastomosis where a suture line leak is a real possibility. It wasn’t mentioned, but it was brought out in the paper, that I was given to review, that not only double-barrel truly diverting stomas were used, but also loop stomas, loop ileostomies, and cecostomies. I think that if proximal decompression is necessary, a totally diverting type of stoma is probably preferable. It doesn’t seem reasonable to accept cecostomy or a loop stoma to be as pro- ficient in diverting the fecal stream. The fact that the overall leak rate was two to three times higher in patients with a proximal stoma than in those without one raises several questions. Were these proximal stomas preferred in anticipation of a leak and imperfect anastomosis? If so, why was the anastomosis carried out to begin with? Why wasn’t the attempted, anastomosis abandoned and the proximal end brought out as an end-colostomy? On the other hand, are the surgeons in the high stoma group, who have twice the leak and mortality rate as the low stoma group surgeons, capable of performing a safe anastomosis? What happens when they close their proximal stomas? Do they have a higher leak rate there also? One set of figures seemed contradictory to the conclusion of the paper. In Table I, I noted that an anastomotic leak without a proximal stoma carried a 26 percent mortality rate as opposed to one with a proximal stoma with only an 8 percent mortality rate. One would draw the conclusion here that a proximal stoma is protective and should be included in the operative procedure. Finally, how should a subclinical or asymptomatic leak be treated?

Steven E. Hedberg (Boston, MA): I’ve been analyzing the records of all patients operated on by three surgeons at our hospital who use staplers frequently, actually in about half of their last 774 gastrointestinal anastomoses that we’ve studied. When I saw Dr. Fielding’s abstract, I was prompted to reorganize our data to see how the results compared with ours. In colorectal anastomoses, as expected, the rates of covering stoma varied widely among the three surgeons who were studied, but the morbidity and mortality rates due to sepsis were inversely proportional to the rate of covering stoma rather than the reverse, as Dr. Fielding has indicated. We saw no clinically anastomotic failures when a covering stoma was used. Excluded from this analysis were four patients who had Hartmann procedures in whom the bowel was so full above the proposed stoma that the covering stoma would not have diverted sufficiently. These other patients were adequately pre- pared.

When we took all patients together with left colon, sigmoid, and rectal anastomoses in whom a covering stoma was a potential, the results were even more suggestive, with the covering stoma rate varying from 2.4 to 5 to 24 percent, actually inversely proportional to the related mortality. All of the patients who died did so as a result of anastomotic failure. No patients died as a result of having had a stoma.

529

Fielding et al

All 20 stomas were created for the usual indications: three for obstruction, four for sepsis, two for steroids, seven patients had heavy preoperative radiation to the anastomotic area, one had a failed stapler, and so on. It’s interesting that of the three patients who died, two of them also had indications for creating a covering stoma, but they were ig- nored. The only other risk factor that I could find in the third patient who died, interestingly, was that he had not be given any perioperative antibiotics for some reason or other.

Finally, it is important to note that the disadvantages attributed to stomas were not apparent in our series. There was no significant prolongation of total hospitalization whether stomas were created or not, and this includes the second hospitalization in the 21 percent of patients who had stomas. Furthermore, although high morbidity and mortality rates are often advanced as a reason for not doing stomas, it should be noted that in the last 56 consecutive colostomy closures performed in our hospital on the service of one of these surgeons, going back to 1967, there has not been a single anastomotic failure, wound infection, or dehiscence, and no morbidity whatsoever even though all of the patients had primary closure.

In conclusion, it may be true that real men don’t eat quiche, but it is not true that real surgeons don’t make covering colostomies. I hope the somewhat intimidating thrust of Dr. Fielding’s statistics will not deter any surgeon here from making a covering colostomy when he thinks there may be an indication. I compliment Dr. Fielding on

an interesting report. I just don’t think the statistics apply on this side of the Atlantic.

L. Peter Fielding (closing): What this study has shown, I think, is that there are very wide differences not only in surgical skill, but also in the indications by which people do one thing or another. I don’t believe that the surgeons in the high stoma group used the normally accepted indications for stoma placement. I agree that in the high-risk patient group, which probably represents only a little more than 5 percent of the total population, something other than a primary anastomosis without coverage should be performed.

One of the other thoughts that occurred to me while I was hearing Dr. Hedberg’s data was that he was looking at a very mixed patient population as far as the overall indications for stomas are concerned. I very carefully pointed out that I have reported an entirely elective group of procedures. In these circumstances, I think that if a patient is being operated on by an experienced and competent surgeon who understands the true indications, on which we would all agree, then the number of such patients requiring a stoma should be less than 10 percent.

Finally, I entirely agree with Dr. Cahow; if there is a doubt about an intraperitoneal anastomosis, the two ends should be brought out separately. In my study I didn’t mention the overall extra hospital stay associated with stoma closure. It was about 25 percent longer than for those patients who did not have a covering stoma.


Documents

Covering stoma for elective anterior resection of the rectum: An outmoded operation?