In vitro steroid resistance correlates with outcome in severe alcoholic hepatitis


  • Research conducted using a grant from the Medical Council on Alcohol (MCA) and the David Telling Charitable Trust

  • Potential conflict of interest: Dr. Dayan received grants from Cerimon Pharmaceuticals.


Steroids improve the outcome in alcoholic hepatitis (AH), but up to 40% of patients fail to respond adequately. Interleukin-2 (IL-2) exacerbates steroid resistance in vitro. We performed a prospective study to determine if intrinsic steroid sensitivity correlates with response to steroids in individuals with severe AH and if IL-2 receptor blockade can reverse this. Peripheral blood mononuclear cells (PBMCs) were isolated from 20 patients with AH and a Maddrey's score >32. Patients were treated with oral prednisolone plus full supportive measures. Clinical resistance to oral steroid treatment was defined as a drop in serum bilirubin of <25% within 7 days or death within 6 months. In vitro steroid resistance was measured in PBMC using the dexamethasone suppression of lymphocyte proliferation assay and repeated after the addition of the anti-IL-2 receptor (anti-CD25) monoclonal antibody, basiliximab. Suppression of lymphocyte proliferation <60% was considered to indicate steroid resistance. In all, 82% (9/11) of in vitro steroid-resistant patients were dead at 6 months as compared to 21% (2/9) of steroid-sensitive patients (P = 0.03). Similarly, 91% (10/11) of in vitro steroid-resistant patients failed to show a significant fall in bilirubin at day 7 as compared to 44% (4/9) of steroid-sensitive patients (P < 0.05). Basiliximab improved the maximal proliferation count in 91% (10/11) of in vitro steroid-resistant patients (P = 0.003). Conclusion: Clinical outcome of steroid therapy in this patient cohort correlated with in vitro steroid resistance. IL-2 blockade improved in vitro steroid sensitivity. This suggests that intrinsic lack of steroid sensitivity may contribute to poor clinical response to steroids in severe AH. IL-2 receptor blockade represents a possible mechanism to overcome this. (HEPATOLOGY 2011;)

Alcohol-related liver disease is placing an increasing burden on health service resources. The overall mortality in patients with alcoholic hepatitis (AH) is 15%, but this rises to 50% in those patients categorized as having severe disease.1

The diagnosis of AH is determined clinically but mathematical derivations are used to score the severity of the condition, aid treatment decisions, and act as prognostic tools. A score of over 32 in Maddrey's discriminant function (MdF) in a patient with clinical AH is considered indicative of severe disease (severe alcoholic hepatitis, SAH) and is often used as a threshold for the commencement of steroids in these patients. Other prognostic markers such as the Glasgow Alcoholic Hepatitis Score (GAHS) and Lille scores are now widely used, and validated, as alternative prognostic markers.2-5.

Alcoholic hepatitis has been established as an important precursor to the formation of cirrhosis.6 Evidence of a cytotoxic T-cell response playing an important role in the development of AH7 supports the use of steroid therapy as an appropriate treatment choice in this patient group to dampen hepatic inflammation. Indeed, high-dose steroid therapy is currently the only pharmacological intervention that has been shown to improve outcome in SAH, and is especially effective in patients with encephalopathy.8 However, using an “early change in bilirubin level (ECBL),” defined as a serum bilirubin level at 7 days lower than the bilirubin level on the first day of treatment, it has been reported that 27%-40% of patients with SAH fail to respond to steroid treatment.9 Alternative drugs such as pentoxifilline (a phosphodiesterase inhibitor) and theophylline have failed to show any benefit in vivo when used in patients unresponsive to steroid treatment.10 Theophylline has recently been shown to enhance steroid suppression of lymphocytes in vitro in SAH.11 However, its use in vivo in steroid-resistant SAH has not been investigated. There is therefore an urgent need for treatment modalities able to improve the response to steroids in SAH.

Failure to respond adequately to steroids is not confined to SAH. Steroid resistance rates of around 30% are reported across a variety of inflammatory diseases including asthma,12, 13 inflammatory bowel disease,13, 14 and rheumatoid arthritis.15 Our group, and others, have shown that measurement of the ability of steroids (dexamethasone) to suppress lymphocyte proliferation in vitro (the dexamethasone suppression of lymphocyte proliferation test, DILPA) correlates with the response to steroids in vivo in severe asthma and inflammatory bowel disease.13, 16 We have also shown that in vitro lymphocyte steroid resistance is detectable in up to 30% of healthy individuals and that this parameter is stable over time,17 suggesting that steroid resistance may be an intrinsic property of certain individuals which only becomes apparent when they develop an inflammatory disease that requires pharmacological treatment with steroids.

The mechanism(s) underlying this common form of lymphocyte steroid resistance is unknown, although several ideas arising from in vitro studies have been proposed18-20 and in the clinical setting, mediators of inflammation may exacerbate this trait.20 Interleukin 2 (IL-2), a key growth factor secreted by T cells, has been shown to antagonize the response to steroids in vitro, reducing the degree of lymphocyte suppression when cells are cultured in the presence of high-dose dexamethasone.21 T cells expressing higher levels of the high affinity IL-2 receptor, CD25, demonstrate steroid resistance.22 Furthermore, inhibition of IL-2, either by cytokine neutralization or receptor blockade, has been shown to enhance sensitivity to steroids in vitro,16 raising the possibility that in vivo blockade of the IL-2 pathway might represent a treatment strategy in steroid-resistant patients. Two monoclonal antibodies targeting CD25 are currently available (and used as part of immunosuppressive regimes for transplantation), one chimeric (basiliximab), and one humanized (daclizumab).

We hypothesize that, as seen in other inflammatory diseases, intrinsic resistance to steroids (indicated by testing in vitro the percentage of lymphocyte suppression in the presence of high-dose dexamethasone) may play a role in individuals who fail to respond to steroids in vivo in SAH. If confirmed, this raises the possibility that IL-2 receptor blockade might be able to reverse this. We report here a prospective study to test these hypotheses in which in vitro lymphocyte steroid resistance and the effects of CD25 blockade (with basiliximab) in vitro were measured in a consecutive series of patients presenting to our unit with SAH (MdF >32). In vitro steroid response values were measured on admission and then compared to the clinical response to steroids in vivo, as indicated both by the early biochemical response (drop in bilirubin >25% in the first 7 days of treatment) and 6-month mortality rates.


AH, alcoholic hepatitis; DILPA, dexamethasone suppression of lymphocyte proliferation test; GAHS, Glasgow Alcoholic Hepatitis Score; Imax, maximal proliferation count; MdF, Maddrey's discriminant function; PBMC, peripheral blood mononuclear cells; PHA, phytohemagglutinin; SAH, severe alcoholic hepatitis.

Patients and Methods


Consecutive patients aged 18-75 presenting to our unit with decompensated chronic liver disease, an MdF >32, and a history of excess alcohol consumption were recruited to the study. Those with overt signs of sepsis or serum creatinine levels >400 μmol/L were excluded. Other exclusion criteria were evidence of nonalcohol-related liver disease, current or recent treatment (in the last 3 months) with oral or intravenous steroids or other immunosuppressants, documented human immunodeficiency virus (HIV) infection, or any history of autoimmune disease. In cases where patients were too confused or encephalopathic to give consent, a relative was asked to sign an assent form to allow entry into the study. Phlebotomy was performed within 48 hours of starting steroids. Patient demographics (age, sex, alcohol history) were documented as well as serum biochemistry results taken on days 0 and 7. GAHS and Lille model prognostic scores were calculated as described.2, 23, 24 All patients were treated daily with 40 mg prednisolone orally for a minimum of 10 days and full supportive care. All patients either had undergone liver biopsy within the 6 months prior to inclusion or underwent biopsy during the current hospital admission to exclude alternative causes of liver disease.

Outcome Measures.

Primary outcome was mortality at 6 months. Fall in bilirubin in the first 7 days following treatment was a secondary outcome measure. A suppression of lymphocyte proliferation of <60% of the maximal proliferation count (Imax) was used as a criterion of in vitro steroid resistance as described.16, 17 Imax = 1 − (cpm with dexamethasone − cpm with phytohemagglutinin [PHA] alone) × 100% (cpm = count per million).


In all, 20-40 mL of blood was taken from each patient within 48 hours of starting steroid therapy. PBMCs were isolated by Ficoll-paque Plus density gradient centrifugation of heparinized venous blood and cell viability assessed by the Trypan blue dye exclusion test.

A total of 4 × 105 PBMCs were resuspended in RPMI 1640 media solution (Invitrogen) and cultured in triplicate in a round-bottom, 96-well plate containing 10% heat-inactivated fetal calf serum and 20 μg/mL PHA as described.16, 17 To study the effects of IL-2 blockade on steroid resistance, 10 μg/mL final concentration of basiliximab (Simulect, Novartis) was added to a triplicate of cultures at baseline. Cells were cultured in the presence or absence of dexamethasone 10−6 M for 42 hours. Ten μL of 3H thymidine (Amersham International, Amersham, UK) was then added to each well and left for a further 6 hours. The plate was harvested onto a glass fiber filter paper (Wallac Oy, Turku, Finland) using a cell harvester apparatus (Tomtec, Orange, CT) and the incorporated radiolabel was counted using a Micro β emission scintillation counter (Wallac) expressing triplicate culture data as counts per minute. In all but five individuals tritiated thymidine incorporation after PHA stimulation alone was >10,000 cpm. Where the induction of proliferation was inadequate (cpm with PHA alone <10,000), the assay was repeated within 3 months and the data included in the analysis if on repeat testing the value was >10,000 (two individuals). Three individuals were excluded from the analysis because of repeated failure of the proliferation assay. Of these, one-third died within 6 months. There was no difference between the median proliferated cell counts in either the steroid-resistant or steroid-sensitive groups (P = 0.84).

Data Analysis.

Comparison of baseline differences and treatment response in the steroid-resistant and steroid-sensitive groups as defined by in vitro and in vivo measurements were conducted using nonparametric statistics (Mann-Whitney U test or Wilcoxon signed rank test). Proportional analysis of percentage drop in serum bilirubin versus Imax was measured using Fisher's exact test.


Baseline Demographics, Clinical Response to Steroids, and Outcome.

In all, 20 patients were recruited into the study over 18 months (14 men and 6 women were recruited). All patients were required to have an MdF score of >32 as defined by the study inclusion criteria. There was no correlation between baseline bilirubin, MdF, Lille score, or GAHS and mortality at 6 months (P = 0.45, P = 0.54; P = 0.70, and P = 0.97, respectively; Fig. 1) in this cohort of SAH.

Figure 1.

There were no significant differences between MdF, Lille score, GAHS, and baseline bilirubin levels (day 0) when patients were grouped depending on outcome defined as alive or dead at 6 months after onset of steroid treatment for SAH (P = 0.54, 0.70, 0.96, 0.45, respectively). Clinical response to steroids, as defined by a drop in serum bilirubin levels over 7 days of treatment with oral prednisolone, was significantly associated with a better outcome (P = 0.005).

A drop in serum levels of bilirubin in the first 7 days of treatment with steroids (clinical steroid sensitivity) has been shown to correlate with outcome in SAH.25 Consistent with this, we saw a strong correlation between this parameter and 6-month mortality in our patient cohort (Fig. 1), indicating that our cohort is similar to those previously studied in SAH.

In Vitro Measurement of Steroid Resistance and Outcome.

Subjects were categorized as steroid-resistant by in vitro criteria (Imax <60%). The overall prevalence of clinical steroid resistance in this patient cohort was high (68%)—higher than the values seen in other inflammatory conditions and the in vitro steroid resistance seen in the general population (about 30%).13, 15, 26, 27

No statistical differences in baseline MdF, Lille score, GAHS, or baseline (day 0) bilirubin were seen between the in vitro steroid-resistant and steroid-sensitive groups (Fig. 2).

Figure 2.

Graphical representation of in vitro steroid response (Imax) versus prognostic scores MdF (Maddrey), GAHS, and Lille score. There was no significant difference in prognostic scores between in vitro steroid-resistant and steroid-sensitive groups (P = 0.32, P = 0.56, P = 0.94, respectively). Median baseline (day 0) bilirubin (mmol/L) was similar in both groups (P = 0.88).

However, in vitro steroid resistance, as indicated by Imax <60%, was significantly associated with outcome in response to steroid therapy as determined by mortality at 6 months (P = 0.03) (Fig. 3). 82% (9/11) of in vitro steroid-resistant patients were dead at 6 months as compared to 21% (2/9) of steroid-sensitive patients (P = 0.03). In patients who survived for 6 months following their treatment, only 2 of 11 (21%) had an Imax value of lower than 60%. Consistent with this finding, patients who had a serum bilirubin fall of < 25% in the first 7 days of steroid treatment also had a lower Imax (Fig. 4). 91% (10/11) of in vitro steroid-resistant patients failed to show a significant fall in bilirubin at day 7 as compared to 44% (4/9) of steroid-sensitive patients (P < 0.05).

Figure 3.

Imax scores are significantly higher (P = 0.03) in those patients surviving 6 months or more after steroid treatment for SAH.

Figure 4.

In vitro steroid sensitivity (Imax) was significantly higher in those patients demonstrating clinical steroid sensitivity as defined by percentage drop in serum bilirubin (Br) (mmol/L) over 7 days of steroid treatment (P = < 0.05).

IL-2 Inhibition and In Vitro Steroid Resistance.

In those patients demonstrating in vitro steroid resistance as measured by an Imax value <60% (n = 11), competitive inhibition of IL-2 at the high-affinity CD25 receptor with 10 μg/mL basiliximab improved lymphocyte suppression in the presence of high-dose dexamethasone, P = 0.002 (Fig. 5). Basiliximab improved Imax in 91% (10/11) of in vitro steroid-resistant patients (P = 0.002).

Figure 5.

Basiliximab (10 μg/mL) was added to cell cultures in 18/20 patients. The median Imax value in the presence of dexamethasone alone was 13.20%. The addition of basiliximab to culture increased the median Imax value to 57.45%. The improvement in lymphocyte suppression in cells cultured with the addition of basiliximab was statistically significant (P = 0.0001).


We have shown here that in a cohort of patients with SAH (MdF/Maddrey score >32 at baseline) treated with a standard steroid regime, clinical outcome (survival at 6 months) correlates with an in vitro measure of lymphocyte steroid resistance (DILPA). We have previously provided evidence that steroid resistance is, at least in part, an intrinsic property in a given individual, independent of the presence of disease. Resistance to steroids only becomes apparent when the individual develops a disease that requires steroid pharmacotherapy.28 Our data therefore suggests that such intrinsic steroid resistance plays a significant role in the failure to respond to steroid therapy in SAH. This finding is consistent with previous reports in other conditions, such as ulcerative colitis,13, 16, 29 asthma,12 rheumatoid arthritis,15 and a very recent report in AH.11 In this study, early bilirubin change correlated with in vitro steroid sensitivity (Imax). Our study adds to this report by demonstrating that in vitro steroid resistance also correlates with a hard clinical endpoint, namely, death. This was possible in our study as we had more patients (n = 20 versus n = 12) and more deaths (11 versus 4), and we would anticipate that extension of the study of Kendrick et al.11 to a further follow-up and/or a larger cohort would lead to similar conclusions.

No correlation was seen between measures of baseline disease severity (MdF score, Glasgow score, Lille score) and outcome in response to steroids in this cohort (Fig. 1). Although some previous studies relating baseline disease severity in AH to clinical response have shown a correlation with outcome, these studies have included all grades of disease severity and not specifically correlated outcome in the most severe group treated with steroids. The lack of correlation with disease severity in our cohort emphasizes that the failure to respond adequately to steroids in some individuals is not simply explained by differences in baseline disease severity and that the role played by intrinsic steroid resistance in determining outcome is independent of disease severity. Consistent with previous reports,1, 25, 30-37 we did observe in our cohort a correlation between fall in bilirubin by day 7 (a measure of early response rather than disease severity) and long-term outcome (Fig. 1).

The separation in outcome between individuals determined to be steroid-resistant or steroid-sensitive at baseline was not complete (Fig. 3). Hence, measurement of in vitro steroid sensitivity should not be considered a robust predictive marker for use in clinical management. Rather, the present finding provides evidence for an important factor that contributes to outcome, and which might represent a target for pharmacotherapy to improve overall outcomes. In this context, we noted that that addition of basiliximab to in vitro cell cultures, competitively targeting CD25, a key component of the high-affinity IL-2 receptor,38 improved steroid sensitivity in all individuals with low Imax on the DILPA test, consistent with previous reports in ulcerative colitis.16, 29, 39 The mechanisms involved in steroid resistance are unknown but IL-2 may play an important immunological role. Combination of IL-2 and IL-4 has been shown to reduce glucocorticoid receptor-binding affinity and consequent T-cell response to steroids.40 In murine and human studies using T-cell receptor-activated cell culture assays similar to the PHA cell culture in this study, IL-2 has been shown to induce steroid resistance in vitro41, 42 and data from our laboratory has demonstrated that blockade of IL-2 signaling with basiliximab can increase in vitro steroid sensitivity in PHA-stimulated PBMCs.16, 29 The data presented here suggest that adjunctive CD25 blockade might be expected to improve outcomes in steroid-resistant AH but caution is required before translating this finding into the in vivo setting. However, there is clearly a need for new intervention strategies. In patients with AH, immunomodulators other than steroids have not been successful at improving outcome; a trial of high-dose infliximab (anti-tumor necrosis factor [TNF]) at 10 mg/kg was stopped early due to increased mortality in the treatment group43 and Etanacept44 has also been proven to be ineffective at enhancing immunosuppressive treatment and leads to a poorer outcome. Sharma et al.45 have recently reported improved MdF at 28 days in patients with SAH receiving one dose (5 mg/kg) of infliximab as monotherapy. In this particular study, a reduction in serum bilirubin at day 7 was significantly associated with a better outcome. However, even in the absence of steroid use in this study, the immunosuppressive profile of infliximab alone may inhibit its clinical use in AH. Overall, five patients in the study (26%) developed infection. Three patients recovered with treatment but two patients (10%) died (one with pneumonia leading to sepsis and the other of disseminated tuberculosis).

The prospective study design, inclusion of consecutive cases, biopsy confirmation of the diagnosis, complete follow-up of all cases to 6 months, and the use of an objective primary outcome measure (survival at 6 months) represent strengths of the current study. In all cases the measurement of steroid resistance was performed before the clinical outcome was known. Potential weaknesses include the lack of a strictly controlled treatment regime, but all subjects were treated at a single center where a standard treatment protocol exists, and the managing clinicians were unaware of the results of the steroid sensitivity measurement results.

The overall mortality rate in the present cohort was high—around 50% at 6 months. However, it should be noted that many of these individuals survived their inpatient treatment (2/11; 18%) but died later of complications of decompensated liver disease either at home or during a subsequent hospital admission. A recent review of mortality in AH showed an overall mortality rate of 34.19%, with a median observation time of 160 days (range, 21-720). The three most common causes of death were hepatic failure, gastrointestinal bleeding, and infection.46 Rates of intrinsic (in vitro) steroid resistance within our cohort were also high, at 55% (Imax <60%), which contrasts with previous series rates of 25%-30% in other diseases.13-15 This may reflect an enrichment of steroid resistance within this cohort either by predisposing to the development of severe disease or due to a contribution from the presence of AH.

In conclusion, we provide the first evidence that intrinsic lymphocyte steroid resistance is an important factor in determining the outcome of steroid therapy in SAH. This element, and the possible role of IL-2 signaling in determining steroid resistance, should be taken into account in developing strategies to improve outcome with steroid therapy in this condition.