In 1979, Bynum et al 8 coined the term “ischemic hepatitis” to refer to a liver injury characterized by a centrilobular liver cell necrosis with a sharp increase in serum aminotransferase activity in the setting of cardiac failure. They proposed the expression “hepatitis” because of some clinical similitudes (anorexia, malaise, jaundice, tender hepatomegaly) with infectious hepatitis, and they proposed the expression “ischemic” because they assumed that the liver cell necrosis resulted only from the failure of hepatic blood perfusion. This last assertion was provocative. Indeed, until then, it had been widely held that the liver cell necrosis observed in patients dying from congestive heart failure was the result of passive congestion of the liver and not of liver ischemia 41,51. Then, 2 years after that milestone study, Arcidi et al 4 published an authoritative study destined to become one of the most widely quoted references in the field. They maintained that “...hypoxia sufficient to produce centrilobular necrosis requires the presence of shock.” Accordingly, the liver injury related to cardiac and circulation failure became recognized as “shock liver,” and now it is largely agreed that ischemia, not passive congestion of the liver, is the mechanism responsible for the liver cell necrosis 25.
During a 10-year period, we prospectively identified and collected data for 142 episodes of “ischemic hepatitis” in the intensive care unit (ICU) of a general hospital. Of this large series, only some particular reports 29–32 have been published. As a final report, we present here the clinical and hemodynamic findings in the whole series, and we compare these results with those already published. We have become more and more aware that liver ischemia, that is, a decrease in hepatic blood flow, is not the sole hemodynamic mechanism responsible for the liver cell necrosis, and accordingly we have proposed renaming this syndrome “hypoxic hepatitis”28.
PATIENTS AND METHODS
Case Identification and Study Setting
Cases of hypoxic hepatitis were consecutively identified during a 10-year period (April 1983 to March 1993) in the ICU of a 600-bed general hospital. Episodes of hypoxic hepatitis were identified according to 3 widely accepted criteria 20,21,34,50: 1) a clinical setting of cardiac, circulatory, or respiratory failure; 2) a dramatic but transient increase in serum aminotransferase activity reaching at least 20-fold the upper limit of normal (ULN); 3) the exclusion of other putative causes of liver cell necrosis, particularly viral or drug-induced hepatitis. Liver biopsy was not required for the diagnosis of hypoxic hepatitis, in agreement with other studies showing that a histologic confirmation is unwarranted 20,34 and even inadvisable 21 when the 3 criteria listed above are met. Cases with high levels of serum aminotransferase activity following cardiac or liver surgery were not included.
Definitions of the Clinical Syndromes
Congestive heart failure referred to the association of 1) biventricular heart failure recognized by a cardiologist, 2) clinical signs of left- and right-sided heart failure or current treatment for heart failure, and 3) cardiomegaly with pulmonary artery congestion on chest X-ray. Acute cardiac failure corresponded to an acute cardiac event responsible for cardiac failure in the absence of chronic heart failure. Chronic respiratory failure referred to a chronic pulmonary disease with respiratory impairment identified by a pulmonologist based on clinical history, physical examination, chest X-ray, and pulmonary function tests. An acute exacerbation of chronic respiratory failure corresponded to an episode of increased respiratory compromise without an objectively documented cause according to Seneff et al 55. Circulatory shock referred to noncardiogenic shock and encompassed septic, toxic, hemorrhagic, and hypovolemic shock. A clinical state of systemic hypoperfusion was indicated by cutaneous signs of hypoperfusion (cold, clammy, cyanotic extremities) requiring treatment by inotropic drugs. A shock state was defined as a fall of systemic arterial pressure (SAP) below 90 mm Hg requiring urgent treatment by inotropic drugs and/or fluid infusion.
Hemodynamic Assessment and Hepatic Blood Flow Measurement
All patients with hypoxic hepatitis underwent an elementary hemodynamic evaluation including SAP, central venous pressure (CVP), and arterial blood gas analysis. CVP was monitored through an internal jugular vein catheter, while a radial artery catheter was inserted to collect blood samples and monitor blood pressure. Arterial blood gases were analyzed on a Corning 175 or 288 pH/blood gas analyzer.
Normal values for the elementary hemodynamic evaluation were as follows:EQUATION
For these data, with the exception of lactates, the initial value at the start of the episode of hypoxic hepatitis was used. For lactates, the more elevated value registered during the first hours of the hypoxic hepatitis episode was used.
In 61 cases of hypoxic hepatitis, a more detailed hemodynamic evaluation was carried out through insertion of a Swan-Ganz catheter (Edwards Laboratories) via the internal jugular vein. Pulmonary artery catheterization was performed when needed for monitoring and treatment. Cardiac output was measured by thermodilution method, and cardiac index (CI) was calculated by dividing cardiac output by body surface. Oxygen delivery (DO2) was calculated as follows: DO2 = CI × 10 × 1.38 × Hb × SaO2. Systemic vascular resistance (SVR) was calculated according to the formula: SVR = (MAP − CVP)/CI × 80, where MAP represents the mean arterial pressure.
Normal values for the pulmonary artery catheterization assessment were as follows, according to Shoemaker 58:EQUATION
A low DO2 of 330 mL/min.m2 was considered the critical threshold for aerobic metabolism in accordance with several studies 14,49,56. Only data derived from pulmonary artery catheterization performed within 24 hours after the clinical onset of hypoxic hepatitis were taken into account. The first recorded data were used in the statistical analysis.
During an 18-month period, hepatic blood flow was assessed in some patients with and without hypoxic hepatitis. These hepatic blood flow measurements were performed in 13 control patients, in 22 consecutive patients with hypoxic hepatitis (18 related to congestive heart failure and 4 related to chronic respiratory failure), in 40 consecutive patients with decompensated chronic heart failure without hypoxic hepatitis, and in 14 consecutive patients with severe hypoxemia free of hypoxic hepatitis. The controls were 13 patients in the ICU without cardiac disease or hemodynamic instability. The 40 patients with decompensated congestive heart failure without hypoxic hepatitis had clinical signs of systemic hypoperfusion necessitating inotropic treatment by dobutamine. The 14 patients with hypoxemia free of hypoxic hepatitis were consecutively admitted to the ICU for exacerbation of chronic respiratory failure not requiring immediate mechanical ventilation. They exhibited severe arterial hypoxemia (PaO2 below 60 mm Hg) at the time of hepatic blood flow measurement. Hepatic blood flow was measured by the technique of galactose clearance at low concentrations as described by Henderson et al 26,27. Practical details of the method and validation tests in our center have been reported elsewhere 29. Briefly, galactose (Janssen Chemica, Belgium) in 5% solution was perfused at a constant rate of 60 mL/h (corresponding to 50 mg/min) after a 10 mL-bolus (500 mg). Blood samples were collected at time 0 (before bolus) and 5 times during the steady-state period from 60 to 100 min. Galactose concentrations were measured by an enzymatic test and the absorbance of the NADH produced during the reaction was measured at 334 nm by spectrophotometry (Perkin Elmer, Lambda 2). This method provides accurate determinations of low concentrations of galactose between 1 and 5 mg/dL 26,27. Galactose clearance was calculated by dividing the galactose infusion rate (equal to the elimination rate during steady state) by the peripheral arterial concentration of galactose calculated as the mean of 5 measurements during steady state. The data on hepatic blood flow reported here were all recorded during the first 12 hours following admission to the ICU for hemodynamic instability, or following diagnosis of hypoxic hepatitis.
Data were prospectively collected on individual files containing the following: 1) clinical data including medical history; current therapy; reason for admission; information from cardiac, respiratory, and liver examination on admission; 2) pertinent blood laboratory analyses including bilirubin, creatinine, prothrombin activity, hemoglobin, glycemia, serum aspartate aminotransferase activity (ASAT), serum alanine aminotransferase activity (ALAT), and serum lactic dehydrogenase activity (LDH); 3) hemodynamic data; 4) therapeutic intervention and short-term outcome.
Biochemical and hemodynamic data were recorded on admission, 12 hours after admission, then every day until discharge from the ICU or death. A retrospective review of the hospital records of all cases was undertaken in 1998 to analyze the long-term outcome and to collect lacking data.
Survey of the Literature
The survey of the literature encompassed a personal library of journals and reprints collected for more than 20 years, the search for relevant references cited in the bibliography of identified articles, and a MEDLINE (National Library of Medicine, Bethesda, MD) search using the key words “hypoxic hepatitis,” “shock liver,” and “ischemic hepatitis.” Professor Felicity Hawker and Professor Hiroshi Yasuda kindly provided us with additional data on their series 24,67. For comparison between published series and the current series, we did not take into account autopsy series where hypoxic hepatitis was retrospectively identified from histopathologic material, pediatric series, and series of heat stroke, even if this last clinical syndrome represents a particular form of hypoxic hepatitis. Moreover, only adult series comprising at least 4 cases were included, to avoid selection of atypical cases published individually.
Means were compared using the Mann-Whitney rank sum test. One-way ANOVA was used to compare means of 3 or more groups and the Tukey method was used to compare all pairs of means. To compare groups that did not fit gaussian distribution, Kruskal-Wallis statistics were used, with the Dunne multiple comparison test for comparison of all pairs of means. The chi-square test was used to compare qualitative data.
One hundred forty-two episodes of hypoxic hepatitis were consecutively identified in 139 patients (3 having 2 episodes) during a 10-year period (April 1983 to March 1993) in the ICU of a 600-bed general hospital. The mean age was 68.5 years (range, 15–98 yr) and the male:female ratio was 2 (93:46). Among the 142 episodes, 108 occurred in patients urgently attending the emergency ward before transfer to the ICU. The other 34 episodes occurred in patients already hospitalized. These 142 episodes of hypoxic hepatitis accounted for 0.9% of the 15,619 admissions to the ICU during the same period. The prevalence of hypoxic hepatitis was higher in the coronary care unit (89 episodes/6,603 admissions, 1.3%) than in the surgical and medical ICU (53 episodes/9,016 admissions, 0.6%) (p < 0.001).
Underlying Conditions Leading to Hypoxic Hepatitis
One hundred thirty-eight episodes could be separated into 4 main groups based on clinical features, as described above: 80 episodes occurred during decompensated congestive heart failure, 20 episodes during acute cardiac failure, 19 episodes during exacerbated chronic respiratory failure, and 19 episodes during circulatory shock (Figure 1). Only 4 episodes could not be classified into these 4 main groups. Two of them resulted from extreme hypoxemia in the setting of sleep apnea syndrome and 2 were massive hepatocellular necrosis secondary to diffuse metastatic infiltration of the liver.
Hypoxic Hepatitis Related to Congestive Heart Failure
Congestive heart failure was the underlying condition in 80 of the 142 episodes (56%) (see Figure 1). This group included 54 men and 24 women, 2 men having 2 episodes of hypoxic hepatitis each. The mean age was 70 years (range, 32–90 yr). The most obvious cause of congestive heart failure was ischemic heart disease in 60 episodes (75%). For 74 of the 80 episodes, congestive heart failure had been recognized before admission and was being treated. A progressive deterioration of the cardiac function (dyspnea, ankle edema) had been noted shortly before admission in 74 episodes. In every episode except 1, the occurrence of hypoxic hepatitis was triggered by 1 or more acute events, most often of cardiogenic origin. The most frequent acute events were cardiac arrhythmia and acute pulmonary edema observed in 42 (52.5%) and 27 (34%) episodes, respectively. Acute myocardial infarction was observed in only 10 of the 80 episodes occurring in the setting of congestive heart failure (12.5%), and pulmonary embolism accounted for 4 episodes. Hypoxic hepatitis related to congestive heart failure could also be precipitated by noncardiogenic acute events: bacterial infection and hyperthermia were present in 7 episodes, severe anemia in 2, and lower limb arterial thrombosis in 2.
Hypoxic Hepatitis Related to Acute Cardiac Failure
Acute cardiac failure was the underlying condition in 20 episodes of hypoxic hepatitis (14%), in 9 men and 11 women with a mean age of 68 years (range, 34–98 yr). None of these patients had a medical history of cardiac failure, and congestive heart failure could be ruled out easily by clinical examination and chest X-ray. The causes of acute cardiac failure included myocardial infarction (n = 8), pulmonary embolism (n = 5), pericardial tamponade (n = 3), thoracic trauma (n = 3), and sudden arrhythmia (n = 1).
Hypoxic Hepatitis Related to Chronic Respiratory Failure
Exacerbated chronic respiratory failure was the underlying condition in 19 episodes of hypoxic hepatitis (13%). This group included 16 men (1 with 2 separate episodes) and 2 women. The mean age was 67 years (range, 50–87 yr). By far the main cause of chronic respiratory failure was silicosis, related to coal mining in 11 patients (1 with 2 episodes) and stone cutting in 1. Other causes of chronic respiratory failure were chronic obstructive pulmonary disease in 5 episodes and cryptogenic fibrosing alveolitis in 1. All patients had long-lasting chronic respiratory failure. Fifteen patients were on chronic treatment with theophylline and inhaled sympathomimetics, and 11 were on oral corticotherapy. Twelve of the 18 patients had been admitted previously for bouts of acute respiratory failure. In 9 patients, respiratory function tests had been performed during the year, apart from the episodes of hypoxic hepatitis. In 8, forced expiratory volume was less than 35% of the predicted value, reflecting severe chronic obstructive pulmonary disease (stage III) according to the American Thoracic Society 3. All patients were admitted to the emergency ward for respiratory distress. None of these patients had evidence of chronic heart failure. Three patients had atrial fibrillation, but with heart rate not exceeding 120 beats per minute, and 1 had a brief cardiac arrest secondary to severe hypercapnia. For the other episodes, no contributing acute event could be identified except for exacerbation of respiratory insufficiency, possibly caused by bronchial infection as defined by Seneff et al 55. Mechanical ventilation was required in 14 episodes.
Hypoxic Hepatitis Related to Circulatory Shock
Protracted circulatory shock was the underlying condition in 19 episodes of hypoxic hepatitis (13%), in 13 men and 6 women with a mean age of 64 years (range, 15–80 yr). Toxic/septic shock was the main cause of circulatory shock in all episodes, whereas hypovolemia (upper gastrointestinal hemorrhage in 3 and multiple hematoma in 2) could have played a role in 5. Blood cultures yielded pathogens in 9 episodes, whereas tissue necrosis was present in 9 episodes affecting muscles in 7 episodes (femoral artery thrombosis in 3, multiple hematoma in 2, and rhabdomyolysis in 2) and the digestive tract in 2 (mesenteric infarct in 1, intestinal volvulus in 1). It is noteworthy that alcoholic cirrhosis was an underlying condition in 5 episodes in this group, and that in 4 of them, the shock state resulted from cirrhosis-related complications—variceal hemorrhage with concomitant sepsis in 3 and spontaneous peritonitis in 1.
Clinical, Laboratory, and Histologic Data
The clinical findings in the whole series and in the 4 main groups are reported in Table 1. A shock state, as defined above, was observed in 78 cases (55%), but clinical signs of systemic hypoperfusion were observed in 86% of the cases on admission. Clinical signs of right-sided heart failure such as tender hepatomegaly, ankle edema, and hepatojugular reflux were observed in about 50% of the cases (see Table 1). Clinical signs of liver injury remained generally in the background. Overt jaundice was rarely observed (23 episodes; 16%) and occurred late in the course of hypoxic hepatitis. Some degree of mental impairment was observed in more than half of the episodes, but, given brain hypoxia, this is a common feature in critically ill patients regardless of the cause.
Pertinent laboratory data in the whole series and in the 4 main groups are reported in Table 2. Hypoxic hepatitis was heralded by a striking elevation of serum ASAT, ALAT, and LDH that rapidly recovered in about 7–14 days (Figure 2). Serum aminotransferase activities and LDH were often grossly elevated from the initial measurement. Values of serum aminotransferase activities already above 5 ULN were observed in 77 episodes (54%) on the first evaluation. In a few episodes, the first measurement of serum aminotransferase activities was the highest, so the true peak value could have been missed. This remained unusual however, as the peak values of serum ASAT, ALAT, and LDH could be identified in 87%, 93%, and 82%, respectively (see Table 2). The complete pattern of the course of serum enzyme activities could be drawn from only 25 episodes where 3 conditions were met: initial serum aminotransferase activities were normal or less than 2-fold the ULN on the initial measurement; survival was at least 8 days; and a complete set of data was available (see Figure 2). The ASAT peak was generally more elevated than the ALAT peak, but this was not a constant rule. The supremacy of 1 serum aminotransferase peak activity, either ASAT or ALAT, could be inferred from 129 episodes in this series; the ASAT peak was the more elevated in 97 episodes (75%). The increase in serum enzyme activities was not significantly different between the 4 main groups of patients, except for admission levels in the chronic respiratory failure group and peak levels in the circulatory shock group (see Table 2).
Another biochemical hallmark of hypoxic hepatitis was the fall of prothrombin activity. The nadir of prothrombin activity was observed as early as the first day, and the complete recovery took about 1 week (see Figure 2). Prothrombin activity fell under 50% in 113 episodes (79.5%) and below 20% in 20 episodes (14%). A mild elevation of serum bilirubin was a common finding, but rarely progressed to overt jaundice (see Table 2). Serum bilirubin higher than 5 mg/dL was observed in 23 episodes (16%) and higher than 10 mg/dL, in only 5 episodes (3.5%). Serum creatinine increased above 2 mg/dL in 93 episodes (65.5%) and above 5 mg/dL in 19 episodes (13.4%) (see Table 2). Renal function impairment was significantly more severe in hypoxic hepatitis in the setting of circulatory shock (see Table 2). Hypoglycemia, which has been regarded by some investigators as a distinct feature 19, was rarely observed. Glycemia below 70 mg/dL was observed in only 4 episodes. Conversely, as pointed out by Gitlin and Serio 21, hyperglycemia was a common observation: glycemia levels higher than 200 mg/dL at admission was observed in 39 episodes (27.5%).
Liver histology was assessed by percutaneous biopsy or postmortem puncture in 56 cases and showed centrilobular liver cell necrosis in each instance.
The elementary hemodynamic data recorded on admission in the whole series and in the 4 main groups are reported in Table 3. The hemodynamic assessment by pulmonary artery catheterization performed in 61 episodes within the first 24 hours is reported in Table 4.
Hypoxic Hepatitis Related to Congestive Heart Failure
Clinical signs of systemic hypoperfusion were present in 74 of the 80 episodes (92.5%), but a shock state was observed in only 33 episodes (41%) (see Table 1). Patients with hypoxic hepatitis related to congestive heart failure had low SAP associated with high CVP and moderately decreased PaO2 (see Table 3). Cardiac index was decreased, while SVR was increased (see Table 4). As a result of low CI, DO2 was depressed and close to the critical value of 330 mL/min.m2 (discussed in Patients and Methods).
Hypoxic Hepatitis Related to Acute Cardiac Failure
Twenty episodes of hypoxic hepatitis were considered hypoxic hepatitis related to acute cardiac failure. A shock state (70% of cases) was significantly more frequent in this group than in the congestive heart failure group (see Table 1). In other respects, the hemodynamic pattern in this group did not differ from the pattern observed in the congestive heart failure group (see Tables 3 and 4), and shared the common features of cardiac failure. The role of hypoxemia (median PaO2, 71 mm Hg) was negligible and did not account for the severely depressed DO2 (median, 271 mL/min.m2) far below the acknowledged critical value. The low level of DO2 was explained only by the fall of CI.
Hypoxic Hepatitis Related to Chronic Respiratory Failure
Nineteen episodes were caused by an acute exacerbation of chronic respiratory failure in the absence of left-sided heart failure. A shock state, not resulting from the hemodynamic collapse induced by mechanical ventilation, was observed in 10 episodes (53%) (see Table 1). The main hemodynamic feature in the episodes of hypoxic hepatitis related to chronic respiratory failure was the severe arterial hypoxemia, significantly more depressed compared with hypoxic hepatitis related to congestive heart failure (see Table 3). In this group, CI tended to be increased, and, therefore, the low DO2 close to the critical value could be explained only by extreme hypoxemia (see Table 4).
Hypoxic Hepatitis Related to Circulatory Shock
Nineteen episodes of hypoxic hepatitis were in the group related to circulatory shock, wherein a shock state was a constant feature (see Table 1). Protracted shock, lasting for more than 6 hours, was observed in 10 of the 19 episodes. Reflecting the severity of shock, blood lactates were strikingly increased (median, 9.8 mEq/L). CVP was low and PaO2 was normal (see Table 3). Hemodynamic assessment through pulmonary artery catheterization showed that CI was not decreased and even was frequently increased as a result of decreased SVR. Subsequently, DO2 was generally preserved and significantly higher than in hypoxic hepatitis related to congestive heart failure (see Table 4).
Hepatic Blood Flow Measurement
The results of hepatic blood flow are reported in Table 5. Compared with controls, patients with congestive heart failure had significantly lower hepatic blood flow, but patients with congestive heart failure and hypoxic hepatitis had more depressed hepatic blood flow than patients with congestive heart failure without hypoxic hepatitis (p = 0.001). In contrast, hepatic blood flow in patients with severe hypoxemia without hypoxic hepatitis was not different from hepatic blood flow in controls. However, when chronic respiratory failure was accompanied by hypoxic hepatitis, hepatic blood flow was more decreased and significantly lower than hepatic blood flow in control patients.
The 1-month mortality rate was 52.8% (75/142), and the 1-year survival was 28.3% (39 of the 138 episodes where long-term outcome could be determined).
Survey of the Literature
By surveying the literature we identified 31 series including 479 patients. These series are reported in chronologic order in Tables 6 and 7. Two series comprising at least 4 cases 12,46 were not reported in the tables because their cases were reported in other publications 34,37. Cases of hypoxic hepatitis were identified from clinical registries in most series but also by watching for markedly elevated serum aminotransferases in the laboratory of the hospital, and in 1 series from histopathologic records (see Table 6). Generally, series comprising cases identified through the surveillance of markedly elevated serum aminotransferases in the laboratory included more patients in a shorter time (see Table 6). Only 5 studies were prospective, 1 with cases identified through clinical records 48 and 4 with cases identified through aminotransferase surveillance in the laboratory 34,36,62,66. The noteworthy clinical data are reported in Table 6. The male:female ratio was 2.7 (233 male:86 female) and the mean age for the 333 patients with available data was 71.2 years. In 1 large series 66 including 68 cases, mean age was not reported but median age was 75 years. Primary heart disease was underlying hypoxic hepatitis in most cases. From available data, primary heart disease was present in 81% (372/458), most often as congestive heart failure (63%; 243/385). Acute myocardial infarction was reported in 16% of the cases (70/432), more frequently in older series 7,22,38,57.
Conversely, chronic respiratory failure was rarely underlying hypoxic hepatitis and was present in only 11% of the cases (52/458). Moreover, chronic respiratory failure was generally associated with left-sided heart failure. Hypoxic hepatitis occurred in isolated chronic respiratory failure in only 7 cases. Circulatory shock was the underlying condition in 16.5% of the cases (76/458), but in 1 study was the most frequent cause of hypoxic hepatitis 66. A shock state or severe systemic hypotension was reported in 64% of the cases (237/369).
The mean peak levels of serum ASAT, ALAT, LDH, bilirubin, and creatinine and the mean nadir value of prothrombin activity are reported in Table 7. In several studies, serum enzyme activities were reported in IU/L without giving the normal range for the laboratory 10,22,37,44,52,54,62,63,67. For these studies, a value of 40 IU/L for aminotransferase levels and a value of 300 IU/L for LDH level were considered as the ULN. Mean values of ASAT peaks ranged from 21 to 247 ULN; of ALAT peaks, from 19 to 87 ULN; of LDH peaks, from 3.1 to 34 ULN; of bilirubin peaks, from 2.2 to 14 mg/dL; of creatinine peaks, from 2 to 5.1 mg/dL; and of prothrombin activity nadirs, from 16% to 54% (see Table 7).
The inhospital mortality rate was 52% (233/446).
The current study demonstrated that, in contrast with a still largely held opinion, hypoxic hepatitis is not a rare condition. One hundred forty-two episodes were prospectively identified during 10 years, a prevalence of 0.9% of all ICU admissions in our general center. The underrated frequency of hypoxic hepatitis may be explained by the numerous studies including cases retrospectively recruited in departments of internal medicine, where it is unusual for these patients to be hospitalized. For example, when Birgens et al 6 coined the term “shock liver,” they reported just 5 cases identified during 13 years in “different hospitals in Copenhagen,” and when Bynum et al 8 coined the term “ischemic hepatitis,” they reported 7 cases identified in 5 years. Conversely, studies carried out in ICUs have provided figures similar to those in the current study. So, in an Israeli hospital, Fuchs et al 19 observed 32 cases of hypoxic hepatitis among 2,155 consecutive admissions to the ICU (prevalence, 1.5%) during 39 months, while the prevalence in the internal medicine wards was only 0.03% (3/9,977). Moreover, studies that prospectively identified cases through a systematic search for dramatic elevation of serum aminotransferase activity at the level of the hospital laboratory have shown that hypoxic hepatitis was the main cause of such an elevation, accounting for more than 50% of the cases 34,36,62,66.
As already pointed out in other large series, the clinical signs of liver injury remained generally in the background of the clinical signs of the underlying condition, while the dramatic but rapidly resolving elevation of serum enzyme activities, the profound fall of prothrombin activity, and the frequent alteration of renal function formed a triad of biochemical abnormalities that is unusual in cases of viral or drug-induced hepatitis and that strongly suggests the diagnosis of hypoxic hepatitis. It is important to stress that a shock state was not a prerequisite to the occurrence of hypoxic hepatitis. Hypoxic hepatitis occurred in a clinical setting of life-threatening disease, as illustrated by the admission to hospital through the emergency ward in 76% of the cases, but a shock state was present in only 55% of the episodes. This figure is in line with the one reported in most large series (see Table 6) but is in sharp contrast with the authoritative statement of Arcidi et al 4 that a shock state is a prerequisite to the occurrence of hypoxic hepatitis. Subsequently, we believe that the commonly used term “shock liver” should be disregarded.
An important contribution of the current study is better definitions of the underlying conditions leading to hypoxic hepatitis. All but 4 episodes were easily separated in 4 main groups merely on clinical grounds (see Figure 1). The most frequent underlying condition was cardiac failure (100 episodes, 70%) and particularly congestive heart failure (80 episodes, 56%). This was in agreement with the survey of the literature, wherein we found a prevalence of 81% for a primary heart disease and 63% for congestive heart failure (see Table 6). The episodes of hypoxic hepatitis related to congestive heart failure were generally (92%) preceded by a period of progressive deterioration of cardiac function and were precipitated in all cases but 1 by an acute event, more often arrhythmia or acute pulmonary edema, as also reported by others 54. Conversely, acute myocardial infarction was the triggering event in only 18 episodes (12.5%), 10 in the congestive heart failure group and 8 in the acute cardiac failure group. This figure was also in agreement with the data of the literature, where the prevalence of acute myocardial infarction was 16% but could exceed 50% in earlier studies (see Table 6).
A particularity of the current study was the group of 19 episodes of hypoxic hepatitis occurring in patients with severe chronic respiratory failure without left-sided heart failure. With the exception of a series originating from South Wales, where chronic respiratory failure or severe hypoxemia accounted for 16% of the episodes of hypoxic hepatitis 66, isolated chronic respiratory failure without left-sided cardiac failure was rarely an underlying condition leading to hypoxic hepatitis (see Table 6). A possible explanation for this discrepancy could be the high frequency of coal mine pneumoconiosis in our region. Indeed, coal mine exploitation was the main industrial activity until 2 decades ago, and the silicosis mortality rate in Belgium was the highest in western countries until 10 years ago 65. Twelve of the 19 episodes of hypoxic hepatitis related to chronic respiratory failure occurred in patients suffering from coal mine pneumoconiosis.
The last frequent underlying condition leading to hypoxic hepatitis was circulatory shock (19 episodes). By definition, circulatory shock encompassed etiologies of shock other than cardiogenic and included septic, toxic, traumatic, hemorrhagic, and hypovolemic shock. However, hypoxic hepatitis from simple hemorrhagic or hypovolemic origin was never encountered in the current series, even if digestive hemorrhage and extensive hematoma were contributing conditions in 5 episodes. In every episode of hypoxic hepatitis related to circulatory shock, the role of some toxins released from bacterial pathogens or tissue damage could be suspected on the grounds of causal events, clinical signs of sepsis, or bacteriologic results. The prevalence of circulatory shock in other series of hypoxic hepatitis ranged from 0%16,41 to 53%51, with a mean prevalence of 16.5%. We note that in a series from South Wales where the prevalence of circulatory shock was as high as 53% (36/68 cases), sepsis was the cause of circulatory shock in 89%66.
Finally, the main contribution of the current study is to help determine the hemodynamic mechanisms leading to hypoxic hepatitis. Since the often-quoted review of Dunn et al 17, it has been agreed that 3 hemodynamic mechanisms theoretically may result in hypoxic injury of the liver: ischemia due to decreased hepatic blood flow, venous congestion due to right-sided heart failure, and hypoxemia due to decreased oxygen content in the afferent blood. However, to the best of our knowledge, the role of these putative mechanisms was not assessed in a human study. In 2000, Seeto et al 54 pointed out the high prevalence of venous congestion of the liver in hypoxic hepatitis occurring in the setting of heart failure, but this study was based only on clinical features without hemodynamic assessment. The current study demonstrated that liver ischemia was not the sole hemodynamic determinant of hypoxic hepatitis, but that the role of venous congestion was important in most cases related to heart failure, that the role of arterial hypoxemia was crucial in cases related to chronic respiratory failure, and that another mechanism should be proposed in cases related to circulatory shock.
When cardiac failure was the underlying condition of hypoxic hepatitis, the role of ischemia was suggested by the fall of SAP, and was directly supported by the measurements of DO2 and hepatic blood flow. Oxygen delivery that depends on cardiac output and oxygen content of the arterial blood was close to the critical value only because of decreased cardiac output. Hepatic blood flow was depressed in decompensated congestive heart failure with and without hypoxic hepatitis, but it was significantly more depressed when hypoxic hepatitis was present (see Table 5). The role of venous stasis was suggested by the high prevalence of clinical signs of right-sided heart failure (see Table 1) and was directly supported by the elevated CVP observed in all patients with hypoxic hepatitis related to congestive heart failure. Moreover, we have shown that in decompensated congestive heart failure, CVP was significantly higher when hypoxic hepatitis was present 29. It may be postulated that in the case of hypoxic hepatitis related to congestive heart failure, the liver is chronically exposed to some degree of hypoxia by long-lasting venous congestion. At the time of an acute cardiogenic event, even a moderate, brief, and sometimes unrecognized fall in SAP may worsen hypoxia sufficiently to induce liver cell necrosis. This could explain why a shock state was not systematically required, why hypoxic hepatitis might occur in clinically unrecognized cardiomyopathy 12,33, and why hypoxic hepatitis was observed mainly in cardiogenic shock and rarely in hemorrhagic or hypovolemic shock. In hypoxic hepatitis related to acute cardiac failure, the liver was probably less sensitized by long-lasting venous congestion, and therefore the occurrence of hypoxic hepatitis probably required a more severe fall of hepatic blood flow during the acute event. This was supported in the current study by a state of shock significantly more frequent (see Table 1), by blood lactates more elevated (see Table 3), and by DO2 more decreased (albeit not significantly) (see Table 4) in hypoxic hepatitis related to acute cardiac failure than in hypoxic hepatitis related to congestive heart failure. Otherwise, the hemodynamic profile of hypoxic hepatitis occurring in the setting of acute cardiac failure did not differ from that observed in congestive heart failure.
When chronic respiratory failure was the underlying condition of hypoxic hepatitis, the degree of arterial hypoxemia was impressive, with levels of PaO2 below 40 mm Hg in 15 of the 19 episodes. Otherwise, patients with hypoxic hepatitis secondary to cardiac or respiratory failure did not differ regarding the prevalence of shock (see Table 1), the degree of systemic hypotension, and the degree of CVP elevation (see Table 3). If the CVP elevation could be explained easily by pulmonary artery hypertension, a common feature in chronic respiratory failure, severe systemic hypotension was rather unexpected in these patients without left cardiac failure. Therefore, if it was agreed that hypoxic hepatitis generally resulted from the association of systemic hypotension and CVP elevation, it should be admitted that arterial hypoxemia, even severe, was not a prerequisite for the development of hypoxic hepatitis. Thus, the elementary hemodynamic assessment tended to confirm the widely held concept that arterial hypoxemia was not a relevant hemodynamic determinant of hypoxic hepatitis. Eventually, the crucial role of arterial hypoxemia emerged clearly from the hemodynamic assessment through pulmonary artery catheterization. Indeed, although comparable low levels of DO2, close to the critical value, were recorded in patients with hypoxic hepatitis related to chronic respiratory failure and to congestive heart failure, the responsible hemodynamic mechanisms were quite different. In hypoxic hepatitis related to chronic respiratory failure, low DO2 was not the result of a decreased cardiac output as in congestive heart failure, but was secondary to extreme arterial hypoxemia only (see Table 4). Cardiac index tended to be increased, even, in hypoxic hepatitis related to chronic respiratory failure (see Table 4), an observation in agreement with other clinical studies 9,11,39. Since it is largely agreed, from the precursory study of Myers and Hickam 43, that hepatic blood flow represents a fixed percentage of cardiac output, it may be postulated that hepatic blood flow was preserved and even increased in episodes of hypoxic hepatitis related to chronic respiratory failure. To our knowledge, few studies on hepatic blood flow in respiratory failure have been carried out. Some animal studies have shown that hepatic blood flow was strikingly increased following hypoxemia or hypercapnia 1,53. In the current study, hepatic blood flow assessed in patients with chronic respiratory failure and severe hypoxemia (PaO2 less than 60 mm Hg), was normal in the absence of hypoxic hepatitis, but decreased when hypoxic hepatitis was present, probably in relation to low SAP (see Table 5). These results have to be interpreted cautiously because only 4 patients with chronic respiratory failure and hypoxic hepatitis have been studied. In summary, hypoxic hepatitis that occurs in exacerbated chronic respiratory failure is the result of extreme arterial hypoxemia, while venous congestion probably plays a contributing role. Adaptive mechanisms tend to increase CI and consequently hepatic blood flow in order to provide sufficient tissue oxygenation. When these adaptive mechanisms fail to preserve sufficient tissue oxygenation, liver cell necrosis ensues.
The hemodynamic pattern of hypoxic hepatitis occurring in the setting of circulatory shock was again different. In this group, a shock state was a constant feature. As expected, CVP was low and PaO2 was normal. Hence, according to the elementary hemodynamic evaluation, hypoxic hepatitis related to toxic/septic shock appeared obviously linked to the shock state and therefore caused by liver ischemia exclusively. However, when a more subtle hemodynamic assessment was conducted through pulmonary artery catheterization, the role of liver ischemia appeared less evident. Indeed, CI was not decreased and even was generally increased in patients with septic/toxic shock (see Table 4). Consequently, DO2 was preserved and significantly higher than in patients with hypoxic hepatitis caused by cardiac failure (see Table 5). These results were in agreement with the data of the literature. In case of septic/toxic shock, cardiac output and DO2 were generally well preserved and even increased due to low SVR, at least during the early stage 47,64. Splanchnic blood flow has been extensively studied in case of sepsis, and although controversial results have been reported in animal studies, most human studies have shown that splanchnic blood flow and thus hepatic blood flow were generally increased 15,16,23,59. Nevertheless, despite high CI and DO2, intraparenchymatous O2 tension of the liver remained low in the case of traumatic or septic shock 16. This paradox may be explained by the increased needs of the liver for oxygen and by its inability to extract oxygen. Indeed, despite elevated DO2, splanchnic oxygen uptake remains low in the case of sepsis 18,60. The mechanisms explaining the inability of the liver to extract oxygen in the case of sepsis are not fully understood, but endotoxins and proinflammatory cytokines seem to play a determinant role by affecting the cellular metabolism and the microcirculation functioning 68. In summary, hypoxic hepatitis occurring in septic/toxic shock is mainly the result of the liver’s inability to extract and use oxygen. Venous congestion and arterial hypoxemia are not contributing factors, whereas liver ischemia cannot be regarded as the motor of liver hypoxia, considering the usual increase in DO2 and hepatic blood flow.
In conclusion, the elementary hemodynamic assessment completed by a more precise evaluation through pulmonary artery catheterization in 61 patients validated the clinical separation of episodes of hypoxic hepatitis in 3 main groups: cardiac failure (chronic congestive and acute), respiratory failure, and septic/toxic shock. It demonstrated that hypoxic hepatitis results not only from an “ischemic” injury, but also is caused by the association of several hypoxic mechanisms, which differ depending on the underlying condition. Accordingly, the terms “shock liver” and “ischemic hepatitis” are misleading and should be replaced by the more general but more accurate expression “hypoxic hepatitis.” This suggestion is not only for semantic consideration; indeed, we encountered cases of hypoxic hepatitis where the diagnosis was correctly raised but then disregarded because of the absence of shock.
1. Adachi H, Strauss W, Ochi H, Wagner H. The effect of hypoxia on regional distribution of cardiac output in the dog. Circ Res. 1976; 39:314–319.
2. Aloy Duch A, Pont MM, Garcia Restoy E, Llebot Serna JJ. Hepatitis isquemica: descripcion de 11 casos. Rev Clin Esp. 1988; 183:255–258.
3. American Thoracic Society. Lung function testing: selection of reference values and interpretation strategies (Statement). Am Rev Resp Dis. 1991; 144:1201–1218.
4. Arcidi JM, Moore GW, Hutchins GM. Hepatic morphology in cardiac dysfunction. A clinicopathologic study of 1000 subjects at autopsy. Am J Pathol. 1981; 104:159–166.
5. Biasi F, Chiarpotto E, Lanfranco G, Capra A, Zummo U, Chiappino I, Scavazza A, Albano E, Poli G. Oxidative stress in the development of human ischemic hepatitis during circulatory shock. Free Radic Biol Med. 1994; 17:225–233.
6. Birgens HS, Henriksen J, Matzen P, Poulsen H. The shock liver. Acta Med Scand. 1978; 204:417–421.
7. Bloth B, de Faire U, Edhag O. Extreme elevation of transaminase levels in acute heart disease—a problem of differential diagnosis? Acta Med Scand. 1976; 200:281–288.
8. Bynum TE, Boitnott JK, Maddrey WC. Ischemic hepatitis. Dig Dis Sci. 1979; 24:129–135.
9. Cargill RI, Kiely DG, Lipworth BJ. Left ventricular systolic performance during acute hypoxemia. Chest. 1995; 108:899–902.
10. Cellarier G, Bonal J, Bouchiat C, Talaro PH, Dussarat GU. Foie ischemique aigu. Presse Med. 1995; 24:1418–1420.
11. Chabot F, Mertes PM, Delorme N, Schryen FU, Saunter CG, Polu J-M. Effect of acute hypercapnia on alpha atrial natriuretic peptide, renin angiotensin II, aldosterone, and vasopressin plasma levels in patients with COPD. Chest. 1995; 107:780–786.
12. Cohen J, Kaplan M. Left-sided heart failure presenting as hepatitis. Gastroenterology. 1978; 74:583–587.
13. Colivicchi F, Guerrera C, Bevilacqua E, Guerrera E, Melina D. L’epatite ischemica: contributo casistico e revisione della letteratura. Minerva Med. 1995; 86:379–386.
14. Creamer JE, Edwards JD, Nightingale P. Hemodynamic and oxygen transport variables in cardiogenic shock secondary to acute myocardial infarction, and response to treatment. Am J Cardiol. 1990; 65:1297–1300.
15. Dahn MS, Lange P, Wilson RF, Jacobs LA, Mitchell RA. Hepatic blood flow and splanchnic oxygen consumption measurements in clinical sepsis. Surgery. 1990; 107:295–301.
16. Dahn MS, Wilson RF, Lange P, Stone A, Jacobs LA. Hepatic parenchymal oxygen tension following injury and sepsis. Arch Surg. 1990; 125:441–443.
17. Dunn GD, Hayes P, Breen KJ, Schenker S. The liver in congestive heart failure: a review. Am J Med Sci. 1973; 265:174–189.
18. Edwards JD. Oxygen transport in cardiogenic and septic shock. Crit Care Med. 1991; 19:658–663.
19. Fuchs S, Bogomolski-Yahalom V, Paltiel O, Ackerman Z. Ischemic hepatitis. Clinical and laboratory observations of 34 patients. J Clin Gastroenterol. 1998; 26:183–186.
20. Gibson PR, Dudley FJ. Ischemic hepatitis: clinical features, diagnosis and prognosis. Aust N Z J Med. 1984; 14:822–825.
21. Gitlin N, Serio KM. Ischemic hepatitis: widening horizons. Am J Gastroenterol. 1992; 87:831–836.
22. Haiat R, Grivaux M, Soulie J, Diallo A, Chevalley J. L’elevation massive du taux des S-GOT en milieu cardiologique. Revue de la litterature a propos de 6 cas. Sem Hop Paris. 1970; 46:153–163.
23. Hanneman L, Meier-Hellmann A, Reinhart K. Regional blood flow in sepsis. Clin Intensive Care. 1992; 3(Suppl):28–30.
24. Hawker F. Ischaemic hepatitis [abstract]. Proceedings of the 9th NSW ANZICS meeting. Sydney, Australia, 1992.
25. Hawker F. Liver dysfunction in critical illness. Anaesth Intensive Care. 1991; 19:165–181.
26. Henderson JM, Fales FW. Continuous flow fluorometry of low galactose concentration in blood or plasma. Clin Chem. 1980; 26:282–285.
27. Henderson JM, Kutner MH, Bain RP. First-order clearance of plasma galactose: the effect of liver disease. Gastroenterology. 1982; 83:1090–1096.
28. Henrion J, Luwaert R, Colin L, Schmitz A, Schapira M, Heller FR. Hepatite hypoxique: etude prospective, clinique et hemodynamique de 45 episodes. Gastroenterol Clin Biol. 1990; 14:836–841.
29. Henrion J, Descamps O, Luwaert R, Schapira M, Parfonry A, Heller F. Hypoxic hepatitis in patients with cardiac failure: incidence in a coronary care unit and measurement of hepatic blood flow. J Hepatol. 1994; 21:696–703.
30. Henrion J, Minette P, Colin L, Schapira M, Delannoy A, Heller FR. Hypoxic hepatitis caused by acute exacerbation of chronic respiratory failure: A case-controlled, hemodynamic study of 17 consecutive cases. Hepatology. 1999; 29:427–433.
31. Henrion J, Colin L, Schmitz A, Schapira M, Heller FR. Ischemic hepatitis in cirrhosis. Rare but lethal. J Clin Gastroenterol. 1993; 16:35–39.
32. Henrion J, Colin L, Schapira M, Heller FR. Hypoxic hepatitis caused by severe hypoxemia from obstructive sleep apnea. J Clin Gastroenterol. 1997; 24:245–249.
33. Henrion J, De Maeght S, Schapira M, Ghilain J-M, Maisin J-M, Gerard R, Heller FR. Hypoxic hepatitis: a difficult diagnosis when the cardiomyopathy remains unrecognized and the course of liver enzyme follows an atypical pattern. A report of two cases. Acta Gastroenterol Belg. 1998; 61:385–389.
34. Hickman PE, Potter JM. Mortality associated with ischaemic hepatitis. Aust N Z J Med. 1990; 20:32–34.
35. Hilpert F, Barbare JC, Souied G, Ricome JL, Auzepy P. Formes pseudo-hepatitiques du foie cardiaque. Ann Cardiol Angeiol. 1979; 28:325–330.
36. Johnson RD, O’Connor ML, Kerr RM. Extreme serum elevations of aspartate aminotransferase. Am J Gastroenterol. 1995; 90:1244–1245.
37. Kanel GC, Ucci AA, Kaplan MM, Wolff HJ. A distinctive perivenular hepatic lesion associated with heart failure. Am J Clin Pathol. 1980; 73:235–239.
38. Killip T 3rd, Payne AM. High serum transaminase activity in heart disease. Circulation. 1960; 21:646–660.
39. Kontos H. Mechanism of circulatory responses to systemic hypoxia in anesthetized dogs. Am J Physiol. 1970; 209:397–401.
40. Kumada T, Himura Y, Iwata T, Nakamura Y, Fukuda Y, Fujiwara H, Kawai C. Liver function in congestive heart failure: abnormal elevation of serum hepatic enzyme and hepatic venous flow velocity. Jpn Circ J. 1989; 53:165–174.
41. Losowsky MS, Ikram H, Snow HM, Hargreave FE, Nixon PGF. Liver function in advanced heart disease. Br Heart J. 1965; 27:578–584.
42. Mastrapasqua G, Scapinello A, Madia D, Zappala G, Beltramello G, Virone C, Confortin L, Costa F. L’epatite ischemica. Descrizione di 4 casi e revisione della letteratura. Minerva Gastroenterol Dietol. 1993; 39:93–97.
43. Myers JD, Hickam JB. An estimation of the hepatic blood flow and splanchnic oxygen consumption in heart failure. J Clin Invest. 1948; 27:620–627.
44. Naschitz JE, Yeshurun D, Shahar J. Cardiogenic hepatorenal syndrome. Angiology. 1990; 41:893–900.
45. Nouel O, Henrion J, Bernuau J, De Gott C, Rueff B, Benhamou JP. Fulminant hepatic failure due to transient circulatory failure in patients with chronic heart disease. Dig Dis Sci. 1980; 25:49–52.
46. Potter JM, Hickman PE. Cardiodepressant drugs and the high mortality rate associated with ischemic hepatitis. Crit Care Med. 1992; 20:474–478.
47. Quezado ZM, Natanson C. Systemic hemodynamic abnormalities and vasopressor therapy in sepsis and septic shock. Am J Kidney Dis. 1992; 20:214–222.
48. Rashed KA, McNabb WR, Lewis RR. Ischaemic hepatitis in the elderly. Gerontology. 2002; 42:245–249.
49. Rashkin MC, Bosken C, Baughman RP. Oxygen delivery in critically ill patients. Relationship to blood lactate and survival. Chest. 1985; 87:580–584.
50. Rawson JS, Achord JL. Shock liver. South Med J. 1985; 78:1421–1425.
51. Richman SH, Delman AJ, Grob D. Alterations in indices of liver function in congestive heart failure with particular reference to serum enzymes. Am J Med. 1961; 30:211–225.
52. Ross R. Hepatic dysfunction secondary to heart failure. Am J Gastroenterol. 1981; 76:511–518.
53. Scholhot J, Shirashi T. The reaction of liver and intestinal blood flow to a general hypoxia, hypocapnia and hypercapnia in the anesthetized dog. Pfeugers Arch. 1970; 318:185–201.
54. Seeto RK, Fenn B, Rockey DC. Ischemic hepatitis: clinical presentation and pathogenesis. Am J Med. 2000; 109:109–113.
55. Seneff MG, Wagner DP, Warner RP, Zimmerman JE, Knaus WA. Hospital and 1-year survival of patients admitted to intensive care units with acute exacerbation of chronic obstructive pulmonary disease. JAMA. 1995; 244:1852–1857.
56. Shibutani K, Komatsu T, Kubal K, Sanchala V, Kumar V, Bizzarri DV. Critical level of oxygen delivery in anesthetized man. Crit Care Med. 1983; 11:640–643.
57. Shields LH, Shannon RE. Acute hepatic congestion as a factor in elevated serum glutamic oxaloacetic transaminase titers above 400 units. Am J Med Sci. 1958; 236:438–446.
58. Shoemaker WC. Relation of oxygen transport patterns to the pathophysiology and therapy of shock states. Intensive Care Med. 1987; 13:230–243.
59. Takala J, Ruokonen E. Blood flow and oxygen transport in septic shock. Clin Intensive Care. 1992; 3(Suppl):24–27.
60. Takala J. Determinants of splanchnic blood flow. Br J Anaesth. 1996; 77:50–58.
61. Tri VH, Scheinmann N, Thomas M. Les insuffisances hepatiques suraigues d’origine cardiaque. Ann Med Interne. 1977; 128:965–968.
62. Van Riel JM, Kerremans AL. Ischemische hepatitis. Ned Tijdschr Geneesk. 1992; 136:2078–2081.
63. Viallet M, Pourrot M, Andre-Fouet X, Gayet C. Foie cardiaque aigu. Problemes diagnostiques. Nouv Presse Med. 1981; 10:3549–3551.
64. Vincent JL. Cardiovascular alterations in septic shock. J Antimicrob Chemother. 1998; 41(Suppl A):9–15.
65. Wagner GR. Asbestosis and silicosis. Lancet. 1997; 349:1311–1315.
66. Whitehead MW, Hawkes ND, Hainsworth I, Kingham JGC. A prospective study of the causes of notably raised aspartate aminotransferase of liver origin. Gut. 1999; 45:129–133.
67. Yasuda H, Fujino H, Tagawa K, Unuma T. Severe liver damage in acute circulatory failure. A clinical study of 40 cases. Nippon Naika Gakkai Zasshi. 1989; 78:629–634.
© 2003 Lippincott Williams & Wilkins, Inc.
68. Zhang H, Vincent J-L. Oxygen extraction is altered by endotoxin during tamponade-induced stagnant hypoxia in the dog. Circ Shock. 1993; 40:168–176.