To the Editor:
We read the recently published study by Lo Scalzo et al. (Reference Lo Scalzo, Ratti, Jefferson, Bernardini and Cerbo4). The study is a short version of a more extensive report published on the Italian Ministry of Health Web site (1). A careful evaluation of both documents reveals several critical issues that cast doubts on the correctness of their conclusions. Usually, decisions about clinical practice should be based on levels of evidence in literature and on the strength of recommendations issued in the guidelines of international scientific societies, many of which rely on the Oxford system (Oxford Center for Evidence based Medicine) (http://cebm.jr2.ox.ac.uk/docs/level.html). As far as diagnostic tests are concerned, the Oxford system assigns the highest level of evidence to systematic reviews supported by homogeneous results. The same system assigns the second highest level of evidence to the “independent blind comparison of an appropriate spectrum of consecutive patients all of whom have undergone both the diagnostic test and the reference standard,” while the expert's opinion ranks at the lowest level. Surprisingly, the authors flatly dismissed all sequential studies that fulfill the criteria of independent blind comparison, based on the fact that they were not randomized. At the same time, they cited as source data the results of a questionnaire presented to the attendees of an Italian meeting, never published in literature, which is plainly an “expert opinion.”
As early as 2006, there was a strong (and not “anecdotal” as Lo Scalzo writes) evidence of the benefits of wireless capsule endoscopy (WCE) in the diagnosis of small bowel (SB) diseases. To date, more than 1,000 studies concerning WCE have been published, homogeneously showing the diagnostic dominance of WCE over the other standard methods used before its introduction, and the comparability of its diagnostic yield with that of the most recent diagnostic techniques. Many of these studies represent reference points for the scientific community and served as a basis for Guidelines formulation (Reference Rey, Ladas and Alhassani6). Even if we limit our attention to the single work (Reference de Leusse, Vahedi and Edery2) considered as methodologically adequate, when only SB lesions are considered, a statistically significant difference in terms of diagnostic yield emerges in favor of WCE when compared with push-enteroscopy.
Sequential comparative studies were also dismissed by the authors on the assumption that such studies might be biased because of the time interval between the performance of WCE and the comparator test. The argument is that angiodysplasias might quickly modify their morphology. According to this assumption, vascular lesions might vary, or even disappear, between procedures, and “time” might represent a relevant factor affecting the results. The authors should have provided references for this assertion. They did not, and to the best of our knowledge, there is no reference confirming this assumption in the literature. A careful evaluation of the document published on the Italian Ministry of Health Web site reveals the reason for this misunderstanding. When Lo Scalzo refers to vascular lesions, she refers to Dieulafoy's lesions (Reference Norman, Gostout and Balm5), which constitute only a tiny minority (less than 1 percent) of all vascular lesions and have a peculiar clinical history as they bleed intermittently and are not visible in the interval between bleeds. This does not apply to other vascular malformations which remain unmodified over time or even worsen (Reference Gordon, Watkinson and Hodgson3). Such a misconception might have influenced the elaboration of the entire report.
Another unfounded statement is that WCE has not been sufficiently compared with other comparative methods capable of exploring the entire SB mucosa. The “real” comparator should be able to explore the entire SB. Only intraoperative enteroscopy can consistently explore the entire SB. However, this procedure requires a surgical intervention, and no ethics committee would have authorized a study comparing WCE with this procedure. Push-and-pull enteroscopy could also potentially explore the entire SB. However, a complete SB examination is not guaranteed even when a double intubation (oral and anal), is provided. Studies that compared the completion rate of WCE and push-and-pull enteroscopy consistently showed a superiority of WCE. As far as radiological tests are concerned, although these tests can visualize entirely the SB, they cannot detect flat vascular lesions and comparative studies are consistently in favor of WCE.
The intent to add context-specific data to literature review is appealing. However, the questionnaires sent to the Centers performing WCE were inaccurate: that is, definitions of technical problems and adverse events were missing. Among adverse events, only “intestinal occlusion” was mentioned. All other adverse events were listed in an entry named “Others.” Furthermore, the authors considered a capsule retained when it was not expelled within 1 day, whereas retention by definition occurs when the WCE capsule is not excreted within 15 days. This might have led to misinterpretations and to an overestimation of the retention rate.
As far as costs are concerned, the authors only mention that WCE is expensive. Information about cost-effectiveness and costs of alternative procedures are not provided. A cost assessment of the diagnostic-therapeutic strategies including WCE would have been much more informative than the analysis of the single procedure.
In conclusion, we believe that the study by Lo Scalzo et al. is the result of a subjective and biased analysis of the literature. The criteria used to select and refuse published studies are arbitrary and do not comply with the accepted methodology for the evaluation of scientific evidence for diagnostic procedures. The context-specific data are hardly representative of the Italian situation because the survey was unclearly structured, reflecting a lack of clinical insight by the authors.
CONFLICT OF INTEREST
R de Franchis has received honoraria from Given Imaging and travel expenses from unidentified sources. G Costamagna has received payments for educational presentations from Given Imaging and grants or pending grants from unidentified sources. The other authors report having no potential conflicts of interest.
