Steatosis is usually an incidental finding and hepatomegaly is often the only finding on physical examination. Despite widespread clinical use of imaging methods, ultrasound, computed tomography, or magnetic resonance imagining can only to some extent detect the degree of steatosis. Saadeh et al demonstrated the limitation of radiological modalities in a study where only steatosis >25–30% could be reliably detected radiologically 22. In addition, none of these modalities was able to either distinguish the grading or the type of liver fat or to detect the individual pathologic features important to establish steatohepatitis such as necro-inflammatory changes, hepatocyte ballooning, and fibrosis 23.

There is consensus that the gold standard of diagnosis is = 2 liver biopsies, as a single biopsy can result in substantial misdiagnosis and staging inaccuracies 24. However, due to the risk of fatal bleeding after liver biopsy (0.4%), it is not routinely performed in patients without apparent complicated liver disease 25. A uniform quantitative grading for steatosis and steatohepatitis has been suggested, combining the identified key pathologic features 26. Besides quantitative grading, steatosis can be classified qualitatively into microvesicular and macrovesicular forms. Most frequent is the macrovesicular one, and is frequently associated with obesity, non-insulin-dependent diabetes, some dyslipidemias, and alcohol abuse. Microvesicular steatosis is usually related to more acute conditions such as acute viral infections, metabolic disorders, and various toxins but also to acute fatty liver of pregnancy. The histopathologic features of steatosis are evaluated in preoperative needle biopsies or operative wedge specimens that are frozen and/or deparaffinized. The staining methods currently used are hematoxylin and eosin with which the fatty changes are assessed by considering the non-stained regions. In addition, specific fat stains such as Oil Red O and Sudan III are used. However, there are several problems in clinical application of these staining methods. The conventional techniques potentially underestimate the extent of fatty infiltration as they fail to identify microvesicular forms of steatosis 27. Also the fat specific stains have pitfalls, for example, in Oil Red O-stained liver tissue, the quality and quantity of the staining are highly operator-dependent and false-positive results or overestimation of the severity are possible because of unspecific sinusoidal staining 28.

Patients with nonalcoholic fatty liver disease have an increased prevalence of non-insulin-dependent diabetes, but the actual role of diabetes in postoperative recovery is unclear as studies report contradictory results. Non-insulin-dependent diabetes was identified as independent and significant variable predicting major postoperative complications in a cohort of 209 patients undergoing liver resection 29. However, contrary to this study, a study including 525 diabetic and non-diabetic patients requiring hepatectomy, reported no difference in perioperative morbidity or mortality and no effect was observed in long-term prognosis 30. Although the impact of diabetes on postoperative complications remains unclear, an increased rate of wound infections was reported in an impressive cohort of over 20,000 patients 31. From the analysis of a large database, was concluded that a subset of diabetic patients having steatosis in the analysis of the liver specimen, poorly tolerate major liver resection 32. Thus, when contemplating major hepatectomy in patients with diabetes mellitus, surgeons should be aware that the presence of diabetes mellitus may lead to higher incidence of hepatic failure postoperatively.

Obesity is crucially linked with steatosis 14. In the past, obesity has been also linked to increased perioperative technical complications leading to prolonged postoperative recovery. However, Dindo et al prospectively investigated a cohort of 6336 patients undergoing elective major surgery and found no increase in postoperative morbidity and mortality between obese and non-obese, not even in morbidly obese patients 33. However, the authors did not discriminate patients who underwent exclusively liver surgery. Unfortunately, reports referred to perioperative outcome of obese patients requiring hepatectomy are scarce in the literature.

In contrast to general obesity, body fat accumulation (subcutaneous or intraabdominal) has been reported to be independently associated with postoperative morbidity in a prospective study of 139 patients who underwent gastric or colorectal cancer surgery 34.

Only a few studies have focused on liver steatosis as a risk factor for postoperative complications after liver resection (Table 1). However, there are some general problems in the reporting of these studies. The histopathology methods for diagnosis of steatosis are not frequently, if ever, mentioned. So, the reliability of diagnosis of steatosis remains uncertain, rendering the comparison of the results difficult. Uniform grading together with a more detailed description of the staining methods used and the number and sort of biopsies taken are important to compare different studies 15. In the absence of a uniform classification system accepted by most to better stratify morbidity by severity, scarce data is reported concerning the incidence of major complication rate after liver resection in steatotic patients. Finally, mortality is usually assessed as in-hospital mortality or 60 days mortality after operation, and there are no studies evaluating 5- or 10-year survival of steatotic patients after surgey.

In 1998, Behrns et al evaluated in a retrospective study of 135 patients, the safety of major resection in patients with hepatic steatosis 35. They reported an increased postoperative mortality, morbidity, and blood transfusion rates together with longer operative time in the presence of steatosis of ≥ 30%. Belghiti et al, in a large database including 478 liver resections, detected the presence of hepatic steatosis ≥ 30% in 37 (7.7%) patients, and they demonstrated that steatosis was an independent risk factor for postoperative complications but not for mortality 5. In the analysis of a cohort of 1803 liver resections, the histological analysis demonstrated that only 55% of the liver specimens had normal parenchyma and 18% had liver steatosis 6. In contrast, the presence of steatosis was not a risk factor for postoperative outcome after hepatectomy. Later, Kooby et al reviewed the above-mentioned cohort to perform a case-matched study involving 325 steatotic patients compared with 160 with normal liver who were matched according to age, co-morbidities, and the extent of resection 36. Interestingly, they showed that steatosis was a predictor of postoperative complications, but mainly due to minor complications. Major postoperative complications, eg, requiring major therapeutic intervention or mortality, were not affected by the presence of steatosis. To note, information regarding the presence of concomitant risk factors, such as nutritional parameters, kidney function and cholestasis, were not included in this analysis. Furthermore, the histologic analysis did not differentiate between the two types of liver steatosis or other histological features such as steatohepatitis or fibrosis. Therefore, we decided to investigate the influence of the quantitative and qualitative analysis of steatosis in postoperative outcome after major liver resection 12. In our study, each patient with liver steatosis was matched with a similar patient with normal liver who required major hepatectomy according to the following variables: age, gender, ASA (American Society of Anesthesiologists) score, diagnosis, extent of hepatectomy, and the need of bilio-enteric bypass. We found that steatosis was the most significant preoperative risk factor for postoperative complications and should be considered for the planning of a major liver resection. Interestingly, neither the type nor the grading of liver steatosis influenced postoperative outcome after major hepatectomy. A novel finding in this study was that the presence of preoperative jaundice in a steatotic liver was an independent risk factor for mortality after surgery. Consequently, efforts should be made either to postpone surgery until normalization of bilirubin or to minimize the extent of liver resection in those patients (e.g., economic hepatectomy combined with ablation therapies, two-stage hepatectomy, etc).

Although blood tests intended to evaluate preoperative hepatic function in cirrhotic patients requiring liver resection, these techniques frequently underestimate the underlying liver disease 38. In an attempt to solve this problem, many different methods have been used to evaluate hepatic functional reserve. Among all, the Child-Pugh clinical score is the most validated and used method for cirrhotic patients. Major hepatic resections are safely performed in patients scored as Child- Pugh A but this classification failed to be an adequate predictor of the extent of the liver resection that could be safely performed. In contrast, patients classified as Child- Pugh B and C have unacceptable morbidity and mortality risks and therefore are not suitable for major liver resections 39.

Several methods have been proposed to evaluate hepatic function prior to liver surgery including those based on liver perfusion and biliary excretion (e.g. indocyanine green), on microsomal or cytosolic function (e.g. aminopyrine breath test, galactosyl elimination capacity) and on functional imaging (e.g. 99m-Tc-galactosyl-human serum albumin scintigraphy) 840414243. Redealli et al established in a prospective study that a galactosyl elimination capacity < 4.0 mg/min/Kg is a strong predictor of complications after hepatic surgery in patients with HCC and liver cirrhosis 41. Among all, the most commonly used is the indocyanine green (ICG) clearance test widely used in eastern countries for cirrhotic patients with hepatocellular carcinoma requiring liver surgery 4044. The ICG clearance test evaluated the retention rate of the substance after 15 minutes wich is actively transfer into the hepatic parenchyma, which seems to be the only mechanism involved in metabolism. Finally it is secreted into the bile 8. The ICG test values are determined by blood sampling or as the percent retention determined by pulsed spectrophotometry in a similar fashion to an oxygen saturation monitor 45. However, the grater limitation of this test is the variation of results associated with changes in hepatic blood flow. While patients classified as Child-Pugh score A with ICG retention rate <14 % are eligible for large resection, those with results between 15% and 20% are only suitable for limited resections 46.

Some authors combined the results of Child-Pugh score, ICG retention rate at 15 minutes and other biochemical results to define the magnitude of resection, or to decide the use of a preoperative portal embolization to achieve a compensatory hypertrophy of the future liver remnant 147. For example, Wo et al sustained that in patients with ICG retention rate >50%, total bilirubin >4 mg/dl or protrombine time <75%, surgery should not be performed 48. Taking these parameters into account there have been low mortality rates attributed to liver failure and acceptable morbidity 464950.

However, the grater limitation of these tests is the variation of results associated with changes in hepatic blood flow and others entities affecting hepatic function leading some authors to discuss the accuracy of this method 5152. For example, Herold et al found unequal results when comparing tests for hepatic metabolism as galactosyl elimination capacity or aminopyrine breath test with ICG test in patients affected with chronic hepatic diseases 42. Currently, there is no functional test capable to define on its own the degree of functional reserve and the extent of resection. Therefore, a general algorithm used to evaluate the functional liver reserve in patients with chronic hepatitis or cirrhosis in many centers include clinical evaluation in the search of co-morbidities, blood testing including platelet count and coagulation studies, ICG test, hepatic volume measured by imaging and Child-Pugh classification 4447.

The presence of severe portal hypertension is an accepted contraindication for liver resection 53. To exclude this situation, the grade of portal hypertension must be evaluated during preoperative work-up with platelet count, measurement of spleen by imaging studies and endoscopy to assess the presence of esophagus varices 44. In case of esophagic varices grade III, the patient must be endoscopically treated before surgery 54. Some authors recommended the systematic measurement of portal pressure gradient to select patients more accurately 1153. Bruix et al demonstrated that a portal pressure gradient higher than 10 mmHg was associated with increased morbidity and reduced survival 53. Furthermore in those with associated high bilirubin mortality was increased 11. As a consequence, a systematic preoperative measurement of portal pressure gradient has been used in many centers to better select cirrhotic patients with liver tumors 1153. Others proposed to delay or avoid hepatectomy in virus related cirrhotic patients with high hepatic transaminases to reduce postoperative mortality 47.

Another important factor to take into account when hepatic regeneration is desired is cholestasis. In this specific situation, most authors agree to perform a preoperative biliary percutaneous drainage of the hemi-liver that will remain after surgery 55. Nevertheless, many variables such as the type and etiology of the underlying liver disease, the extent and location of the lesions, make the experience of the treating team of primary importance to define the extent of the liver resection in each particular patient 6. Co-morbidities tend to start or enlarge perioperative complications, so it is sensitive to achieve ASA score I-II prior to surgery because it reduces risk 47. Moreover, there are entities related to cirrhotic status that can be resolved during surgery to prevent further complications. For example, some authors recommended to perform splenectomy during hepatectomy to reduce postoperative medical complications in patients with severe hyperesplenism 5657.

Evidence from liver transplant suggests that the donor graft volume must represent ≥ 0.8% of the total body weight of the receptor to reduce the risks of acute hepatic failure 58. However, the regeneration of chronically ill liver is not equal than that observed in normal organs 8395960, thus surgeons must be extremely cautious when defining the extent of functional remaining parenchyma to avoid postoperative hepatic failure. Makuuchi et al proposes preoperative right portal vein embolization (PPVE) as a useful attempt to increase the volume of the future remaining liver 61. Azulay et al demonstrated that resecability can be achieve when PPVE is indicated in cirrhotic patients when the estimated future remnant liver parenchyma ≤40% 62. When the liver does not regenerate after PPVE, most agree that major hepatectomy should be contraindicated to avoid severe postoperative liver failure 63. Today, most authors recommend to perform PPVE when the remaining liver is ≥ 25–30% of the whole organ in normal liver and ≥ 40% in injured livers 40474962. In eastern countries, where ICG test is widely used in cirrhotic patients, PPVE is recommended when the non-tumoral parenchyma to be removed is ≥ 40% of the organ together with ICG test of 10% to 20% 50.

A comparative analysis of morbidity and mortality in cirrhotic patients undergoing liver surgery is very limited as a result of the absence of a standard manner to report postoperative complications. Moreover, most of the studies involve cirrhotic as well as other patients with different hepatic disorders such as fibrosis, cholestasis or hepatitis. Finally, there is no description of severity degree of postoperative complications and often the cause of death is not reported (Table 2).

The evolution of hepatic surgery together with the introduction of novel surgical strategies and modern transection techniques has led to a dramatic reduction in mortality rates from 10–32% 76465 to current series displaying mortality rated of <8% (Table 2). Although post-operative hepatic failure has also been reduced to <5% 465066, most of the different surgical teams evaluate failure in their own fashion.

Many liver surgeons advocated to the conventional approach for the surgical treatment of these patients to reduce hemorrhagic complications. However, some authors have suggested that the laparoscopic approach when feasible can reduce complications related to the abdominal wall or ascitis 676869. The use of systematic drainage of abdominal cavity after hepatic surgery in cirrhotic liver has no proven benefits 7071 and there is also an association with infections, abdominal wall complications and prolonged hospital stay 72.