The American Journal of Surgery (2013) 205, 188-193

Clinical Science Improved outcomes of incarcerated femoral : a multivariate analysis of predictive factors of bowel ischemia and potential impact on postoperative complications

Cristina Alhambra-Rodriguez de Guzma´n, M.D.a, Joaquı´n Picazo-Yeste, M.D., Ph.D.a,*, Jose Marı´a Tenı´as-Burillo, M.D., Ph.D.b, Carlos Moreno-Sanz, M.D., Ph.D.a aDepartment of General and Digestive Surgery, Hospital General La Mancha, Centro Avenida de la Constitucio´n, 3.13600 Alca´zar de San Juan, Ciudad Real, Spain; bDepartment of Epidemiology and Preventive Medicine, Hospital General La Mancha, Ciuda Real, Spain

KEYWORDS: Abstract Femoral hernia; BACKGROUND: Although much of the literature focuses on risk factors for intestinal resection in Ischemia; groin , little is known specifically for the femoral type. This study identifies clinical and Intestinal obstruction; analytic parameters associated with in patients with an incarcerated femoral Risk factors hernia. METHODS: Eighty-six patients with an incarcerated femoral hernia were included in an analytic, longitudinal, observational, retrospective cohort study. Clinical presentation, the duration of symptoms, analytic and radiologic studies, complications, and mortality rates were analyzed. RESULTS: Eight (9.3%) patients underwent intestinal resection. Factors related to intestinal ische- mia were oral anticoagulants intake (odds ratio 5 9.6) and a duration of symptoms longer than 3 days (odds ratio 5 2.1). There was no relationship between leukocytosis (P 5 .02) or radiographic signs of intestinal obstruction (P 5 .28) and . CONCLUSIONS: Patients with a duration of symptoms longer than 3 days and, interestingly, those having oral anticoagulant therapy appeared to be at a higher risk for developing intestinal ischemia. A remarkable reduction in morbimortality can be achieved through an earlier referral to the hospital, quick preoperative workup, and urgent operation. Ó 2013 Elsevier Inc. All rights reserved.

Hernia surgery is one of the most frequently performed the prevalence, resulting disability, recurrence, and socio- operations in the Western world. Despite this universally economic implications of hernias. The incidence of femoral acknowledged fact, scant attention has been paid to hernia is reported to be 2% to 8% of all groin hernias in adult patients. This type of hernia, which is very rare in children, is most commonly observed between the fourth and seventh * Corresponding author. Tel.: 134-629-149031; fax: 134-926-580669. E-mail address: [email protected] decades of life and is 4 to 5 times more common in women. Manuscript received December 18, 2011; revised manuscript February In addition, for reasons yet unknown, right-sided presenta- 1 2, 2012 tion is more common than left-sided presentation.

0002-9610/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2012.03.011 C. Alhambra-Rodriguez de Guzma´n et al. Outcomes of incarcerated femoral hernia 189

Although the incidence of femoral hernias is low, they Methods are of high clinical importance because they often present with strangulation. The incidence of strangulation reaches Study design and patients 38% in some reports, leading to high morbidity and 1–4 mortality rates. The importance of an early diagnosis We conducted an analytic, longitudinal, observational, of incarceration is paramount because a delay increases retrospective cohort study with 86 adult patients undergo- the risk of ischemia and necrosis of the incarcerated tissue, ing emergency treatment for incarcerated femoral hernias at thus augmenting the need for intestinal resection with its 5–7 La Mancha Centro General Hospital, Ciudad Real, Spain, consequently higher morbidity and mortality rates. between 1995 and 2009. All patients underwent a routine The only satisfactory treatment for incarcerated femoral preoperative workup including blood analysis, an electro- hernia is an urgent operation. Over the past several years, cardiogram, chest and abdominal radiographs, and anes- many published studies have examined the efficacy of thesiology consulting. Because incarcerated femoral hernia different techniques, which have undergone a is mainly determined by a clinical diagnosis, other radio- remarkable evolution with the development of prosthetic logic studies (ie, ultrasonography, a computed tomography materials. Currently, the Lichtenstein femoral plug repair is scan, and magnetic resonance imaging) were deemed the most widely used technique. It is reliable, safe, and unnecessary. Attempts for the manual reduction of femoral easily reproducible and has become the gold standard for hernias are forbidden in our hospital. A hematologist was both elective and emergency hernia repair. Although pre- consulted for patients on anticoagulant therapy, proceeding peritoneal hernia repair is as reliable and safe as the 8 for immediate reversal (ie, vitamin K, plasma infusion, and/ anterior approach, it requires specific training, and some or activated prothrombin complex). All operations were surgeons find it more difficult to perform. With regard to performed as soon as possible once the patient was complications, available studies have found no differences prepared for surgery. Depending on the surgical team between anterior and preperitoneal techniques with respect 9 preference, the anterior or preperitoneal approach was used. to wound infection, seroma formation, or hematomas. The following patient information was recorded: sex, age, Although prostheses are widely used in elective groin personal health history, a history of cardiopulmonary dis- hernia repair, the use of synthetic materials is considered eases, a history of intake of oral anticoagulants, the presence controversial in emergency operations because of the 10 of diabetes mellitus, clinical presentation (ie, tumor, pain, inherent risk of sepsis. Moreover, some surgeons consider and vomiting, and fever), time until the operation, the presentation of strangulated hernia to be an absolute laboratory results (ie, white cell blood count), radiologic contraindication for the use of prosthetic meshes because studies (ie, signs of ), postoperative med- of a higher risk of infection.11 However, others such as 12 ical complications (ie, heart failure and pneumonia), wound Pans et al consider the risk of sepsis to be overestimated. infection and the presence of fistula, the surgical technique They insist that with careful attention to local antisepsis and used (anterior vs preperitoneal approach), the length of the use of systemic antibiotic therapy there should be no hospital stay, and mortality. additional risk of local infection when inserting a prosthesis during an emergency herniorraphy procedure.13–15 Late admission to the hospital is perhaps the most Analysis important factor in the evolution of an incarcerated femoral hernia because the prolonged duration of symptoms leads to All continuous data were expressed with central ten- a higher risk of intestinal ischemia and necrosis of the dency measures as means and dispersion measures as incarcerated tissue. Some reports show that bowel obstruc- standard deviations. Qualitative data are expressed as tion significantly increases the mortality rates associated absolute and relative frequencies (percentages). Statistical with hernias. Several studies have analyzed the risk factors analyses were performed using either the t test, chi-square for intestinal resection in groin hernias but not specifically test, or Fisher exact test. Differences were considered to be for the femoral type. Our special concern regarding abdom- significant at P , .05. inal wall pathology led us to develop and maintain a reliable To identify predictive factors for bowel resection caused database during the last 16 years. The awareness of a by ischemia, a predictive model was created with the aid of surprisingly low rate of bowel resection made us investigate multiple logistic regression analysis. The dependent variable this issue. As far as we know, this article reports on one of was the presence or absence of bowel ischemia. First, we the largest single-institution series dealing with incarcerated performed a simple logistic regression to determine the femoral hernias. To anticipate bowel necrosis and its relationship between each potential predictor variable with catastrophic consequences, we investigated the relationship bowel resection caused by ischemia. Those variables with a P between several preoperative clinical, radiologic, and lab- value %.25 were selected for the multiple logistic regression oratory findings and the occurrence of irreversible ischemia analysis. The Hosmer-Lemeshow test was performed for in- in incarcerated femoral hernias. This article also focuses on ternal calibration of the model. The capability of discrimina- potential strategies aiming to achieve the best control of tion of the model was calculated by the area under the intestinal viability upfront bowel necrosis. receiver operating characteristic curve. All analyses were 190 The American Journal of Surgery, Vol 205, No 2, February 2013 performed with SPSS software (version 18.0; SPSS Inc, Chi- 10 (11.6%) patients were taking oral anticoagulants for the cago, IL). treatment of atrial fibrillation (7 cases), previous deep venous thrombosis (2 cases), and dilated cardiomyopathy (1 case). The most common complaint was pain in 77 cases Results (89.5%) followed by the presence of a femoral bulge in 64 cases (74.4%), nausea and vomiting in 44 cases (51.2%), During the study period, a total of 86 adults with and fever in 2 patients (2.3%). Leukocytosis above 11,000 incarcerated femoral hernias were included in the analysis, was present in 31 cases (36%), and radiographic signs of 20 men (23.3%) and 66 women (76.7%). The average age of bowel obstruction were observed in 37 patients (43%). the patients was 68.4 6 18.3 years (range 27.6 to 93.9 In all 86 cases, the repair was buttressed with a years). Twenty-nine (33.7%) patients suffered from cardi- polypropylene mesh. Forty-six cases (53.5%) were oper- opulmonary disease; 4 (4.7%) had diabetes mellitus; and ated using the anterior approach (ie, placing a plug into the

Table 1 Characteristics of patients with and without bowel resection Total No bowel resection Bowel resection Variables (N 5 86) (N 5 78) (N 5 8) P value Sex (%) Male 20 (23.3) 17 (21.8) 3 (37.5) .38† Female 66 (76.7) 61 (78.2) 5 (62.5) Age (y) 68.4 6 18.3 67.7 6 18.8 75.7 6 10.4 .24‡ Duration of symptoms 2.2 6 1.3 2.1 6 1.2 3.3 6 1.5 .017‡ Concomitant disease (%) Without 56 (65.1) 53 (67.9) 3 (37.5) .121† With 30 (34.9) 25 (32.1) 5 (62.5) Cardiopulmonary diseases (%) Without 57 (62.3) 52 (66.7) 5 (62.5) 1† With 29 (33.7) 26 (33.3) 3 (37.5) Anticoagulants oral (%) Without 76 (88.4) 71 (91.1) 5 (62.5) .047† With 10 (11.6) 7 (8.9) 3 (37.5) Diabetes mellitus (%) Without 82 (95.3) 74 (94.9) 8 (100) 1† With 4 (4.7) 4 (5.1) 0 (0) Femoral tumor (%) Without 22 (25.6) 21 (26.9) 1 (12.5) .674† With 64 (74.4) 57 (73.1) 7 (87.5) Pain (%) Without 9 (10.5) 7 (8.9) 2 (25) .196† With 77 (89.5) 71 (91.1) 6 (75) Nausea and vomiting (%) Without 42 (48.8) 40 (51.3) 2 (25) .266† With 44 (51.2) 38 (48.7) 6 (75) Fever (%) Without 84 (97.7) 78 (100) 6 (75) .008† With 2 (2.3) 0 (0) 2 (25) Signs of bowel obstruction at abdomen radiography (%) Without 49 (56.9) 46 (58.9) 3 (37.5) .282† With 37 (43.1) 32 (41.1) 5 (62.5) Leukocytosis* (%) Without 55 (63.9) 53 (67.9) 2 (25) .023† With 31 (36.1) 25 (32.1) 6 (75) Length of stay (d) 5.6 6 3.6 4.8 6 2.6 13.3 6 3.1 ,.001‡ Surgical technique (%) Anterior 46 (53.5) 44 (56.4) 2 (25) .138† Preperitoneal 40 (46.5) 34 (43.6) 6 (75) *The presence of leukocytosis: values above 11,000; absence: values below 11,000. †The Fisher exact test. ‡t test for independent samples. C. Alhambra-Rodriguez de Guzma´n et al. Outcomes of incarcerated femoral hernia 191

Table 2 The duration of symptoms of patients at the moment of hospital admission Duration of symptoms Total (N 5 86) No bowel resection (N 5 78) (%) Bowel resection (N 5 8) (%) P value* %3 days 73 69 (94.5) 4 (5.5) .016 .3 days 13 9 (69.2) 4 (30.8) *The Fisher exact test was used unless otherwise indicated. ), whereas 40 patients (46.5%) underwent a wound infection. There were 9 complications (10.5%); 2 modified open preperitoneal hernioplasty as previously patients developed pneumonia, and 7 presented with wound described by our group.8 Briefly, the procedure consists infection (62.5% vs 5.1%, P , .001). In 6 (86%) of 7 com- of an open preperitoneal approach (ie, the Nyhus approach) plicated wounds, the infections were located in the superfi- to insert a patch of polypropylene individually fashioned in cial site, and all healed with conservative treatment. The an approximately 12 ! 10 cm ‘‘M’’-shaped piece to con- only patient with deep site infection (she had received a form to each patient’s anatomy and placed without fixation Lichtenstein plug repair with bowel resection through the and covering all potential hernial orifices. The aim here is ) developed a chronic cutaneous fistula. She to create 3 prolongations (flaps) in the mesh for proper was scheduled for removal of the plug 6 months after the self-anchorage in the preperitoneal space. Because the sper- first operation. The inguinofemoral area was reinforced matic cord is ‘‘parietalized’’ under the central flap, a slit in with a new preperitoneal mesh with an uneventful recovery. the prosthesis is unnecessary, thus avoiding its weakening. Fifty percent of patients who underwent bowel resection Bowel resection because of ischemia was required in 8 developed wound infection, whereas this complication cases (9.3%). Demographic and clinical characteristics of arose in only 3.8% of cases with no bowel resection. patients with or without bowel resection are given in Table Two patients died, accounting for a global mortality rate of 1. Patients who required bowel resection were older (75.7 2.3%. Both patients belonged to the bowel resection group 6 10.4 vs 67.7 6 18.8 years, P 5 .24) and had a longer (P 5 .008). Indeed, mortality and general complications hospital stay (13.3 vs 4.8 days, P , .001). Seventy-three were both significantly associated with bowel resection. (85%) of the 86 patients were admitted to the hospital within the first 3 days after the onset of symptoms. The du- ration of symptoms was longer in patients who required Comments bowel resection (3.3 6 1.5 vs 2.1 6 1.2 days, P 5 .02) with a significant cutoff point at 3 days (Table 2). Our results indicate 2 independent risk factors for bowel A multivariable logistic regression model identified resection in patients with incarcerated femoral hernia. First, 2 variables that were independent risk factors for patients taking oral anticoagulants have a much higher risk bowel resection because of ischemia (Table 3): oral antico- of having bowel ischemia. Although it is not easy to find a agulants intake (odds ratio 5 9.6 [95% confidence interval, physiopathological reason for this association, we suggest 1.5 to 60.8], P 5 .016) and a duration of symptoms .3 days that patients receiving these medications may have some (odds ratio 5 2.1 [95% confidence interval, 1.1 to 3.7], underlying disease that predisposes them to bowel ische- P 5 .015). Patients whose blood analysis showed over mia. If this hypothesis is considered plausible, it raises 11,000 leukocytes presented almost a 3-fold risk of bowel serious doubts about the opinion that anticoagulant thera- resection because of ischemia compared with those with pies may protect against ischemic complications. More- lower values although the difference was not statistically over, reversion of anticoagulant effects before surgery significant (P 5 .176). Two patients had fever; both be- should be undertaken as quickly as possible. Considering longed to the group requiring bowel resection because of that our favorable results may be related to this way of ischemia (P 5 .008). proceeding, we should emphasize the role of operating as Severe postoperative complications are listed in Table 4. soon as possible, even more in the presence of anticoag- Bowel resection was associated with a higher likelihood of ulant therapy. In our experience, this management has led complications, including both medical condition and us to frequently find cyanotic but viable intestinal loops within the hernia sac. Second, we found a relationship between the prolonged duration of symptoms (.3 days) up Table 3 Risk factors for bowel resection because of ischemia to the time of surgery and a higher risk of bowel resection, Risk factors P value OR (95% CI)* which is consistent with the findings reported by Kurt et al.16 Being female and/or being over the age of 65 years . Duration of symptoms 3 days .015 2.1 (1.15–3.67) have been reported by several authors to be risk factors for Oral anticoagulants .016 9.6 (1.53–60.76) incarcerated femoral hernia,16,17 but our study found no CI 5 confidence interval; OR 5 odds ratio. evidence of this. *Odds ratio derived from logistic regression; values in parentheses Ge et al18 retrospectively analyzed 182 patients who had are 95% confidence intervals. undergone emergency surgery for incarcerated groin hernia, 192 The American Journal of Surgery, Vol 205, No 2, February 2013

Table 4 Postoperative complications of patients with and without bowel resection Variable Total (N 5 86) No bowel resection (N 5 78) Bowel resection (N 5 8) P value* General complications (%) Without 77 (89.5) 74 (94.9) 3 (37.5) ,.001 With 9 (10.5) 4 (5.1) 5 (62.5) Pneumonia (%) Without 84 (97.7) 77 (98.7) 7 (87.5) .18 With 2 (2.3) 1 (1.3) 1 (12.5) Wound infection (%)) Without 79 (91.9) 75 (96.2) 4 (50) .001 With 7 (8.1) 3 (3.8) 4 (50) Mortality (%) Without 84 (97.7) 78 (100) 6 (75) .008 With 2 (2.3) 0 (0) 2 (25) *The Fisher exact test was used unless indicated otherwise.

28 of them with the femoral type. They observed a rate of results published by Kulah et al19 (17.6%) although the dif- bowel resection for femoral hernias of approximately 46%. ferences are not statistically significant (P 5 .91). Suppiah This resection rate, the highest we have found in the liter- et al17 found morbidity and mortality rates of 21.4% and a ature, is consistent with several other reports3,6,16–19 in 3.6%, respectively, after emergency or elective femoral her- which the rate ranges from 20.8% to 38.5%. In contrast, nia surgery, both of which exceed the rates observed in our our study showed a significantly lower resection rate of study. only 9.3% (Table 5). Alterations in skin color and the presence of leukocy- This marked difference must be highlighted because it tosis are usually considered to be signs of strangulated raises doubts about previously asserted associations. In our femoral hernia. However, we were not able to show an series, 73 (85%) patients experienced a duration of symp- association between leukocytosis and bowel ischemia in toms of 3 days or less before seeking specialized medical our study. Although there was a clearly higher rate of bowel help, perhaps because of quick referrals on the part of their resection in patients with leukocyte values above 11,000, general practitioners. This suggests that minimizing the this association was not found to be statistically signifi- preoperative period in patients with incarcerated femoral cant.12 Our data likewise showed no association between hernias may be a way to lower bowel resection rates. the suspicion of intestinal obstruction as seen in abdominal The development of bowel necrosis and subsequent radiography and bowel ischemia. This is consistent with bowel resection has been associated with longer hospital observations made by Sarr et al,20 who found no preopera- stays and worse outcomes for patients with incarcerated tive clinical parameters including the presence of continu- femoral hernia; indeed, some studies have shown that ous abdominal pain, fever, peritoneal signs, leukocytosis, bowel resection has a direct effect on morbidity and or acidosis, or a combination thereof, proved to be sensi- mortality.5 This supposition is confirmed in our study be- tive, specific, and predictive for strangulation. cause patients who underwent bowel resection had longer The principal limitation of our study is its retrospective postoperative hospital stays, mainly because of surgical nature, and further prospective studies are needed for a better wound infections. Moreover, we observed a mortality rate assessment of the proposed risk factors. However, some of of 25% after bowel resection, which is higher than the the key issues discussed in this article may be integrated into

Table 5 Literature reports of bowel resection and mortality in femoral hernia and comparison with this study Patients with emergency Reference Study period femoral hernia Resection rate (%) P value* Mortality rate (%) P value† A´lvarez et al3 1992–2001 77 16 (20.8) .065 3 (3.9) .9 Derici et al6 1998–2006 19 4 (21) .3 1 (5.3) .95 Kurt et al16 1997–2001 13 5 (38.5) .014 dd Suppiah et al17 2000–2004 28 8 (28.6) .025 1 (3.6) .75 Ge et al18 1999–2009 28 13 (46.4) ,.001 dd Kulah et al19 1996–2001 42 17 (41) .0001 3 (7) .4 This study 1995–2009 86 8 (9.3) d 2 (2.3) d *P value: result of the comparison of the resection rate between this study and the rest. †P value: result of the comparison of the mortality rate between this study and the rest. C. Alhambra-Rodriguez de Guzma´n et al. Outcomes of incarcerated femoral hernia 193 a strategy to achieve the highest advance before the devel- 4. Gallegos NC, Dawson J, Jarvis M, et al. Risk of strangulation in groin opment of irreversible bowel necrosis. In that sense, we hernias. Br J Surg 1991;78:1171–3. should underline the following: improving the level of 5. Alimoglu O, Kaya B, Okan I, et al. Fermoral hernia: a review of 83 cases. Hernia 2006;10:70–3. suspicion of incarcerated femoral hernia at primary care 6. Derici H, Unalp HR, Bozdag AD, et al. Factors affecting morbidity providers (maybe through specific medical instruction); and mortality in incarcerated hernias. Henia 2007; stating clearly that any attempt of manual hernia reduction 11:341–6. should be avoided at all levels of health care (ie, general 7. Chamary VL. Femoral hernia: intestinal obstruction is an unrecog- practitioners, emergency room staff members, residents, and nized source of morbidity and mortality. Br J Surg 1993;80:230–2. 8. Picazo JS, Seoane JB, Moreno C, et al. Description of M-shaped pre- surgeons) when a femoral hernia is suspected; and preparing peritoneal hernioplasty for inguinocrural hernias. Am J Surg 2003;185: the patients for surgery as soon as possible, especially if they 108–13. are receiving anticoagulant therapy or we think that the 9. Muldoon RL, Marchant K, Johnson DD, et al. Lichtenstein vs anterior duration of symptoms is about 3 days or more. preperitoneal prosthetic mesh placement in open inguinal repair: a pro- spective, randomized trial. Hernia 2004;8:98–103. 10. Campanelli G, Nicolosi FM, Pettinari D, et al. Prosthetic repair, intes- Conclusions tinal resection, and potentially contaminated areas: safe and feasible? Hernia 2004;8:190–2. Similar to previously published results, our study found 11. Nyhus LM, Pollak R, Bombeck CT, et al. The preperitoneal approach that physical signs, leukocytosis, and radiologic data of and prosthetic buttress repair for recurrent hernia. The evolution of a technique. Ann Surg 1988;208:733–7. intestinal obstruction are not useful for anticipating intes- 12. Pans A, Desaive C, Jacquet N. Use of preperitoneal prosthesis for tinal ischemia. Our main finding was that there is a strong strangulated groin hernia. Br J Surg 1997;84:310–2. association between the intake of oral anticoagulants and 13. Wysocki A, Kulawik J, Pozniczek M, et al. Is the Lichtenstein opera- the prolonged duration of symptoms in patients with tion of strangulated groin hernia a safe procedure? World J Surg 2006; incarcerated femoral hernia and the presence of bowel 30:2065–70. 14. Dahlstrand U, Wollert S, Nordin P, et al. Emergency femoral hernia re- necrosis. In these circumstances, a prompt operation is pair. a study based on a National Register. Ann Surg 2009;249:672–6. necessary. We also showed that the average rate of bowel 15. Karatepe O, Adas G, Battal M, et al. The comparison of preperitoneal resection among patients with incarcerated femoral hernia and Lichtenstein repair for incarcerated groin hernias: a randomized can be reduced from those in previously reported series if controlled trial. Int J Surg 2008;6:189–92. the aforementioned factors are properly controlled. 16. Kurt N, Oncel M, Ozkan Z, et al. Risk and outcome of bowel resection in patients with incarcerated groin hernias: retrospective study. World J Surg 2003;27:741–3. 17. Suppiah A, Gatt M, Barandiaran J, et al. Outcomes of emergency References and elective femoral hernia surgery in four district general hospitals: a 4-year study. Hernia 2007;11:509–12. 1. Hachisuka T. Femoral hernia repair. Surg Clin N Am 2003;83: 18. Ge BJ, Huang Q, Liu L, et al. Risk factors for bowel resection and outcome 1189–205. in patients with incarcerated groin hernias. Hernia 2010;14:259–64. 2. Oishi S, Page C, Schwesinger W. Complicated presentations of groin 19. Kulah B, Duzgun AP, Moran M, et al. Emergency hernia repairs in hernias. Am J Surg 1991;162:568–70. elderly patients. Am J Surg 2001;182:455–9. 3. A´ lvarez-Pe´rez JA, Baldonedo-Cernuda RF, Garcı´a-Bear I, et al. Pre- 20. Sarr MG, Bulkley GB, Zuidema G. Preoperative recognition of intes- sentation and clinical evolution of incarcerated groin hernias in adults. tinal strangulation obstruction: prospective evaluation of diagnostic Cir Esp 2005;77:40–5. capability. Am J Surg 1983;145:176–82.