Patient factors in responders and non-responders treated with steroids for acute alcohol-associated hepatitis

Abstract

BACKGROUND

Steroids remain the primary treatment for severe alcohol-associated hepatitis (AAH), though there is little available tools to predict patient response to steroids. It was hypothesized that phosphatidylethanol (PEth) value will inversely correlate with response to steroid therapy based on Lille score in AAH.

AIM

To assess the relationship of patient factors, focusing on pre-steroid therapy PEth value, to steroid therapy response in AAH.

METHODS

A retrospective case control study was performed on patients who received ≥ 4 days of steroid therapy for AAH at our hospital between July 1, 2019 and June 30, 2022. A total of 2087 patients were screened for AAH and those treated with steroids were included for statistical analysis utilizing independent sample t-test and for categorical variables using the χ² test.

RESULTS

No correlation was found between PEth value, pre-steroids abstinence length, or number of drinks per week pre-steroids and response to steroids. Non-responder status significantly correlated with older age (P = 0.024), lower albumin (P = 0.003), and higher bilirubin (P = 0.010) pre-steroids. Our study suggests that age, pre-steroid albumin, and pre-steroid bilirubin levels may predict nonresponse to steroid therapy. Non-responders have increased incidence of death and higher medical costs.

CONCLUSION

Identifying non-responders through these identified factors should prompt early referral for liver transplantation. Future prospective studies with larger population size are needed to assess the efficacy of combined pre-steroid age, albumin, bilirubin and other biochemical markers as predictors of steroid response.

Key Words: Acute alcohol-associated hepatitis; Steroids; Transplant; Steroid response; Phosphatidylethanol

Core Tip: Steroid therapy for alcohol-associated hepatitis (AAH) remains the primary therapy despite poor predictability of efficacy. Phosphatidylethanol (PEth) value and other biomarkers were analyzed as predictors of steroid therapy response in the treatment of AAH. PEth value was a poor predictor of steroid therapy response, though pre-steroid biochemical markers of liver response to alcohol consumption significantly correlated with steroid response. Identification of non-responders to steroids can direct these patients to earlier enrollment in liver transplantation and avoid unnecessary medical cost.

INTRODUCTION

Alcohol-associated hepatitis (AAH) presents as an acute-on-chronic liver failure in alcohol-associated liver disease[1]. Without intervention, AAH poses significant risk of mortality, estimated at 30%-60% by the American Association for the Study of Liver Diseases[1]. Current treatment of AAH is limited to steroid (typically prednisolone) therapy, typically administered at 40 mg/day over 28 days[2-4]. According to the largest multicenter clinical trial to date, the Steroids or Pentoxifylline for Alcoholic Hepatitis study[2], patients with AAH treated with prednisolone experience reduced mortality compared to placebo for 28 days with no reduced mortality at 90 days or 1 year.

However, steroid therapy can pose multiple risks to patients, including elevation of blood glucose, gastrointestinal bleeding, and exacerbation of infections[3,4]. Additionally, several contra-indications for steroid therapy exist, hence only about half of all AAH cases are suitable candidates for steroid therapy[3]. Even among suitable patients, responses to steroid therapy is approximately 50%[3,4]. The Lille score, which incorporates age, prothrombin time, albumin, creatinine, and change in bilirubin at day 4 or day 7 is used to monitor the response to steroid therapy at 4 days and/or 7 days[3-6]. To date, no reliable tool has been found to predict patient response to steroid therapy for AAH prior to initiation of steroids[1,7].

Phosphatidylethanol (PEth), a biomarker obtained from dried blood, objectively estimates alcohol consumption directly correlating with blood ethanol concentration[8-12]. PEth can reliably detect alcohol consumption up to 3-4 weeks following the cessation of alcohol consumption[12]. Factors that can predict response to steroid therapy in AAH could reduce medical costs and minimize wait time to initiate alternative treatment modalities such as liver transplantation for non-responders. A case control study aimed to assess the relationship of patient factors, focusing on pre-steroid therapy PEth value, to steroid therapy response in AAH. It was hypothesized that PEth level will inversely correlate with response to steroid therapy based on Lille score in AAH.

MATERIALS AND METHODS

Study design

A case control study was conducted using electronic medical records from our institution after obtaining approval from the institutional review board at the University of South Florida. Patient informed consent was exempt by our institutional review board. We reviewed medical records of consecutive adult patients who underwent a hepatology consult and were diagnosed with AAH between July 1, 2019 and June 30, 2022. The objective of this study was to assess the relationship of pre-steroid patient factors, focusing on PEth levels, in response to steroid therapy in patients with AAH.

Study population

The study included adult patients, aged 18 and older, with a confirmed diagnosis of AAH as defined by a Maddrey’s Discriminant Factor (MDF) greater than or equal to 32 (n = 2087). To be eligible for our study, patients with AAH were required to have received steroid therapy for at least 4 consecutive days. Patients were excluded from our study if they had an MDF score less than 32, received less than 4 consecutive days of steroid therapy, or if they were below the age of 18 at the time of steroid treatment.

Data collection

Our study divided our sample based on their response to steroid therapy, which was assessed using the Lille score. The Lille score was calculated based on the collection of the following clinical lab values: (1) Albumin at patient admission; (2) Bilirubin at patient admission; (3) Bilirubin day 4 or day 7 of steroid treatment; (4) Creatinine at patient admission; and (5) Prothrombin time at patient admission[3-6].

Patients were then categorized as responders if their Lille score was less than 0.45 or non-responders if their Lille score was greater than or equal to 0.45. Outcome measures included the patient's disposition within 6 months following steroid therapy, which was categorized as successful treatment, relapse to alcohol, referral to hospice, liver transplantation, or death. Relapse was defined as a positive PEth test (PEth value greater than 20) or self-reporting alcohol use within 6 months following steroid therapy, and successful treatment was defined as discharge from the hospital and the absence of relapse, referral to hospice, liver transplantation, or death for 6 months following steroid therapy.

Patient demographics (age, gender, and ethnicity) and pre-treatment information including PEth levels, days of alcohol abstinence prior to steroid therapy, and MDF were collected. Pre-steroid albumin, pre-steroid bilirubin, pre-steroid creatinine, and pre-steroid prothrombin time were collected along with bilirubin values on day 4 and/or day 7 of steroid treatment. The number of standardized alcoholic drinks per week before the initiation of steroid therapy was also collected.

Statistical analysis

De-identified data were analyzed by a biostatistician from the University of South Florida Division of Evidence Based Medicine using Statistical Package for the Social Sciences 25 (IBM, Armonk, NY, United States). Differences between responders and non-responders for continuous variables were assessed using the independent sample t-test and for categorical variables using the χ² test. P values were deemed statistically significant if they were less than 0.05.

RESULTS

A total of 2087 patients screened with AAH, 64 patients were identified (52% male, 48% female), predominantly white (72%) who were treated with steroid therapy. A total of 64% of patients were responders to steroid therapy. PEth was performed prior to steroid therapy in 52 of the 64 patients previously identified. No correlation existed between pre-steroid PEth value, abstinence length, or standardized drinks consumed in one week and response to steroid therapy (Table 1).

Table 1 Pre-steroid patient factors and response to steroid, n (%).

Pre-steroid factors	Nonresponse (n = 36)1	Response (n = 16)1	P value
PEth (continuous) (ng/mL)	612	398	0.210
PEth (> 20 ng/mL)2	16 (31)	36 (69)	0.456
Length of stay (days)5	49	25	0.123
Standard drinks/week	25	52	0.136
	Nonresponse (n = 23)3	Response (n = 41)3
Age (years)	49	42	0.024
Albumin (g/dL)	2.1	2.6	0.003
Bilirubin (mg/dL)	24.5	16.9	0.010
Death	2 (9)	0	0.001
Transplanted	4 (17)	1 (2)
Rehab/intensive care units	4 (17)	0
Readmission/relapse4	7 (30)	18 (44)
Discharge)	6 (26)	22 (54)
Ethnicity			0.254
White	17 (27)	29 (45)
Black	1 (2)	6 (9)
Hispanic/Latino	5 (8)	4 (6)
Other	0 (0)	2 (3)

¹n = 52, phosphatidylethanol (Peth) performed prior to steroid initiation.

²Positive PEth value > 20 ng/mL.

³n = 64, alcohol-associated hepatitis treated with steroid.

⁴Low mortality risk (i.e., mild infection, headache, pain).

⁵Length of abstinence from alcohol consumption subjectively reported by patient.

Nonresponse: Post steroid day 4 or day 7 Lille score ≥ 0.45; Peth: Phosphatidylethanol; Response: Post steroid day 4 or day 7 Lille score < 0.45.

Non-responder status correlated significantly with pre-steroid increased age (P = 0.024), lower albumin (P = 0.003), and higher bilirubin (P = 0.010) (Table 1). Disposition 6 months post-steroid for non-responders showed statistically significant increased mortality, more liver transplantations performed, more intensive care units (ICU) and rehab admissions, and fewer discharges. Responders were more likely to have readmission or relapse (Table 1). No statistical significance was found between ethnicity and response to steroid therapy (P = 0.254) (Table 1).

DISCUSSION

PEth value was a poor predictor of non-responder status to steroids in patients with AAH (P = 0.210). A longer time of alcohol abstinence and a fewer number of average standardized drinks per week were associated with nonresponse to steroid therapy, though statistically insignificant (P = 0.123 and P = 0.136 respectively). Non-responders to steroid therapy were more likely to display abnormal laboratory findings of liver function, including lower albumin and elevated bilirubin, despite evidence of lower ethanol consumption. Thus, steroid therapy was more likely to benefit patients presenting with greater biochemical liver reaction to alcohol consumption, irrespective of the amount of alcohol consumption. Steroids act by reducing inflammation through cytokines, including tumor necrosis factor-α, interleukin (IL)-6, IL-8, and intercellular adhesion molecule 1[13,14]. Abundant inflammation is associated with AAH and in a study by Mathurin et al[15], successful steroid treatment for AAH was associated with patients with a white blood cell count > 5500/cm and increased polymorphonuclear leucocyte infiltration on liver biopsy. A reduction in bilirubin at 1 week following steroid use is also a reliable and specific biomarker for response to steroid therapy and is incorporated in the Lille score calculation[16].

Identifying reliable factors to identify non-responders to steroid therapy for AAH remains a difficult task, though early identification may lead to earlier placement on the liver transplantation list and reduction of medical cost and potential harm caused by steroid therapy. Non-responders showed a statistically significant higher rate of death, ICU or rehab admission, and fewer discharges from the hospital causing longer length of stay, and higher medical cost than responders[17].

Despite historical institutional restrictions of 6-month minimum alcohol abstinence for liver transplantation eligibility in AAH patients, earlier liver transplantation can be a safe and serve as a definitive treatment for non-responders to steroid therapy[18]. Factors associated with alcohol abstinence post-liver transplant include older age, longer pre-transplant abstinence, presence of social support residing in the patient’s dwelling, and insight into the disease process[19]. We found younger age had a statistically significant association with response to steroid therapy for AAH. Aging is associated with alterations to steroid receptor expression and function and dysfunction of immune cell function in an asthma model, resulting in poorer therapeutic response to steroid therapy[20]. Therefore, older patients identified as having higher rates of nonresponse to steroid therapy in this study may also be better candidates for early liver transplantation. Despite some institutional support for reduced abstinence restrictions for early liver transplantation in AAH, there continues to be resistance within the transplantation community due to fear of ethical violations for transplantation selection[21,22].

Limitations of the study include the retrospective nature, relatively small sample size, as well as the study being performed at a single institution. Confounding variables such as comorbidities and social determinants of health such as healthcare coverage and economic status were not considered in the study. Prospective randomized controlled trials assessing inflammatory biomarkers such as erythrocyte sedimentation rate, C-reactive protein, and procalcitonin, as well as other liver enzymes as predictive factors in determining response of AAH to steroid therapy may yield more reliable and generalizable results.

CONCLUSION

Predicting the response of AAH to steroid therapy may allow for earlier enrollment onto the liver transplantation list for patients unlikely to respond to steroid therapy. PEth does not serve as a reliable predictor for steroid response. Biochemical markers of liver response to alcohol use, including bilirubin and albumin, as well as age may better predict steroid response in AAH. Prospective studies including larger population sizes will help to assess the efficacy of combining patient factors, such as pre-steroid age, albumin, bilirubin, and inflammatory markers to predict steroid response before administration.