Discussion
This study highlights several features of the debate pertaining to the use of mesh in the repair of PEH. First, the main reason for the debate is that the reported recurrence rates of PEH vary widely for both primary repair and mesh repair. Table 2 shows the reported recurrence rates after primary repair varying from 0% to 59%, and for mesh repair, from 0% to 54%. Second, the consequences of recurrence also vary. The reoperation rates for both primary and mesh repair range from 0% to 57%, and the rate of severe adverse outcomes from these reoperations (defined as the need for gastric or esophageal resection) varies from 0% to 33%. In addition, the rate of recurrences being symptomatic also varies from 0% to 67%. Even if there is not PEH recurrence, erosions and stricture because of mesh have been reported in up to 0.5% of cases, although it is very likely that this outcome is underreported. Even without recurrence or erosions, patients with prosthetic mesh repair may have significant problems with dysphagia,[59] which may not be as significant with biological mesh.[60] Given these wide ranges of outcomes, it is not surprising that no consensus has emerged as to whether or not to use mesh in the repair of PEH.
Interestingly, there has only been 2 randomized controlled clinical trials comparing mesh repair with primary repair of PEH. Frantzides et al[61] performed a randomized trial of 35 patients with PEH, with 17 patients receiving primary repairs and 18 patients receiving mesh repairs with polytetrafluoroethylene. Postoperative complications were similar, and at 6 months, 3 primary repair patients (17%) had recurrences, whereas none of the mesh patients did. Oelschlager et al[25,42,55] randomized patients to primary repair and mesh repair using a biological procine small-intestine submucosa mesh. Initially, their first report showed fewer recurrences at 6 months for the mesh group compared with the primary repair group,[25] but their follow-up report at nearly 5 years showed similar recurrences.[42] Interestingly, few of these recurrences in either group were symptomatic.[55] Therefore, in the randomized trials published to date, no clearly superior technique has emerged.
Proponents of mesh repair will argue that the type of mesh is important, and a wide variety of meshes have been used. These include prosthetic mesh made of polypropylene, polytetrafluoroethylene,[19] "DualMesh,"[62] and "TiMesh."[32] The other major type is biological mesh with procine small-intestine submucosa and aceullar human dermis.[12,17,19] Erosions, dysphagia, and strictures have been associated with prosthetic meshes.[12,17,19,48] This has led some authorities to recommend biological mesh as an alternative.[17,60]
Recent reports have focused on the consequences of having to reoperate on patients with recurrent PEH. These can be quite challenging operations and these challenges can be magnified by the presence of mesh. Zaninotto et al[30] and Parker et al[49] reported the need to perform resections of the stomach and/or esophagus in >=30% of patients undergoing reoperations for recurrence after initial repairs with mesh. Another report of mesh erosion into the aorta is quite dramatic.[36] However, as with other aspects of the data in this debate, it is unclear whether these data are the norm or aberrations. Nevertheless, anyone who had to reoperate on a patient who has had prior surgery in the esophageal hiatus can attest to its difficulty. Compounding this confusion is that gastric or esophageal resection may occur in the absence of mesh.[49] In the long term, there is a concern about mesh erosion. The reported incidence is low at 0.5% or less.[48] However, it is likely that this may be underreported. The consequences of mesh erosion or stricture are quite severe, often necessitating reoperation.
When assessing utility scores with either mesh or primary repair, at wide ranges of recurrence and reoperation rates, the differences between the two are overall small. For example, the worst score is a mesh recurrence rate of 50%, with 50% of this requiring reoperation, yet the utility score is 97.47, just over 2.5 units worse than the most preferred state of 100. What this means clinically is that the probabilities of this state occurring are not as preferable, and the relative lack of preference lies somewhere between a certainty of a reoperation with a good outcome and a certainty of a perfect outcome. Compare this with the consequences of an esophagectomy that will produce quality of life scores 20% or more lower than the highest possible quality of life scores and, therefore, is a far less preferred outcome state.[56–58] Therefore, it may be that the reason the debate is unresolved is that there is not much difference between mesh and primary repair, once the probabilities of the various outcome states and their relative utility scores are rolled back to the original decision to use or not use mesh. Although cost was not part of this decision analysis, one wonders whether the additional costs of mesh, especially biological mesh, is worth whatever modest improvement in recurrence rate exists.
Obviously, there are limitations to this analysis. First, the data are mostly from case series, with only 2 randomized trials reported. Therefore, it is difficult to determine how much of the variation is because of the use of mesh or not, compared with other factors, such as patient selection and surgeon's skill. Second, there is a wide variety of repair techniques used for both primary repair and, especially, mesh repair. It is impossible to control for these variations, and in fact, if the "ideal" repair with any method were found, we may have to revisit the results of this analysis. Third, no attempt was made to control for the other "adjuncts" of PEH repair, namely a Collis lengthening procedure for the short esophagus or a gastropexy. It truly is unclear how much these surgical adjuncts contribute to lowering recurrence and reoperation rates. Fourth, we made no evaluation of symptomatic or quality of life improvement, nor of postoperative untoward events such as dysphagia or bloating. Recent reports have suggested that many recurrences after PEH repair are small, sliding-type hiatal hernias that are asymptomatic.[35,39,55] Therefore, one can argue that the conversion of a symptomatic PEH to an asymptomatic hiatal hernia is truly not a clinical failure. Fifth, the utility scores were based on the authors' best assessment of the given health state. In addition, understanding the utility scores based on preferred health states is not particularly intuitive. One can argue that assigning a utility score of 95 to someone who has had a reoperation may imply that a recurrence requiring a reoperation is not quite the failure that it is. Although this is true, the advantage of assigning a utility score to a health state is that it rolls a myriad of implications of an outcome into a quantifiable number that can be compared with other health states explicitly. Nevertheless, this process may be considered somewhat arbitrary. Last, we have made no attempt at determining cost-effectiveness.
In conclusion, the differences between primary repair and mesh repair are small. So small, in fact, it can be argued that it is clinically inconsequential. This decision analysis demonstrates that there is no compelling reason to choose one over the other. Individual surgeons may have their own data with respect to their own recurrence and reoperation rates, and they can use these data in the sensitivity analysis (Table 4 and Table 5) to make their own determination. Costs may be a significant determination of choice.
Annals of Surgery. 2013;257(4):655-664. © 2013 Lippincott Williams & Wilkins
