Page 177 - Prevention and Treatment of Incisional Hernia- New Techniques and Materials

and rarely abdominal compartment syndrome(90, 91). The use of preoperative pneumoperitoneum or botox might be implemented in some cases although evidence is limited(44, 92-94).
In LIH repair overall postoperative complications are higher compared to smaller incisional hernia repair. The increased morbidity is partly caused by patient characteristics, such as more serious and extensive primary diseases, systemic collagen disease and the increased intra-abdominal and pulmonary pressure after repair. Frequently a large wound bed is created, increasing the risk of wound complications.
Limitations
The rst limitation of this review is the lacking consensus on the de nition of LIH (table 1). The criteria for LIH as proposed by the European Hernia Society (EHS)(21) were used: size of hernial ori ce 10cm or more in any direction. Since some authors only report the hernia surface, articles describing hernias over 100cm2 were also included. Recently a consensus paper on de nition of complex abdominal wall hernias is published, but these detailed criteria are often not mentioned in articles and are especially usable for (future) prospective studies(27). That’s why we used a more simple de nition to di erentiate between ‘simple’ and ‘complex’ LIH in this review.
Secondly, the follow-up between the studies included in this review varied from 1 to 10 years. Due to the delay between hernia repair and the development of a recurrence the period of follow-up is important. Since short term follow-up might cause underestimation of recurrence rate, only articles with a mean follow-up of at least 1 year were included. Still, comparing techniques for recurrence rates is di cult with di erent follow-up periods. For this reason the recurrence hazard per year for every repair technique was calculated. This model assumes an equal hazard for getting a recurrence during each month of follow-up of the study. But this is not consistent with the natural pattern of recurrence, and as a consequence the monthly or yearly hazard does not resemble the true percentage of recurrence. Furthermore, the assumption in the GLM is that count follow Poisson distributions. Overdispersion is quite common, and so one has to keep in mind that standard errors will be too optimistic. However, we think that the GLM is a useful tool in comparing recurrence rates for studies with di erent follow-up periods.
8
Review treatment large IH
175