Journal of Critical Care 41 (2017) 9–15

Contents lists available at ScienceDirect

Journal of Critical Care

journal homepage: www.jccjournal.org

Clinical Potpourri

Hypoxic hepatitis — its biochemical profile, causes and risk factors of

mortality in critically-ill patients: A cohort study of 565 patients☆
Mohammed M. Aboelsoud, MD a,⁎, Amen I. Javaid, MD b, Mazen O. Al-Qadi, MD c, James H. Lewis, MD d
a
Department of Medicine, Memorial Hospital of Rhode Island/Alpert Medical School of Brown University, 11l1 Brewster St, Pawtucket, RI 02860, USA
b
Department of Medicine, Georgetown University Hospital, 3800 Reservoir Rd NW, Washington, DC 20007, USA
c
Department of Medicine, Yale-New Haven Hospital, 20 York St, New Haven, CT 06510, USA
d
Division of Gastroenterology and Hepatology, Georgetown University Hospital, 3800 Reservoir Rd NW, Washington, DC 20007, USA

a r t i c l e i n f o a b s t r a c t

Keywords: Purpose: A retrospective analysis of critically-ill patients with hypoxic hepatitis (HH) to characterize the
Ischemic hepatitis biochemical profile and to identify predictors of mortality using the Medical Information Mart for Intensive
Shock liver Care III database.
MIMIC-III Methods: HH was defined as a rapid increase in AST/ALT ≥ 800 IU/L after exclusion of other causes. We investigated
Outcomes the correlation between various clinical and laboratory parameters and mortality rates using regression models.
Results: Among 38,645 ICU-patients, 565 (1.46%) were diagnosed with HH; 57.9% were males; median age was
63 years. The unique biochemical profile of HH was confirmed; lactate dehydrogenase (LDH) was higher than
both ALT and AST; AST N ALT for the first 2 days then the ratio is reversed until recovery. All-cause hospital mortality
was 44.1%. All-cause hospital mortality was 44.1%. On multivariate analysis, older age, higher SAPS-II, higher INR,
higher bilirubin, higher LDH, acute kidney injury (AKI), and the need for vasopressors were independently associ-
ated with mortality.
Conclusion: Older age, higher SAPS-II, LDH, INR and bilirubin levels, concomitant AKI and the need for vasopressors
were all factors associated with increased mortality. The diagnosis of HH was an important harbinger of mortality in
this population, which appears to be driven mainly by the severity of the underlying conditions.
© 2017 Elsevier Inc. All rights reserved.

1. Background
Hypoxic hepatitis (HH) is a form of acute liver injury characterized
by rapid, often dramatic, increases in serum aminotransferases, which
can reach into the thousands or tens of thousands, associated with hy-
povolemic, cardiogenic or septic shock or other serious states of hemo-
dynamic instability. The diagnosis should be considered in any of these
settings after the exclusion of drug-induced liver injury, acute viral hep-
atitis, or other causes of acute hepatotoxicity [1-5]. HH is commonly
known as “ischemic hepatitis” or “shock liver”; however, Henrion et
al. proposed the term “hypoxic hepatitis” after demonstrating that a
true shock state was absent in half of their cases when studied hemody-
namically [2]. These investigators suggested that the terms “ischemic
hepatitis” or “shock liver” were in fact, often misnomers [2,6,7]. Histo-
logically, hypoxic hepatitis is characterized by centrilobular liver cell
☆ All authors approved the final version of this manuscript.
⁎ Corresponding author at: Department of Medicine, Memorial Hospital of Rhode
Island, 111 Brewster St, Pawtucket, RI 02860, USA.
E-mail addresses: Mohammed_Aboelsoud@brown.edu (M.M. Aboelsoud),
Amenismail88@gmail.com (A.I. Javaid), Mazen.al-qadi@ynhh.org (M.O. Al-Qadi),
Lewisjh@gunet.georgetown.edu (J.H. Lewis).
necrosis, but liver biopsy is rarely needed to make the diagnosis [5,8,
9]. Although HH may contribute to many cases in which elevated AST
or ALT are observed in an ICU setting [4,10], it is likely often missed,
and has been the subject of many case reports describing the difficulties
encountered in making this diagnosis [11-16]. While an uncommon
condition overall, HH has been estimated to occur in about two of
every 1000 hospital admissions (0.2%), two to three of every 100 ICU ad-
missions (2.5%) and four of every 10 (40%) admissions in which the
aminotransferases are greater than ten times the upper limit of normal
[8]. However, the in-hospital mortality associated with this diagnosis
has been shown to be upwards of 50% [8].
The pathophysiology of hypoxic hepatitis is not fully understood but
different mechanisms have been proposed, including ischemia, venous
congestion, arterial hypoxemia, and inability of the liver to extract and
use oxygen [2]. Complications associated with HH have included sponta-
neous hypoglycemia, respiratory insufficiency due to hepatopulmonary
syndrome, and hyperammonemia [17]. Since there is no specific treat-
ment for HH per se, early recognition is of key importance [18]. Amassing
clinical data regarding HH in literature has been hampered by the relative
rarity of the disease, with most studies being limited by a relatively small
number of patients. The largest study of HH published to date was a Jap-
anese/US cohort reported by Birrer et al. that included 322 patients [19]

http://dx.doi.org/10.1016/j.jcrc.2017.04.040
0883-9441/© 2017 Elsevier Inc. All rights reserved.
10 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15
followed by the Austrian group at the Vienna University Hospitals, who
analyzed 295 patients [20]. Most of the HH studies have been conducted
in Europe and very few were conducted solely on a US-based population
[21]. The most recent US-based study was conducted by the Acute Liver
Failure Study Group (ALFSG) which included only 51 patients [22]. More-
over, results from previous HH studies regarding predictors of mortality
and other outcomes have not always been uniform, and have also been
limited by the number of patients. To overcome some of these limitations,
we reviewed data from the Medical Information Mart for Intensive Care
III (MIMIC-III) critical care database which includes about forty thousand
patients from a single center, and is freely available on line [23,24]. The
aims of this study were to investigate the frequency of hypoxic hepatitis
in a large cohort of ICU patients and to characterize its biochemical profile
as well as the predictors of mortality in this population.
2. Methods
2.1. Study design
We conducted a retrospective cohort study to characterize the bio-
chemical profile of hypoxic hepatitis and the predictors for in-hospital
mortality in a study population composed of critically-ill patients from
a single large medical center database.
2.2. Study population
We used the Medical Information Mart for Intensive Care III (MIMIC-
III) research database developed by researchers from the Laboratory for
Computational Physiology at Massachusetts Institute of Technology
(MIT), Cambridge, MA, USA, and the Department of Medicine at the
Beth Israel Deaconess Medical Center (BIDMC) Boston, MA, USA [23,
24]. This database contains detailed information about intensive care
unit patient stays, including high-resolution vital sign trends and wave-
forms, laboratory data, therapeutic interventions, discharge summaries,
radiology reports, and International Classification of Diseases, 9th Revi-
sion (ICD-9) codes for all patients admitted to the BIDMC ICU between
2001 and 2012 [23,24]. Patients were de-identified in a Health Insur-
ance Portability and Accountability Act (HIPAA)-compliant manner.
The institutional review boards of BIDMC and MIT approved the use of
the MIMIC-III database [23,24].
We included all adult patients (≥18 years) admitted to the medical,
cardiac or surgical ICU units at BIDMC between 2001 and 2012 who met
all of the following criteria for the diagnosis of hypoxic hepatitis: 1) AST
and ALT ≥20-times the upper limit of normal (~800 IU/L), irrespective
of the admission diagnosis; 2) daily AST and ALT values available to re-
view until their recovery or death; 3) no evidence of drug-induced,
chemical or toxic-related liver injury and acute viral hepatitis excluded
by standard serology panels. Patients with suspected acetaminophen
poisoning were excluded by virtue of positive serum acetaminophen
levels or by an appropriate overdose history on reviewing the chart.
Clinical notes and discharge summaries were reviewed to document
the reason for admission, the cause of death (if any) and to rule out
any suspected toxic-related liver injury or accidental hepatic artery liga-
tion in surgical patients. Patients who had undergone liver surgery, in-
cluding resection or orthotropic liver transplant, were also excluded.
The day of diagnosis of HH (designated as Day 0) was defined as the
first evidence of AST and/or ALT increased ≥800 IU/L. Serial ALT, AST,
lactate dehydrogenase (LDH), alkaline phosphatase, albumin, total bili-
rubin, international normalized ratio (INR), lactate, complete blood
counts, blood urea nitrogen and creatinine levels were extracted from
Day −2 to Day +10. Demographic data including age, sex, and ethnicity
were obtained for each patient. Clinical data included vital signs, fluid
intake/output, need for vasopressors, need for mechanical ventilator
support, hospital/ICU length of stays (LOS) and discharge location. Sep-
sis was identified as a documented infection and acute organ dysfunc-
tion based on ICD-9 codes as studied and validated by Angus et al.
[25]. We also used ICD-9 codes to look for other medical comorbidities,
including congestive heart failure, cardiac arrhythmias, peripheral vas-
cular disease, hypertension, chronic respiratory diseases, chronic kidney
disease, and diabetes mellitus. Acute kidney injury (AKI) was defined as
a ≥1.5 times rise in creatinine from baseline or the need for renal re-
placement therapy. Hypotension was defined as blood pressure b 90/
60 mm Hg or mean arterial pressure (MAP) b 65 mm Hg. For grading
of severity, we used the well-validated Simplified Acute Physiology
Score II (SAPS-II) score on admission to the ICU. The SAPS-II score uses
17 variables to predict the risk of death without having to specify a
primary diagnosis [26].
2.3. Statistical analysis
Age, vital signs, laboratory data, SAPS-II and LOS were defined as
continuous variables; while ethnicity, gender, sepsis, vasopressors, me-
chanical ventilation, medical comorbidities and in-hospital mortality
were defined as categorical variables. Continuous variables were report-
ed as the mean with standard deviation (SD) or the median with inter-
quartile range (IQR) when appropriate. Categorical variables are
reported as percentages. The correlation between various clinical and
laboratory data and hospital mortality rates was investigated using uni-
variate analysis. Comparisons between groups for categorical variables
were evaluated using Pearson's chi-square test for contingency, and
for continuous variables, a two-sided t-test was used. Subsequently, a
multivariate analysis was conducted using a logistic regression model.
The model included potential variables from the univariate analysis
with P b 0.05, as well as those felt to have clinical significance (even if
not reaching the pre-defined statistical significance value). In the
event of co-linearity between variables, only one variable was included.
All statistical tests and/or confidence intervals (CI) were performed
at α = 0.05. All reported P values were two sided and rounded to
three decimal places. Statistical analysis was performed using JMP
Pro by SAS Institute, NC, USA.
3. Results
3.1. Patient characteristics
Out of 38,645 adult ICU admissions at the BIDMC between 2001 and
2012, we identified 746 patients with AST and ALT ≥800 IU/L. Of these,
565 (1.46%) satisfied all of our inclusion criteria for a diagnosis of HH,
including having daily liver tests available to review. We excluded
181 patients for having either a positive serum acetaminophen
level (n = 105), serological evidence of acute viral hepatitis (n = 33)
or having missing or incomplete data (n = 43), including the absence
of daily liver biochemistries. A flow diagram of the study population is
shown in Fig. 1.
The baseline characteristics of the study population are summarized
in Table 1. The median age of the group was 63.2 years [interquartile
range (IQR) 49.2–76.1 years]; 57.9% were males; 65.1% were white, me-
dian SAPS-II score on admission was 47 [IQR 36–62.5]. The underlying
conditions for the ICU admissions are summarized in Table 2. Sepsis,
myocardial infarction, CHF, respiratory failure, arrhythmia and post-
surgical conditions were the most frequent indications for admission
in our cohort. All-cause in-hospital and 28-day mortality rates were
44.1% and 48% respectively. The causes of death are summarized in
Table 3. In many cases the cause of death was not mentioned, or patients
were placed on comfort measures only (31%), followed by septic shock
(18%) and multiple organ failure (13%). Documented hypotension prior
to the diagnosis of HH was found in only 24.4%. However, vasopressors
were used in 63.2% of patients, sepsis was identified in 56.3% of patients;
concomitant AKI developed in 69.4% and 75.2% required ventilatory
support, suggesting that the proportion with unrecorded hypotension
might have been substantially higher.
 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15 11

Table 2
Underlying conditions of 565 hypoxic hepatitis patients (n = 565).

Condition n (%)

Septic conditions (endocarditis, abdominal abscess, C-diff 61 (10.80)

colitis, urosepsis, fungemia, cellulitis, necrotizing fasciitis,
osteomyelitis, others)
Myocardial infarction 49 (8.67)
Emergent abdominal surgery (bowel obstruction/perforation, 44 (7.79)
cholecystectomy, splenectomy, nephrectomy, incarcerated
hernia, gastrectomy)
Congestive heart failure 42 (7.43)
Respiratory failure 37 (6.55)
Pneumonia 35 (6.2)
Arrhythmias (Atrial fibrillation/flutter, heart block, 30 (5.31)
ventricular tachycardia)
Cardiac surgery (CABG, valve repair) 29 (5.13)
Cardiopulmonary arrest 27 (4.78)
GI bleeding 23 (4.07)
Aortic dissection/repair 22 (3.89)
Trauma (fall, motor vehicle accident) 22 (3.89)
Cerebrovascular accident/intracranial hemorrhage 19 (3.36)
Pericardial tamponade 17 (3.01)
Electrolyte imbalance 12 (2.12)
Fig. 1. Flow diagram of the study population. Abbreviations: ICU, intensive care unit; AST Pulmonary embolism 12 (2.12)
aspartate aminotransferase; ALT, alanine aminotransferase; HH, hypoxic hepatitis. Substance abuse (alcohol, opiates, cocaine) 12 (2.12)
Acute limb ischemia 11 (1.95)
Acute renal failure 10 (1.77)
Complications of malignancy (acute leukemia, tumor lysis 6 (1.06)
3.2. Biochemical profile
syndrome, neutropenic fever)
Others 45 (7.97)
We observed that on Day 0, LDH rises first [median 2665 IU, range
Abbreviations: C-diff, clostridium difficile; CABG, Coronary artery bypass graft.
(646–12,610 IU)], followed by increases in AST [median 2384 IU,
range (866–10,308)] and ALT [median 1626 IU, range (681–6837)].
LDH was always higher than ALT and often higher than AST. INR initially
increased on Day 0 [median 2.1, range (1.4–4.5)]. Alkaline phosphatase range (1.5–2.5)] started to decline, while ALT [median 1730 IU, range
[median 101 IU, range (46–333)], albumin [median 2.9 g/dL, range (640–6077)] reached a peak but was still lower than LDH and AST.
(1.8–4.1)] and bilirubin [median 1.2 mg/dL, range (0.4–2.7)] did On Day + 2, there was the start of a rapid decline in both LDH
not show any significant changes on Day 0. [median 739 IU, range (247–5130)] and AST [median 854 IU, range
On day + 1, LDH [median 1735.5 IU, range (341–11,065)], AST (185–4739)] and a slower decline for ALT [median 1269 IU, range
[median 2063.5 IU, range (515–11,253)] and INR [median 1.8, (440–4430)] with a reversal of the initial AST N ALT ratio. In survivors,
all liver tests were observed to continue their decline with ALT becom-
ing greater than AST, and AST quickly nearing its baseline on day + 5
[median AST 77 IU, range (49–156)]. ALT declined more slowly, nearing
Table 1 its baseline after 10 days [median ALT 92 IU, range (53–156)]. INR
Baseline patient characteristics (n = 565). quickly neared baseline on day +5 [median 1.3, range (1.2–2.6)]. Alka-
Variable Value line phosphatase generally remained stable, and serum albumin
Age, median (IQR), years 63 (49.15–76.14)
showed a gradual mild decline. The time trends of these biochemical
Men, n (%) 327 (57.87) parameters are shown in Figs. 2 and 3.
Ethnicity n (%) Median total bilirubin was 1.2 mg/dL on Day 0 (range 0.4–2.7), and
White 368 (65.13) rose only to a median of 1.35 (range 0.5–3.3) on Day +1 and a similar
African American 46 (8.14)
median value and range of 1.35 (range 0.4–3.5) was recorded on Day
Hispanic 25 (4.43)
Asian 18 (3.19) + 2. On Day 0, no patient had a bilirubin N 3 mg/dL and over the next
Other 108 (19.12) 2 days, only 15 and then 39 patients, respectively, exceeded this thresh-
SAPS-II, median (IQR) 47 (36–63) old (6.9% overall for these initial 3 days). Many of these patients had
ICU LOS median (IQR), days 3.9 (1.9–9.21) sepsis, or other conditions associated with hyperbilirubinemia [27,28].
Hosp LOS median (IQR), days 9.51 (3.98–18.79)
Hospital mortality n (%) 249 (44.1)
28 days mortality n (%) 271 (48)
Sepsis, n (%) 318 (56.28)
Hypotension, n (%) 138 (24.42) Table 3
Vasopressors, n (%) 357 (63.18) Causes of in-hospital death in hypoxic hepatitis patients (n = 249).
Mechanical ventilation, n (%) 425 (75.22)
Cause of death n (%)
Acute kidney injury, n (%) 392 (69.38)
Hyper-bilirubinemia N 3 mg/dL, n (%) 126 (22.50) Other (not mentioned or comfort measures) 77 (30.92)
Comorbidities n (%) Septic shock 45 (18.07)
Congestive heart failure 99 (17.52) Multiple organ failure 33 (13.25)
Arrhythmias 124 (21.95) Respiratory failure 29 (11.65)
Peripheral vascular disease 76 (13.45) Cardiogenic shock 25 (10.04)
Hypertension 62 (10.97) Anoxic brain injury 14 (5.62)
Chronic respiratory disease 102 (18.05) Fatal arrhythmia 13 (5.22)
Chronic kidney disease 86 (15.22) Disseminated intravascular coagulation 5 (2.01)
Diabetes Mellitus 171 (30.27) Acute renal failure 3 (1.20)
Pulmonary embolism 3 (1.20)
Abbreviations: IQR, interquartile range; SAPS-II, Simplified Acute Physiology Score-II; LOS,
Hemorrhagic shock 2 (0.80)
length of stay.
12 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15
3.3. Predictors of In-hospital mortality
On univariate analysis, we found associations between in-hospital
morality and older age (P = 0.0001), a higher SAPS-II score on admis-
sion (P = 0.0001), lower albumin (P = 0.0001), higher AST (P =
0.0001), higher LDH (P = 0.0001), higher INR (P = 0.0001) and higher
bilirubin (P = 0.011). The correlation between in-hospital mortality and
various clinical variables is summarized in Table 4. Increased mortality
rates were observed among patients with (vs. without) sepsis [52% vs.
34% (P = 0.0001)], AKI [56% vs. 17% (P = 0.0001)], CHF [56% vs. 42%
(P = 0.011)], HTN [66% vs. 41% (P = 0.0002)], CKD [61% vs. 41% (P =
0.001)], and chronic respiratory disease [52% vs. 42% (P = 0.048)].
Also, patients who required vasopressors [59% vs. 18% (P = 0.0001)]
or mechanical ventilation [52% vs. 19% (P = 0.0001)] had higher mortal-
ity rates compared to those who did not.
On multivariate analysis, only age, SAPS-II, AKI, peak bilirubin, peak
LDH, peak INR and the need for vasopressors were independently asso-
ciated with a higher mortality (Table 5). For every 5-year increase in pa-
tient age, a 19% increase in mortality [OR 1.19, CI 1.1–1.29, (P = 0.0001)]
was seen. Higher SAPS-II scores were also associated with a 9% in-
creased mortality for every 5 point increase [OR 1.09, CI 1.01–1.19, (P
= 0.028)]. Bilirubin was associated with a 7% increased mortality rate
for every 1 mg increase from baseline [OR 1.07, CI 1.02–1.1, (P =
0.008)], which tended to occur late in the course of their illness, and
often well after AST and ALT had declined back towards normal,
representing supervening infection or as part of multisystem failure.
We also found that for every 1000 unit increase in peak LDH, there
was a 7% increase in mortality [OR 1.07, CI 1.01–1.2, (P = 0.023)]. INR
was associated with a 32% increase in mortality for every 1 point in-
crease [OR 1.32, CI 1.12–1.58, (P = 0.0007)]. AKI was associated with
a 3-fold increase in mortality [OR 2.9, CI 1.4–6.1, (P = 0.002)], and the
need for vasopressors was associated with a 2-fold increase in mortality
[OR 2.1, CI 1.1–4.1, (P = 0.026)].
4. Discussion
In this study of a large ICU population from a single center, it was
noted that while hypoxic hepatitis was able to be diagnosed in only
1.46% of patients, it was associated with a high mortality rate. Although
1.46% is similar to that recorded in other studies [1,2,19] it is lower than
the pooled incidence of 2.5% from a recent meta-analysis [8]. However,
since not all patients in the BIDMC ICU had daily liver tests, it is likely
Fig. 2. Trend of the liver enzymes in hypoxic hepatitis patients. Abbreviations: AST, aspartate aminotransferase; ALT, alanine aminotransferase; LDH, lactate dehydrogenase.
that the overall prevalence of HH in this cohort is actually higher. We se-
lected AST/ALT ≥800 IU/L (N20 times the ULN) as a cut-off for the diag-
nosis, while other studies have used values ranging from 400 to
3000 IU/L [8]. While the 800 IU level may have allowed some cases of
HH to go undiagnosed, we believe the 800 IU/L threshold is both reason-
able and adequate to rule out other disorders that can acutely increase
serum aminotransferases in ICU patients to these levels in just a matter
of hours or days [4,10]. Several other investigators have also chosen the
800 IU/L threshold as being optimal [2,5,29-33].
In this ICU population, HH occurred in association with one or more
underlying conditions potentially causing hypoxic injury (Table 2). Nu-
merous studies have reported hypoxic hepatitis can develop in virtually
any patient with hemodynamic instability, including hypovolemic, sep-
tic, or cardiac shock, post-surgical states and respiratory failure as we
observed [1,32,34-41]. Given the unique aspects of the MIMIC-III data-
base, we were able to track the daily pattern of liver enzymes in patients
with hypoxic hepatitis. In doing so, we confirmed the unique injury pat-
tern that has served as the diagnostic hallmark of HH – a rapid and
marked rise in AST N ALT within the first 24–48 h, usually accompanied
by an equally dramatic rise in LDH that exceeds both AST and ALT. Sub-
sequently, there is a rapid fall in the aminotransferase levels by N 50%
within the first 3 days, with a reversal of the AST:ALT ratio [18,34,42].
We observed that AST normalizes first with ALT lagging behind by sev-
eral days, most likely reflecting the much shorter half-life of AST (17 h)
compared to ALT (47 h) [43]. This phenomenon of a rapid decline and
reversal of the AST:ALT ratio seen with HH may be a reason why the di-
agnosis of HH is not considered in many ICU patients. Without the pre-
cise knowledge of the initial or peak values, patients in whom the ALT is
found to be higher than the AST when first recorded, are likely to have a
differential diagnosis that does not even include HH.
We observed an overall all-cause mortality rate of 44.1% in our study
population. Multiple organ failure from sepsis and other conditions, es-
pecially with respiratory and renal failure, contributed heavily to the
high mortality. Among these non-survivors, 69% (173/249) died within
10 days of the diagnosis of HH. These results are in line with other re-
ports, where HH has been associated with poor outcomes and an aver-
age mortality of 51% in the literature [8]. Upon investigating the
predictors of mortality in our study population, we found that older
age, an increased severity of the underlying disease (as indicated by a
higher SAPS-II score), the development of AKI, the need for vasopres-
sors, higher bilirubin, INR and LDH levels, were all independently asso-
ciated with an increased mortality in these patients. Our results are also
 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15 13

Fig. 3. Trend of the albumin, bilirubin and INR in hypoxic hepatitis patients. Abbreviations: INR, international normalized ratio.

consistent with most other studies that indicate that the severity of the of HH was made developed bilirubin N 3 mg/dL, but this usually oc-
underlying disease, and to lesser extent the degree of hepatic impair- curred later in the course of the ICU stay and often accompanied the
ment, can predict mortality in HH patients. For example, Fuhrmann et downward course of their underlying disease states. The inclusion of
al. showed that higher INR, the presence of septic shock and increased hyperbilirubinemia in our predictors of mortality is best considered a
severity of the underlying disease were independent risk factors of mor- reflection of these conditions.
tality [3]. In another study, also by Furhmann et al., HH was associated A summary of the overall mortality and risk factors of mortality asso-
with an increased mortality in critically-ill patients receiving vasopres- ciated with HH as reported in the literature is shown in Table 6. The dif-
sors. In that study, age, SAPS-II score, and male gender were associated ferences may be attributed in part, to different study designs, as well as
with increased mortality in HH patients who were not receiving vaso- different study populations. While older age and the severity of the un-
pressor support [30]. Higher INR, need for renal replacement therapy derlying disease as well as the degree of hepatic impairment (INR, bili-
and septic shock were demonstrated as predictors of mortality in HH rubin and LDH) are associated with an increased mortality, this might
patients by Raurich et al. [34]. Chavez-Tapia et al. observed that me- be best explained by the fact that older and sicker patients are expected
chanical ventilation was associated with a higher mortality in his to have higher mortality rates than are younger ones. Concomitant
study of HH in the cardiac ICU [32]. And in a recent study by Drolz et acute kidney injury is a common finding with HH [20,31,34]. Indeed,
al., it was demonstrated that AKI, higher INR and septic shock were all we demonstrated a prevalence of 70% of AKI among the HH patients
associated with a higher mortality rate [20]. While factors leading to in our cohort. The association between AKI and HH might be most easily
mortality vary among different studies, only our analysis appears to explained by the fact that both processes are hypoxic in nature.
have demonstrated LDH as a useful predictor of mortality. Most studies of HH have investigated only the acute comorbidities
Although jaundice has also been demonstrated to be associated with and therapeutic interventions that have been used in its management.
increased complications and mortality in HH patients [31], this is usual- Relatively few investigators have studied patients' chronic comorbidi-
ly a late occurrence and may be due to the cholestasis seen with sepsis, ties. In our study, while the univariate analysis found that CHF, CKD,
blood product transfusions, heart failure, and other causes [27,28]. In HTN and chronic respiratory diseases were all associated with increased
our series, hyperbilirubinemia did not accompany the initial rapid rise mortality, the significance of these co-morbidities was not confirmed in
in AST, ALT or LDH in the majority of patients, and is not considered a the multivariate logistic regression model when combined with other
common feature of HH. Overall, 22% of the patients in whom a diagnosis variables, suggesting that it is the acute rather than chronic comorbidi-
ties that contribute most to the mortality.
Apart from supportive care measures, specific therapeutic options
Table 4
Univariate analysis for in-hospital mortality. for HH are limited. A few studies have shown that artificial liver support
might be of benefit in selected cases [44,45], but this is not widely avail-
Variable Odds ratio (CI) P value
able. Statins have been reported to have a protective effect against HH
Age 1.21 (1.15–1.28) b0.0001⁎ [33], but data are lacking to recommend their empiric use. The mainstay
Gender 1.19 (0.86–1.67) 0.2942 of management for cases of HH remains correction and reversal of the
SAPS-II 1.32 (1.25–1.40) b0.0001⁎
Sepsis 2.03 (1.44–2.86) b0.0001⁎
underlying cause.
Vasopressors 6.47 (4.33–9.85) b0.0001⁎ To our knowledge, this study represents the largest single series of
Ventilatory support 4.84 (3.07–7.85) b0.0001⁎ HH patients to date. In a recent meta-analysis of HH studies that includ-
Acute kidney injury 6.03 (3.93–9.52) b0.0001⁎ ed 1782 patients, the largest single cohort was 322 patients [19]. We are
Congestive heart failure 1.75 (1.13–2.72) 0.011⁎
aware of the potential limitations inherent in the retrospective nature of
Arrhythmias 1.48 (0.98–2.2) 0.055
Peripheral vascular disease 1.58 (0.97–2.58) 0.0623 our study, but feel that statistical power derived from a cohort of 565
Hypertension 2.77 (1.61–4.90) 0.0002⁎
Chronic respiratory diseases 1.55 (1.01–2.38) 0.0482⁎
Chronic kidney disease 2.19 (1.38–3.52) 0.0009⁎ Table 5
Diabetes mellitus 0.95 (0.66–1.37) 0.8018 Multivariate analysis for predictors of in-hospital mortality.
AST per 100 u 1.01 (1.006–1.015) b0.0001⁎
Variable Odds ratio (CI) P value
ALT per 100 u 1.0 (0.99–1.01) 0.656
LDH per 1000 u 1.13 (1.08–1.19) 0.0001⁎ Age per 5 years 1.19 (1.1–1.29) 0.0001
Alkaline phosphatase per 100 u 0.94 (0.86–1.03) 0.169 SAPS-II per 5 points 1.09 (1.01–1.19) 0.028
Bilirubin 1.08 (1.02–1.12) 0.0107⁎ Acute kidney injury 2.9 (1.4–6.1) 0.002
Albumin 0.39 (0.28–0.53) b0.0001⁎ Vasopressors 2.1 (1.1–4.1) 0.026
Platelets 0.99 (0.99–1.01) 0.45 Bilirubin per 1 mg 1.07 (1.02–1.1) 0.008
INR 1.52 (1.32–1.78) b0.0001⁎ LDH per 1000 u 1.07 (1.01–1.2) 0.023
INR per 1 point 1.32 (1.12–1.58) 0.0007
Abbreviations: CI, confidence interval; INR, international normalized ratio; SAPS-II, simpli-
fied acute physiology score II; AST, aspartate aminotransferase; ALT, alanine aminotrans- Abbreviations: CI, confidence interval; SAPS-II, Simplified Acute Physiology Score-II; LDH,
ferase; LDH, lactate dehydrogenase; INR, international normalized ratio. lactate dehydrogenase; INR, international normalized ratio.
14 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15

Table 6
Summary of most recent hypoxic hepatitis (HH) studies.

Study Year Liver Sample Population Number of Incidence Mortality Risk factors for mortality
enzymes size HH patients % %
cutoffa

Henrion [2] Belgium 2003 N800 15,619 ICU 142 0.9 52.8 Not studied
Birrer [19] Japan/USA 2007 N400 32,209 ICU 322 1.1 45 Not studied
Raurich [34] Spain 2011 N1000 7674 ICU 182 2.4 61.5 INR, RRT, septic shock
Taylor [22] USA 2012 N1000b 1147 ALFSG 51 4.4 25 Advanced encephalopathy, high serum phosphate
Chavez-Tapia [32] Mexico 2014 N800 1649 CICU 18 1.1 55.6 Mechanical ventilation
Galvin [46] Ireland 2015 N1000 Case – 111 – 55 Not studied
series
Bjornsson [10] Iceland 2016 N500b Case – 26 – 35 Not studied
series
University of Vienna Austria 2016 N800 1948 ICU 295 15 58 AKI, INR, septic shock, SOFA N 10, vasopressors,
group [3,20,30,31,33] age, SAPS-II, male gender, jaundice
Current study USA 2016 N800 38,645 ICU 565 1.46 44.1 Age, SAPS-II, AKI, vasopressors, LDH, INR and
bilirubin

Abbreviations: ALFSG, acute liver failure study group; CICU, cardiac intensive care unit; INR, international normalized ratio; SOFA, sequential organ failure assessment score; RRT, renal
replacement therapy; SAPS-II, simplified acute physiology score II; AKI, acute kidney injury.
a
AST/ALT.
b
ALT only.

patients followed daily with liver-related and other tests allowed us to
confidently assess the biochemical profile and clinical outcomes of the
patients. We recognize that since only critically-ill patients from a single
institution using an aminotransferase cut-off of 800 IU/L were included,
we need to be circumspect about generalizing these results to a broader
patient population. It is likely that the 1.46% figure we arrived at is an
underestimate of the true prevalence of HH, as many patients meeting
the 800 IU AST/ALT threshold did not have daily liver tests and were ex-
cluded from further analysis, and we may have missed the diagnosis in
patients who were below the 800 IU threshold. Nevertheless, we feel
confident that the biochemical injury pattern we observed among
these patients confirms the unique ratios and fold-elevations of AST,
ALT, and LDH that have been used to characterize HH previously. The
rapid reversal and recovery of the AST: ALT ratio is an important re-
minder that knowledge of the initial biochemical values is essential to
ensure that HH remains in the differential diagnosis of acute liver injury
in the ICU setting, and may be even more important than being able to
document the presence of hypotension, which may be missed.
Presented in part at the American College of Gastroenterology (ACG)
meeting, October 18, 2016, Las Vegas NV and at The American Associa-
tion for the Study of Liver Diseases (AASLD) meeting, November 11,
2016, Boston MA.
Author contributions
Mohammed Aboelsoud, conception and design of the study; gener-
ation, collection, assembly, analysis and/or interpretation of data;
drafting and revision of the manuscript.
Amen Javaid, drafting of the manuscript.
Mazen Al-Qadi, conception and design of the study; drafting and
revision of the manuscript.
James H. Lewis, conception and design of the study; analysis and/or
interpretation of data; drafting and critical revision of the manuscript.
References

[1] Fuchs S, Bogomolski-Yahalom V, Paltiel O, Ackerman Z. Ischemic hepatitis: clinical

5. Conclusion
In summary, while hypoxic hepatitis is relatively uncommon, it can
be associated with a high mortality reflective of the serious underlying
conditions that it represents in a critical care setting. Recognition of
the unique biochemical liver test pattern is essential in order to make
the diagnosis of hypoxic hepatitis, and to avoid unnecessary testing
for disorders that are unlikely to be present. For example, although
not all viral hepatidities (such as hepatitis E virus) were sought or ex-
cluded in our cohort, the ultra-rapid onset and offset of AST and ALT,
their ratio whereby AST is higher initially, and in many cases, their ex-
treme elevations, are not at all typical of viral hepatitis. Finally, while
we confirmed that several patient risk factors (such as older age, higher
SAPS-II scores, prolonged INR, higher bilirubin levels, AKI and the need
for vasopressors) are risk factors associated with increased mortality in
patients who present with hypoxic hepatitis, and that higher LDH
values can be added to this list. It should be remembered that mortality
is driven primarily by the severity and course of the patient's underlying
disease that may be heralded by HH initially.
Disclosures
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
No potential conflict of interest was reported by the authors.
and laboratory observations of 34 patients. J Clin Gastroenterol 1998;26(3):183–6.
[2] Henrion J, Schapira M, Luwaert R, Colin L, Delannoy A, Heller FR. Hypoxic hepatitis:
clinical and hemodynamic study in 142 consecutive cases. Medicine 2003;82(6):
392–406.
[3] Fuhrmann V, Kneidinger N, Herkner H, Heinz G, Nikfardjam M, Bojic A, et al. Hypoxic
hepatitis: underlying conditions and risk factors for mortality in critically ill patients.
Intensive Care Med 2009;35(8):1397–405.
[4] Whitehead MW, Hawkes ND, Hainsworth I, Kingham JG. A prospective study of the
causes of notably raised aspartate aminotransferase of liver origin. Gut 1999;45(1):
129–33.
[5] Rawson JS, Achord JL. Shock liver. South Med J 1985;78(12):1421–5.
[6] Henrion J. Hypoxic hepatitis. Liver Int 2012;32(7):1039–52.
[7] 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, hemo-
dynamic study of 17 consecutive cases. Hepatology (Baltimore, Md) 1999;29(2):
427–33.
[8] Tapper EB, Sengupta N, Bonder A. The incidence and outcomes of ischemic hepatitis:
a systematic review with meta-analysis. Am J Med 2015;128(12):1314–21.
[9] Birgens HS, Henriksen J, Matzen P, Poulsen H. The shock liver. Clinical and biochem-
ical findings in patients with centrilobular liver necrosis following cardiogenic
shock. Acta Med Scand 1978;204(5):417–21.
[10] Bjornsson HK, Olafsson S, Bergmann OM, Bjornsson ES. A prospective study on the
causes of notably raised alanine aminotransferase (ALT). Scand J Gastroenterol
2016;51(5):594–600.
[11] Sampanis N, Gavriilaki E, Paschou E, Kavlakoudis G, Lysitska A, Dontsos C, et al. Is-
chemic hepatitis during hemodialysis in a patient with end-stage renal disease.
Eur J Intern Med;24:e70.
[12] Henrion J, Deltenre P, Peny MO, Dumoulin P, Laurent Y, Brenard R. Fatal hypoxic
hepatitis in a patient with hereditary hemorrhagic telangiectasia (Rendu-Osler-
Weber's disease). Acta Gastroenterol Belg 2010;73(1):61–4.
[13] Assy N, Gefen H, Schlesinger S, Hussein O. The beneficial effect of N-acetylcysteine
and ciprofloxacin therapy on the outcome of ischemic fulminant hepatic failure.
Dig Dis Sci 2007;52(12):3507–10.
 M.M. Aboelsoud et al. / Journal of Critical Care 41 (2017) 9–15
[14] Gitlin N, Serio KM. Ischemic hepatitis: widening horizons. Am J Gastroenterol 1992;
87(7):831–6.
[15] Shovman O, George J, Shoenfeld Y. Ischemic hepatitis in congestive heart failure
after an episode of hypotension. Harefuah 1997;132(7):459–60 (527).
[16] Azzopardi N, Chetcuti S, Sant J, Pocock J. Acute liver impairment in a young, healthy
athlete: hypoxic hepatitis and rhabdomyolysis following heat stroke. Case Rep
Gastroenterol 2012;6(2):563–8.
[17] Trilok G, Qing YC, Li-Jun X. Hypoxic hepatitis: a challenging diagnosis. Hepatol Int
2012;6(4):663–9.
[18] Fuhrmann V, Jager B, Zubkova A, Drolz A. Hypoxic hepatitis — epidemiology, patho-
physiology and clinical management. Wien Klin Wochenschr 2010;122(5–6):
129–39.
[19] Birrer R, Takuda Y, Takara T. Hypoxic hepatopathy: pathophysiology and prognosis.
Intern Med 2007;46(14):1063–70.
[20] Drolz A, Horvatits T, Roedl K, Rutter K, Staufer K, Haider DG, et al. Outcome and fea-
tures of acute kidney injury complicating hypoxic hepatitis at the medical intensive
care unit. Ann Intensive Care 2016;6:61.
[21] Johnson RD, O'Connor ML, Kerr RM. Extreme serum elevations of aspartate amino-
transferase. The American journal of gastroenterology 1995;90(8):1244–5.
[22] Taylor RM, Tujios S, Jinjuvadia K, Davern T, Shaikh OS, Han S, et al. Short and long-
term outcomes in patients with acute liver failure due to ischemic hepatitis. Dig
Dis Sci 2012;57(3):777–85.
[23] Saeed M, Villarroel M, Reisner AT, Clifford G, Lehman LW, Moody G, et al. Multipa-
rameter intelligent monitoring in intensive care II (MIMIC-II): a public-access inten-
sive care unit database. Crit Care Med 2011;39(5):952–60.
[24] Johnson AEW, Pollard TJ, Shen L, Lehman L-wH, Feng M, Ghassemi M, et al. MIMIC-
III, a freely accessible critical care database. Sci Data 2016;3:160035.
[25] Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky MR. Epidemi-
ology of severe sepsis in the United States: analysis of incidence, outcome, and asso-
ciated costs of care. Crit Care Med 2001;29(7):1303–10.
[26] Le Gall JR, Lemeshow S, Saulnier F. A new simplified acute physiology score (SAPS II)
based on a European/North American multicenter study. JAMA 1993;270(24):
2957–63.
[27] Bansal V, Schuchert VD. Jaundice in the intensive care unit. Surg Clin North Am
2006;86(6):1495–502.
[28] Lescot MDPDT, Karvellas MDFRCPCC, Beaussier MDPDM, Magder MDFRCPCS. Ac-
quired liver injury in the intensive care unit. Anesthesiology 2012;117(4):898–904.
[29] Seeto RK, Fenn B, Rockey DC. Ischemic hepatitis: clinical presentation and pathogen-
esis. Am J Med; 109(2):109–13.
[30] Fuhrmann V, Kneidinger N, Herkner H, Heinz G, Nikfardjam M, Bojic A, et al. Impact
of hypoxic hepatitis on mortality in the intensive care unit. Intensive Care Med
2011;37(8):1302–10.
15
[31] Jager B, Drolz A, Michl B, Schellongowski P, Bojic A, Nikfardjam M, et al. Jaundice in-
creases the rate of complications and one-year mortality in patients with hypoxic
hepatitis. Hepatology 2012;56(6):2297–304 (Baltimore, Md).
[32] Chavez-Tapia NC, Balderas-Garces BV, Meza-Meneses P, Herrera-Gomar M, Garcia-
Lopez S, Gonzalez-Chon O, et al. Hypoxic hepatitis in cardiac intensive care unit: a
study of cardiovascular risk factors, clinical course, and outcomes. Ther Clin Risk
Manag 2014;10:139–45.
[33] Drolz A, Horvatits T, Michl B, Roedl K, Schellongowski P, Holzinger U, et al. Statin
therapy is associated with reduced incidence of hypoxic hepatitis in critically
ill patients. J Hepatol 2014;60(6):1187–93.
[34] Raurich JM, Llompart-Pou JA, Ferreruela M, Colomar A, Molina M, Royo C, et al.
Hypoxic hepatitis in critically ill patients: incidence, etiology and risk factors for
mortality. J Anesth 2011;25(1):50–6.
[35] Raurich JM, Perez O, Llompart-Pou JA, Ibanez J, Ayestaran I, Perez-Barcena J.
Incidence and outcome of ischemic hepatitis complicating septic shock.
Hepatol Res 2009;39(7):700–5.
[36] Henrion J, Descamps O, Luwaert R, Schapira M, Parfonry A, Heller F. Hypoxic hepati-
tis in patients with cardiac failure: incidence in a coronary care unit and measure-
ment of hepatic blood flow. J Hepatol 1994;21(5):696–703.
[37] Seeto RK, Fenn B, Rockey DC. Ischemic hepatitis: clinical presentation and pathogenesis.
Am J Med 2000;109(2):109–13.
[38] Raman JS, Kochi K, Morimatsu H, Buxton B, Bellomo R. Severe ischemic early
liver injury after cardiac surgery. Ann Thorac Surg 2002;74(5):1601–6.
[39] Kavoliuniene A, Vaitiekiene A, Cesnaite G. Congestive hepatopathy and hypoxic hep-
atitis in heart failure: a cardiologist's point of view. Int J Cardiol 2013;166(3):554–8.
[40] Rashed KA, McNabb WR, Lewis RR. Ischaemic hepatitis in the elderly. Gerontology
2002;48(4):245–9.
[41] Zhang Y, Liu J, Yu L, Zhou N, Ding W, Zheng S, et al. Prevalence and characteristics of
hypoxic hepatitis in the largest single-centre cohort of avian influenza A(H7N9)
virus-infected patients with severe liver impairment in the intensive care unit.
Emerg Microbe Infect 2016;5:e1.
[42] Ebert EC. Hypoxic liver injury. Mayo Clin Proc 2006;81(9):1232–6.
[43] Dufour DR, Lott JA, Nolte FS, Gretch DR, Koff RS, Seeff LB. Diagnosis and monitoring
of hepatic injury. I. Performance characteristics of laboratory tests. Clin Chem 2000;
46(12):2027–49.
[44] Faybik P, Hetz H, Krenn CG, Baker A, Germann P, Berlakovich G, et al. Liver support in
fulminant liver failure after hemorrhagic shock. Wien Klin Wochenschr 2003;
115(15–16):595–8.
[45] Drolz A, Saxa R, Scherzer T, Fuhrmann V. Extracorporeal artificial liver support in
hypoxic liver injury. Liver Int 2011;31(Suppl. 3):19–23.
[46] Galvin Z, McDonough A, Ryan J, Stewart S. Blood alanine aminotransferase levels
N1,000 IU/l — causes and outcomes. Clin Med (Lond) 2015;15(3):244–7.