A Giant With Clay Feet
The article by Prytz et al1 addresses a controversial topic in the colorectal surgery literature—namely, abdominoperineal procedures for rectal cancer. Interestingly, the study1 refers to the same cohorts of patients analyzed within the same time frame in a previous publication.2
There are at least three categories for discussion: (i) recall bias to determine type of abdominoperineal procedure performed; (ii) questionable indications for abdominoperineal procedures in patients with tumors not involving the sphincters; and (iii) conflicting data in publications addressing the same cohorts of patients.1,2
The first point of discussion refers to the authors’ breakdown of the series into groups based on their perception of what type of abdominoperineal procedure might have been performed. The authors stated that the type of procedure [Extralevator abdominoperineal excision (ELAPE) or non-ELAPE] was documented only in some records, whereas in a large number of the records, the authors simply guessed what type of procedure might have been carried out. Moreover, the authors could not decide what type of abdominoperineal procedure was performed in half of the records. Hence, the readers can hardly rely on the retrospective allocation of cases to one procedure or another. Such recall bias is likely to have led to unwarranted conclusions because of noncomparability of the study arms. The second concern refers to the inclusion in the analysis of patients with a rectal cancer located more than 4 cm from the anal verge.1 Abdominoperineal procedures should be indicated for patients with a suspected involvement of the sphincter muscles. In fact, procedures such as intersphincteric resections may be considered as an option in case of absence of sphincter involvement. Moreover, Prytz et al1 failed to provide the readers with the T-stage stratification of the subgroup of patients with cancer located less than 4 cm from the anal verge to prove comparability. The third category for discussion involves a back-to-back comparison of publications.1,2 In the recent article,1 ELAPE resulted in a significantly increased 3-year local recurrence rate as compared with non-ELAPE with intraoperative perforation as an important risk factor for local recurrence. In the earlier paper,2 ELAPE did not result in fewer intraoperative perforations or involved circumferential resection margins as compared with non-ELAPE. The above mentioned statements seem not only incongruent with common practice and the literature, but also not supported by the analysis of heterogeneous and noncomparable patient subgroups. Figure 1 in the article by Prytz et al1 compared 58 patients and 385 patients with cancer located less than 4 cm from the anal verge having undergone non-ELAPE and ELAPE, respectively; such subgroups are not comparable for size (P < 0.0001). Table 1 shows a mean tumor height of 6.0 and 3.0 cm for non-ELAPE and ELAPE, respectively (P < 0.0001).1 Table 1 of the previous publication2 showed 17 patients (8%) versus 176 patients (34%) with tumor location less than 2 cm from the anal verge in the non-ELAPE and ELAPE arms, respectively (P value not reported). Furthermore, the long-course radiation ratio (non-ELAPE 24% vs ELAPE 36%, P < 0.0001) represents an additional bias.
In conclusion, the readers should be less impressed with the size of the series and more attentive to the quality of the methodology.