Early use of mammalian target of rapamycin inhibitors is an independent risk factor for incisional hernia development after liver transplantation

Authors


Abstract

Incisional hernias (IHs) are common complications after liver transplantation (LT) with a reported incidence of 1.7% to 34.3%. The purpose of this retrospective study was to evaluate the risk factors for IH development after LT with a focus on the role of immunosuppressive therapy during the first month after LT. We analyzed 373 patients who underwent LT and divided them into 2 groups according to their postoperative course: an IH group (121 patients or 32.4%) and a no-IH group (252 patients or 67.6%). A univariate analysis demonstrated that the following were risk factors related to IH development: male sex (P = 0.03), a body mass index ≥ 29 kg/m2 (P = 0.005), LT after 2004 (P = 0.02), a Model for End-Stage Liver Disease (MELD) score ≥ 22 (P = 0.01), and hepatitis B virus infection (P = 0.01). The highest incidence of IHs was found in patients treated with mammalian target of rapamycin (mTOR) inhibitors (54.5%, P = 0.004). A multivariate analysis revealed male sex (P = 0.03), a pretransplant MELD score ≥ 22 (P = 0.04), and the use of mTOR inhibitors (P = 0.001) to be independent risk factors for IHs after LT. In conclusion, immunosuppressive therapy with mTOR inhibitors is an important independent risk factor for IH development after LT. To reduce the incidence of IHs, mTOR inhibitors should be avoided until the fourth month after LT unless their use is deemed to be strictly necessary. Liver Transpl 18:188–194, 2012. © 2011 AASLD.

Incisional hernias (IHs) are common complications after liver transplantation (LT) with an incidence of 1.7% to 34.3%.1-8 In the year 2008, LT was performed 5890 times in the United States, so approximately 100 to 2020 patients will experience an IH in the United States every year.9

Even though LT was first performed by Thomas Starzl in 1963,10 the first study analyzing the risk factors for IHs after LT was not published until 2002.3

The importance of IHs is often underestimated. They not only are an aesthetic problem that can affect the patient's quality of life and ability to work but also are important postoperative complications that can lead to many further complications such as abdominal pain, intestinal incarceration and perforation, and skin necrosis.11-13 Moreover, patients with IHs require surgery as a curative treatment, and this can lead to even more complications (eg, IH recurrence, hematomas, seromas, mesh infections, and foreign body reactions) and a mortality rate that can reach up to 5.3%.14

The kind of incision that is typically used in LT (J-shaped or Mercedes) can facilitate the development of IHs, which occur mainly in the midline of the abdomen and at the junction of the upper midline and transverse incisions.15 Several risk factors that can influence IH development have been identified and described: pre- and post-LT ascites in the recipient, the recipient's body mass index (BMI) and age, male sex, the type of incision, a living donor, rejection, wound infections, reoperation, pulmonary complications, steroids, and immunosuppressive therapy. In the last few years, the use of mammalian target of rapamycin (mTOR) inhibitors (eg, sirolimus and everolimus) as immunosuppressants after LT has gradually increased at the expense of calcineurin inhibitors (CNIs) because of the former drugs' preferential action with respect to renal function, blood pressure, diabetes mellitus, and neurological side effects.4, 16, 17

The purpose of this retrospective study was to evaluate the risk factors for IH development after LT with a focus on the role of immunosuppressive therapy.

Abbreviations:

BMI, body mass index; CNI, calcineurin inhibitor; HBV, hepatitis B virus; HCV, hepatitis C virus; IH, incisional hernia; LT, liver transplantation; MELD, Model for End-Stage Liver Disease; mTOR, mammalian target of rapamycin; OR, odds ratio.

PATIENTS AND METHODS

The present retrospective study was a priori approved by the local accredited medical review ethics committee. Between October 2000 and December 2009, we performed LT 444 times in 409 patients at our institution. Thirty-six of these patients (8.8%) without a minimum follow-up of 3 months were excluded from this study, and the analysis was conducted with 373 patients (91.2%). The mean follow-up was 47 ± 32 months.

No patients were lost to follow-up, which generally consisted of a monthly ambulatory examination during the first year, 4 examinations during the second year, 2 examinations during the third year, and 1 examination per year thereafter. The IH diagnosis was based on a clinical examination and was confirmed with an abdominal wall magnetic resonance imaging scan if necessary. The patients were divided into 2 groups according to their postoperative course: patients who experienced 1 or more IHs (the IH group) and patients who did not develop any IHs (the no-IH group).

Until 2008, the Mercedes incision (a bilateral subcostal incision with an upper midline extension) was preferred; thereafter, the preferred incision was the J-shaped incision (a right subcostal incision with an upper midline extension).

The incision closure technique was not recorded and was based on the surgeon's preferences; in general, however, a subcostal incision was closed with a polydioxanone number 1 running monofilament absorbable suture (PDS, Johnson & Johnson) for the peritoneum and transverse muscle, and this was followed by interrupted and coated number 1 braided polyester nonabsorbable sutures (Ticron, Syneture) or Polyglactin 910 violet braided absorbable sutures (Vicryl 1, Johnson & Johnson) for the other 2 single layers or a 3-layer interrupted suture with Ticron 1 or Vicryl 1. Median incisions were closed with interrupted Ticron 1 or Vicryl 1 sutures.

The closure of the abdominal wall was delayed in 4 cases because of intestinal edema with excessive abdominal pressure (2), bleeding that required only packing (1), or liver donor/recipient mismatching (1). The technique that we used for this delayed closure was previously described by Kirshtein et al.18

According to our policy for IH repairs in the presence of surgical indications, we performed the surgical correction as soon as possible after the IH diagnosis was made according to the patient's medical condition.

As for the surgical technique for IH repairs, we always used mesh interposition according to a sublay reconstruction for an IH less than 10 cm in diameter and according to an inlay reconstruction for a larger IH. We recently started a laparoscopic program for the treatment of IHs after LT; so far, only 2 procedures have been performed with a laparoscopic approach.

In the statistical analysis, we recorded and analyzed the immunosuppressive therapy administered during the first month after LT. The immunosuppressive therapy during the first month after LT was based on CNIs such as cyclosporine (164 patients) and tacrolimus (194 patients) alone or in association with mTOR inhibitors such as sirolimus (60 patients) and everolimus (62 patients) according to local, national, or international protocols. Induction consisted of the administration of basiliximab or daclizumab during the anhepatic phase or thymoglobulin during the anhepatic phase and the first week after LT. No patient was administered everolimus monotherapy until the first month after LT, whereas 4 patients were not given CNIs or mTOR inhibitors (2 in the IH group and 2 in the no-IH group) because of acute severe preoperative and postoperative renal impairment. Steroids were administered and tapered in all patients at different posologies after LT, but most patients received 1 g of methylprednisolone (500 mg in the case of any associated induction therapy) before the reperfusion phase and 20 mg on postoperative day 1; this was followed by the tapering of prednisone (4 mg/week), and after 5 weeks, the drug was stopped.

The amount of missing data for each of the considered variables was less than 10%; specifically, data were missing for the Model for End-Stage Liver Disease (MELD) score (22 patients), the preoperative ascites status (25 patients), the operation length (23 patients), the total ischemia time (31 patients), the main immunosuppression (4 patients), and the albumin level (3 patients).

Statistical Analysis

Continuous data are reported as means and standard deviations and were compared with the 2-sided Student t test. Comparisons between groups for categorical variables were performed with the χ2 test with Yates' correction or Fisher's exact test when it was appropriate. The optimal cutoff for continuous variables was obtained from receiver operating characteristic curves.19

To identify variables that were independent predictors of outcomes, a logistic regression analysis with backward stepwise selection was constructed with those variables with a significance level of P < 0.05.

Patient survival was evaluated with the Kaplan-Meier method and was compared with the log-rank test. Statistical significance was set at P < 0.05. Statistical analysis was performed with SPSS 15.0.

RESULTS

During the entire follow-up period, we observed 121 IH cases (32.4%). The mean time to IH occurrence was 18.4 ± 17.3 months (range = 1.01-107 months; Fig. 1). Patient survival was significantly better in the IH group during the entire follow-up period (P < 0.001).

Figure 1.

IH-free survival versus the time after LT.

Figure 2.

IH-free survival of patients who were administered a CNI or an mTOR inhibitor.

One hundred nine patients (90.1%) were treated with surgical repairs, whereas 12 patients (9.9%) were not treated because of concomitant morbidities [hepatitis C virus (HCV) recurrence (10), respiratory distress (1), and IH with a marked loss of domain (1)]. After the IH surgical repairs, 2 of the 109 patients (1.8%) experienced a mesh infection, whereas IHs recurred in 4 patients (3.7%); all were given mTOR inhibitor therapy after LT and during the surgical repair period.

Table 1 lists pertinent pre-LT characteristics for patients who developed IHs and for patients who did not. A univariate analysis showed the following to be risk factors for post-LT IH development: male sex (P = 0.03), a BMI ≥ 29 kg/m2 (P = 0.005), LT after 2004 (P = 0.02), a MELD score ≥ 22 (P = 0.01), and hepatitis B virus (HBV) infection (P = 0.01).

Table 1. Preoperative Characteristics of Recipients With or Without IHs After LT
CharacteristicIH Group (n = 121)No-IH Group (n = 252)P Value
  • The bold numbers highlight the p values with a statistical significance.

  • *

    Data were available for 116 patients in the IH group and for 235 patients in the no-IH group.

  • Data were available for 114 patients in the IH group and for 234 patients in the no-IH group.

Sex: male [n (%)]99 (81.8)178 (70.6)0.03
Age ≥ 60 years [n (%)]29 (24.0)84 (33.3)0.09
BMI ≥ 29 kg/m2 [n (%)]23 (19.0)21 (8.3)0.005
LT era: 2000-2004/2005-2009 [n (%)]39 (32.2)/82 (67.8)114 (45.2)/138 (54.8)0.02
Liver disease [n (%)]  0.88
 Viral85 (70.2)169 (67.1)
 Alcohol13 (10.7)25 (9.9)
 Cholestatic5 (4.1)12 (4.8)
 Fulminant4 (3.3)7 (2.8)
 Other14 (11.6)39 (15.5)
Hepatocellular carcinoma [n (%)]51 (42.1)121 (48.0)0.34
MELD score ≥ 22 [n (%)]*46 (39.7)60 (25.5)0.01
HCV-positive [n (%)]57 (47.1)127 (50.4)0.6
HBV-positive [n (%)]36 (29.8)45 (17.9)0.01
Human immunodeficiency virus–positive [n (%)]7 (5.8)9 (3.6)0.48
Diabetes [n (%)]27 (22.3)48 (19.0)0.55
Ascites [n (%)]35 (30.7)73 (31.2)>0.99
Albumin (g/dL)3.35 ± 0.73.39 ± 0.70.63
Previous liver surgery [n (%)]4 (3.3)15 (6.0)0.40

No perioperative factors were found to be related to IH development (Table 2). None of the 4 patients with a delayed closure experienced an IH. However, patients who received a transplant from a deceased donor had a higher incidence of IHs in comparison with those who received a transplant from a living donor (34.9% versus 6.2% respectively, P = 0.002). The use of induction therapy was unrelated to IH development (Table 3). However, among those who did receive induction, basiliximab was associated with a higher rate of IHs (43.2%) in comparison with daclizumab (32.3%) or thymoglobulin (12.5%, P = 0.02). The use of basiliximab was also more closely associated with the use of mTOR inhibitors as monotherapy or in combination with CNIs (58.9%) in comparison with daclizumab (16.6%) or thymoglobulin (12.5%, P < 0.001).

Table 2. Postoperative and Donor Characteristics of the IH and No-IH Groups
CharacteristicIH Group (n = 121)No-IH Group (n = 252)P Value
  1. The bold numbers highlight the p values with a statistical significance.

Incision: J-shaped/Mercedes [n (%)]30 (24.8)/91 (75.2)52 (20.6)/200 (79.4)0.44
Operation length (minutes)498 ± 101501 ± 1130.84
Total ischemia (minutes)399 ± 96371 ± 1390.059
Postoperative ascites [n (%)]8 (6.6)12 (4.8)0.62
HCV recurrence [n (%)]32 (26.4)91 (36.1)0.08
Reoperation [n (%)]36 (29.8)91 (36.1)0.27
Re-LT [n (%)]6 (5.0)22 (8.7)0.28
Hospital stay (days)17.4 ± 1019.5 ± 160.22
Follow-up (months)51.9 ± 28.745.7 ± 34.30.08
Deceased/living donor [n (%)]119 (98.3)/2 (1.7)222 (88.1)/30 (11.9)0.002
Donor BMI (kg/m2)25.5 ± 3.225.3 ± 3.70.65
Donor age (years)57.5 ± 16.755.5 ± 19.40.32
Table 3. Immunosuppression Therapy During the First Month After LT
ImmunosuppressionIH Group (n = 121)No-IH Group (n = 252)P Value
  • The bold numbers highlight the p values with a statistical significance.

  • *

    P = 0.02 for the 3 types of induction.

  • P = 0.004 for the 3 immunosuppression regimens.

  • Data were available for 119 patients in the IH group and for 250 patients in the no-IH group.

  • §

    P = 0.01 for the 3 immunosuppression monotherapies.

  • Data were available for 72 patients in the IH group and for 186 patients in the no-IH group.

Induction [n (%)]*54 (44.6)96 (38.1)0.28
 Basiliximab41 (75.9)54 (56.2)0.03
 Daclizumab10 (18.5)21 (21.9)0.08
 Thymoglobulin3 (5.6)21 (21.9)0.02
Main immunosuppression [n (%)]  
 CNI66 (55.5)181 (72.4)0.002
 CNI and mTOR inhibitor47 (39.5)64 (25.6)0.009
 mTOR inhibitor6 (5.0)5 (2.0)0.20
Main immunosuppression [n (%)]  0.002
 mTOR inhibitor53 (44.5)69 (27.6)
 CNI66 (55.5)181 (72.4)
Immunosuppresion monotherapy [n (%)]§   
 Cyclosporine33 (45.8)62 (33.3)0.08
 Tacrolimus33 (45.8)119 (64.0)0.01
 Sirolimus6 (8.3)5 (2.7)0.08

IHs were found more frequently in patients treated with mTOR inhibitor monotherapy (54.5%) versus patients treated with CNI monotherapy (26.7%) or with CNIs and mTOR inhibitors (42.3%, P = 0.004). Moreover, patients who used any mTOR inhibitors (122 in all, either as monotherapy or in combination with a CNI) had an increased incidence of IHs (43.4%) in comparison with patients who received only a CNI (247 patients in all; incidence = 26.7%) with a higher statistical significance (P = 0.002). This difference is confirmed by the cumulative IH-free survival between the two groups, as showed in Fig. 2. Patients who were administered sirolimus as their primary immunosuppression had a higher incidence of IHs (54.5%) than patients who were administered cyclosporine or tacrolimus (34.7% and 21.7%, respectively, P = 0.01).

Patients who always used only CNIs after LT had an IH-free survival rate that was statistically lower than the rate of patients who used mTOR inhibitors (alone or in association with CNIs) within or after the first month post-LT (P < 0.001; Fig. 3).

Figure 3.

IH-free survival of patients who were administered a CNI alone or any mTOR inhibitor within the first month after LT or after the first month.

Forty-six patients (12.3%) were switched from CNIs to mTOR inhibitors after the first month post-LT; 43 were switched to sirolimus, and 3 were switched to everolimus. The mean time to switching was 397 ± 589 days (range = 30-2364 days). When the switch was performed between months 1 and 4, the incidence of IHs was 56% (9/16), whereas when the switch was performed after month 4, the incidence was 10% (3/30, P = 0.002).

A multivariate analysis with a logistic regression model revealed male sex [P = 0.03, odds ratio (OR) = 2.15], a pretransplant MELD score ≥ 22 (P = 0.04, OR = 2.3), and the use of mTOR inhibitors (P = 0.001, OR = 2.5) to be independent risk factors for IHs after LT (Table 4).

Table 4. Multivariate Logistic Regression of Preoperative and Postoperative Risk Factors
Risk FactorP ValueOR95% Confidence Interval
Sex: male0.032.151.1-4.2
MELD score ≥ 220.042.31.3-4
Immunosuppression: mTOR inhibitor0.0012.51.5-4.2

DISCUSSION

IHs after LT are complications with a reportedly growing incidence in recent years.1-7, 15 This growth may be caused by greater attention to this kind of complication, the adoption of the MELD score as a priority criterion for LT, and the recent introduction of new immunosuppressive drugs such as mTOR inhibitors.

The main risk factors for IH development after LT are not different from those for general abdominal surgery. However, candidates for LT often have additional risk factors because of their chronic liver disease and associated mixed protein-energy malnutrition due to hypermetabolism and malabsorption, diabetes, cachexia, and muscle mass loss.4

The overall incidence of IHs in our study was high (32.4%) and reflected the high percentage of patients with perioperative risk factors for IHs. The lesser incidence reported in the literature could be biased by an underestimation of IHs in other studies, by the hypothesis that the incidence of IHs is increasing, or by both.

In our experience, the overall patient survival was longer for the IH population because IH development is most often a late complication that occurs in patients still alive after LT.

In our experience, the statistically significant risk factors related to IH development in a univariate analysis were male sex, a patient BMI ≥ 29 kg/m2, LT performed in or after 2005, a MELD score ≥ 22, HBV positivity, a deceased donor graft, and mTOR inhibitor immunosuppressive therapy. In contrast to previous studies, age, diabetes, preoperative and postoperative ascites, previous and subsequent LT abdominal surgery, and the type of surgical incision were not found to be significant risk factors.

Some of the risk factors in our series were temporally related. The use of mTOR inhibitors and the use of basiliximab-based induction both started after 2005. In other articles,2 the J-shaped incision has been reported to be a protective factor for IH development, but in the present series, we did not find any correlation. This difference could be explained by the fact that when we introduced the J-shaped incision (after 2007), we also increased the use of mTOR inhibitors as immunosuppressants.

The multivariate analysis revealed that male sex, a MELD score ≥ 22, and immunosuppression based on mTOR inhibitors in the first month after LT were independent risk factors for IH development after LT. Among these factors, immunosuppressive treatment is the only feature modifiable by clinicians.

Several series3, 17, 20 have reported immunosuppressive therapy as a risk factor for IHs in univariate analyses, but only 1 small study has described sirolimus as an independent risk factor in a multivariate analysis.4

The use of mTOR inhibitors as immunosuppressants after LT is growing because of their many advantages in comparison with CNIs, especially for renal function preservation.21 However, these advantages are associated with other side effects such as anemia, leukopenia, dyslipidemia, pitting edema, buccal ulcers, and IHs.17, 22 In our analysis, the rate of IHs in patients treated with mTOR inhibitors during the first month after LT was 43.4%, whereas in patients treated with CNIs, the rate was 26.7% (P = 0.002). The risk of IHs was also high (56%) when mTOR inhibitors were introduced 1 to 4 months after LT.

Our group recently described how the early use of everolimus monotherapy preserves renal function after LT despite cyclosporine with similar freedom from efficacy failure.21 In that study, the incidence of IHs was higher in the everolimus group, although this was not statistically significant. After the enrollment of more patients to reach the optimal sample size for evaluating the secondary targets of the study and with longer follow-up, the incidence of IHs was 50.8% (32 of 63 patients) within the everolimus group and 27.0% (10 of 37 patients, P = 0.03) within the cyclosporine group.

A recent experimental study of Wistar rats demonstrated considerable and protracted impairment of wound healing induced by the administration of everolimus.20 In that study, the effect on the development of wound strength persisted for at least 4 weeks after the operation, and this showed that it may have clinical consequences and cause surgical morbidity. mTOR inhibitors reduce the number of inflammatory cells and lead to the decreased expression of vascular endothelial growth factor and nitric oxide, which are mediators of angiogenesis and immune function in skin wounds. It seems evident that the correlation between mTOR inhibitors and IHs is strictly linked to the early use of mTOR inhibitors after LT, that is, when the immune function and angiogenesis associated with wound healing are required.

In a multicenter randomized study, De Simone et al.23 evaluated whether the use of everolimus with CNI reduction or discontinuation would improve renal function in maintenance LT recipients experiencing CNI-related renal impairment. In their series, they introduced everolimus therapy after a mean time of 3 years, and they did not find any IH episodes. These data support the evidence showing that mTOR inhibitor immunosuppressive therapy is associated with IH development in the early postoperative period. In the present study, except for the use of mTOR inhibitors, no perioperative factors were found to be related to IH development (Table 2). However, patients who received a transplant from a deceased donor had a higher incidence of IHs (34.9%) than patients who received a graft from a living donor (6.2%, P = 0.002). This finding is in contrast to the series of Vardanian et al.,7 who found an increase in IHs in recipients of living donor grafts due to a high incidence of reoperation for bile leaks.

Andreoni et al.24 demonstrated that in 12 LT patients who underwent laparoscopic IH repair, not one experienced IH recurrence or serious infections during a mean follow-up of 212.3 ± 125.7 days. Kurmann et al.25 reported 44 patients who underwent IH repair after LT; the recurrence rate was 34% after open repairs (31 patients) and 15% after laparoscopic repairs (13 patients) during mean follow-up periods of 56 and 29 months, respectively. These studies show that laparoscopic IH repairs in the post-LT population result in trends of lower recurrence rates and fewer complications; that is, these patients respond similarly to patients who have not undergone LT.

The incidence of IHs after LT in the literature is 1.7% to 34.3%; surgical repairs are effective, but recurrence often occurs. In our series, 4 of 109 patients (3.7%) experienced IH recurrence, and all had been administered an mTOR inhibitor as an immunosuppressant after LT and after the surgical repair of their IHs. On the basis of this preliminary experience as well as an in vivo study26 demonstrating that the administration of rapamycin dose-dependently impairs vascularization and the incorporation of implanted surgical meshes, we now switch the immunosuppressive therapy from mTOR inhibitors to CNIs 1 month before and 3 months after the surgery to augment wound healing and strength.

In conclusion, the early use of mTOR inhibitors as immunosuppressive therapy is the only modifiable independent risk factor associated with IH development after LT in our series. The beneficial effects of the early use of mTOR inhibitors versus CNIs in terms of renal function preservation, diabetes mellitus, neurological complications, and arterial blood pressure should be balanced against the high incidence of postoperative IHs. We suggest that CNI drugs be used in the early post-LT period if they are not contraindicated and that the use of mTOR inhibitors be delayed until after the fourth month post-LT. At that time, complete wound healing and wound strength development should be achieved, and the potentially negative side effects of CNI drugs can still be reversed.

Ancillary