Portal de Periódicos da CAPES

Total Mesorectal Excision – A Fashionable Fable or a Timely Truth

1999; Karger Publishers; Volume: 16; Issue: 5 Linguagem: Inglês

10.1159/000018748

1421-9883

William H. Isbister,

Anorectal Disease Treatments and Outcomes

Total mesorectal excision (TME) is a technique that is becoming increasingly utilized, and in some countries it is said to be almost universally practiced now [L. Påhlman, pers. commun.]. This trend continues despite cautions that the technique remains unproven and inconclusive [1, 2]. The postoperative morbidity, including anastomotic dehiscence, is considerably higher following TME than is the morbidity following conventional rectal resection and unless it can be shown that TME really confers a survival or improved quality of life benefit for the patient the risks of the technique may not justify its implementation. This paper attempts to reevaluate the most recent data upon which TME is 'justified' [3].In 1982, Heald et al. [4]argued that a meticulous dissection of the mesorectum should be performed in all patients undergoing restorative resection of the rectum. In order to lend some 'academic respectability' to the thesis, an academic 'independent validator and assessor' was invited to re-present Heald's data [3]. It is still, for the most part, upon this paper from Basingstoke that the case for TME is 'justified' and it is now important to review the paper and at least seek answers to some of the questions that arise from it. By way of introduction, MacFarlane et al. [3]stated that 'the temptation to dismiss our low local and overall recurrence rates as a statistical aberration... was scrutinized objectively'. Unfortunately the reader was given little information regarding the validity of the population studied or the analytical techniques used to 'rework' the data and thus it was rather difficult to evaluate the 'new' data objectively. More questions in fact were raised than answered. Was it appropriate to attempt 'to determine the standards of outcome that are acceptable' for the management of patients with rectal cancer on one's own data? Surely this was a task for their peers? Could an oncologist who had been part of the team for 14 years be expected to provide 'further objective corroboration'? How 'objective' would such an involved member of any team be? Patients were 'examined' and CEA tested at regular intervals in order to detect local recurrence. How were they 'examined'? Did they have endo-anal ultrasound, CT or MRI examinations or were they simply clinically examined? The method of examination used to detect local recurrence would have materially affected the time of diagnosis if not the diagnosis itself and could easily be one of the reasons for the low local recurrence rates reported. All deaths were 'documented' but did all the patients who died have postmortem examinations? If not, it would be impossible to state that patients did not die with progressive asymptomatic disease. Since the study end points depended upon recurrence and 'cure', clarification of the above issues was mandatory.The technique of TME places great importance on the anterior and posterior planes of resection, but in the paper little information was given in relation to the lateral resection margins which are known to be important determinants of loco-regional recurrence [5, 6]. The re-analyzed 'actuarial local recurrence rate' was 4%. But what does this mean? Is the figure age-corrected, does it relate to all patients, have postoperative deaths been excluded? 'Recurrence in cases of Dukes' C carcinoma has fallen from 42% to 32% at the 5-year point', according to MacFarlane et al. [3], and this figure is similar to the 30% reported from Gateshead for patients with Dukes' C tumours where TME was performed [7]. These rates are however rather similar to the rates reported from series where TME has not been undertaken [8, 9].TME was said to be more difficult and time-consuming... and 'avoids the autonomic nerves by defining the place within them...'. This 'holy plane' or 'place' [10]was dissected under 'direct vision throughout' according to the authors, but anyone familiar with rectal surgery in a fat male with a narrow pelvis and a large tumour of the mid-rectum would know that it would be impossible to even see to sharply dissect under these circumstances. Is this a clue to population bias? It is known that primary tumour size affects both survival and local recurrence rates following TME resection [7]so that results from a series composed predominantly of small sized tumours would be expected to be lower than those from the population at large. Further evidence of population bias is indicated by the statement that 'more abdominal-perineal excisions were undertaken by other surgeons'. Why were 9 such patients included at all?The supporters of TME emphasize the need to avoid violating the 'fascial envelope', it being assumed that tumour cells respect the 'envelope' and do not traverse it. Can this same assumption be made regarding the plane between the muscularis propria of the rectum and the distal mesorectum where no such fascia exists? In this part of the operation the surgeon is simply sharply dissecting between the fat and the muscle of the rectal wall – do the tumour cells in the mesorectum remain confined to the fat during this process?One of the costs of TME is an increased blood transfusion requirement and this could favour local recurrence. According to Sugarbaker [11], in patients resected for 'cure', too many local recurrences are found for all of the local recurrences to be due to inadequate primary excision of the tumour. Tumour cells are 'spilled' from transected veins and lymphatics during dissection and the more the blood loss the greater the chance of spilling tumour cells. These cells become 'entrapped' in a fibrin matrix which forms at the site of resection in the peritoneally denuded pelvis. Blood transfusion itself may be associated with decreased survival in colorectal surgery also and thus one might expect not only a higher local recurrence rate following TME rather than the low rate claimed, but also a decreased rather than increased rate of survival.No mention is made of the other 'costs' of TME such as increased hospital stay, the morbidity of stoma closure, the extra hospital costs of stoma closure, etc. The cost of performing TME with its added morbidity and 'costs' in patients who did not have islands of tumour cells in their mesorecta was not discussed. These patients would have been labeled Dukes' C in the final analysis but would have become Dukes' C simply because of the finding of extra nodal tumour cells in the mesorectum. This process, known as stage migration [1], was not considered and it is not known how many stage migrated patients there were in the Basingstoke series. These patients are the key to the debate because theoretically all other patients underwent TME without benefit.The surgery for rectal cancer has almost gone 'full circle' [12]. First there was a 5-cm rule which was reduced to 2 cm because it was shown that intramural spread was confined within this length except in the most advanced patients [13]. Now, in order to reduce the high morbidity of TME (16–17% leak rate [3, 7]), it has been suggested that a modified form of TME, subtotal mesorectal excision (SME) (mesorectal excision limited to 5 cm in a caudad direction), be applied to all but the very lowest tumours and that the 5-cm rule be re-introduced for all other rectal resections [12]. Implementation of this rule must necessarily increase the number of patients undergoing abdomino-perineal resection of the rectum and therefore the total number of permanent colostomies fashioned. This might be another rather high cost for the few patients who might benefit from the new TME.How representative of the total population was the population reported? There were 52 Dukes' A (24.4%), 88 Dukes' B (41.3%), and 73 Dukes' C (34.2%) including an unknown number of 'stage migrated Bs'. If TME had not been performed there would have been a higher percentage of Dukes' B and consequently a lower percentage of Dukes' C than actually recorded. More than 65% of the patients in the Basingstoke series were either Dukes' A or B whereas in other population-based studies the percentage of early stage tumours is not so high (table 1).It was not clear why the local recurrence rate after non-curative anterior resections in Basingstoke was so low (6%) when it is known that by definition these patients have residual progressive disease at the time of surgery. Since local recurrence is associated with other metastatic disease in over 80% of patients with progressive disease [18], one would have expected much higher mortality and local recurrence rates in the non-'cured' patients. Why should there be any survival advantage for removing residual local disease at a cellular level when there is already more advanced progressive disease elsewhere in the body.MacFarlane et al. [3]claim a '5-year tumour-free survival of 78%'. This is difficult to interpret when it is not known what the 78% is a percentage of and when the techniques used for detecting the 'tumour-free' status of the patients are not given. Thirty percent of patients operated for 'cure' have micrometastasis in the liver [19], and it is difficult to understand why these patients have not been detected in Basingstoke. It is doubtful therefore whether there is justification for the claim that it 'represents a real increase in patients actually cured of cancer'. It was also claimed that the 'finding is at variance with the notion that cancer is a systemic disease' because there were no recurrences after 5.9 years. This latter finding probably simply reflects the detection methods used and the duration of follow-up since local recurrence rates for carcinoma of the rectum have been shown to rise with time and may reach over 50% at 18 years [20].Five-year survival rates in Basingstoke were better than those reported in a New Zealand National study, where practically all cases were documented (registration rate 92%) and the surgery was performed by a variety of general surgeons [16]. The population in New Zealand, however, is somewhat different, there being twice as many patients with disseminated disease at the time of surgery, so that fewer patients would by any definition have been operated for 'cure'. This together with the variability of surgical skills but not the performance of TME [21]could easily account for the differences in 'crude' and 'relative' 5-year survival rates recorded (table 2). Survival data is very sensitive to the way in which it is calculated [22]and it is possible that if the New Zealand data were 'reworked', to include only those patients operated for 'cure', the survival figures from New Zealand, without TME, could well come closer to the Basingstoke rates.Such questions as why local recurrence rates are high after abdomino-perineal resection of the rectum, when in this situation it is likely that an adequate mesorectal excision will have been performed by even the most 'average surgeon', require explanation although the finding may simply be another example of case-mix bias as has been suggested recently [23]. Why should TME be more successful in eradicating pelvic recurrence than radiotherapy. The lowest local recurrence rates reported following such therapy [24]are several times higher than the Basingstoke rates. It could be argued that a field therapy would be more useful in this respect than local excision. If the intellectual philosophy behind TME is valid, why does the same argument not apply to high ligation of the inferior mesenteric vessels [25]?If it is assumed that those patients with tumour in the distal mesorectum (DMR) actually benefit from TME and this may not be so since circumferential resection margin (CRM) involvement may be an even more sensitive marker for poor prognosis, it is possible to calculate the number of patients needed to resect with TME to achieve this benefit. Of the 213 patients reported from Basingstoke, 73 had Dukes' C tumours. The proportion of these patients who would be expected to have tumour in their mesorecta would be between 20 and 30% [26, 27]so that of the 73 patients with Dukes' C tumours, 22 might benefit from TME. Overall, 213 patients underwent TME so that for every 1 patient who benefited, 9 were subjected to an increased morbidity. In addition, patients who have DMR involvement have a worse outcome after 4 years [27]. TME has no influence on the development of distant metastases either [28], a fact not discussed in papers supporting TME [29], so that the patients who will actually benefit from TME are likely to be very few indeed.Many of the criticisms made in the present paper have been made previously [30]but they have not yet been satisfactorily answered [31]. There may be something in the Basingstoke data which should not be dismissed, but until exact information relating to the statistics used and the population studied are made available and the criticisms which have been made and which are based on other studies are addressed, it is unlikely that this debate will progress further. Factors such as the staging system used [32], the age/sex mix of patients [33]and the inclusion or exclusion of non-cancer-related deaths [34]from the survival analysis may all affect the calculation of survival data.Heald by his own confession is a craftsman [9]and clearly he performs a meticulous pelvic dissection. My concern is not with his craft but its logic. There must be a few cases with disease localized to the margins of a 'manual extraction' and no disease elsewhere who will be rescued by TME but I contend that this number will logically be small and the increased morbidity of TME may not be worth the benefit. Heald may be right, but first he must address and explain criticisms of TME based on other documented data and, secondly, and perhaps most importantly, his thesis relating to the importance of TME should be tested prospectively and independently. There are some colorectal 'enthusiasts' and some colorectal 'nihilists', but I have no doubt that at least some colorectal surgeons have open minds and would be willing to take part in such a multicentre evaluation. Rather than have yet another paper which argues the same case on the same data in the future, let us have a definitive study which will answer the question once and for all. If TME is important and if it genuinely confers a survival advantage and if the cost of this success is not inconsequential, then the decision to proceed with TME must be based on more data than is either available or has been presented to date.Many of the colorectal surgeons of today are like the mastectomists of the sixties. Some do admit that 'surgery has accomplished all it possibly can...' [35]whilst others champion the cause of TME. Time and proper analysis did not really favour the radical mastectomists and I suspect that time and rigorous testing will fail to prove the value of TME. Let us however have a proper prospective randomized trial of TME to settle the matter once and for all.