Naturalistic comparison of clomethiazole and Diazepam treatment in alcohol withdrawal: effects on oxidative stress, inflammatory cytokines and hepatic biomarkers

Abstract

Ethanol is metabolized by alcohol dehydrogenase to acetaldehyde and induces cytochrome P450 2E1 (CYP2E1), which generates reactive oxygen species that cause inflammatory liver damage. Clomethiazole, a drug approved for alcohol withdrawal treatment (AWT) in some European countries, inhibits CYP2E1. We hypothesized that clomethiazole would lead to a faster reduction in oxidative stress, inflammatory cytokines, and liver enzymes compared to diazepam treatment. We analysed respective biomarkers in 50 patients undergoing AWT and 25 healthy individuals but found no statistical difference between the two medication groups over 3–5 days. Hence, our hypothesis was not confirmed during this observation period.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00406-024-01835-7.

Introduction

Alcohol addiction remains a significant global public health concern, with a global prevalence rate of around 3 per cent, which varies by region [1]. Worldwide, 3 million deaths per year result from harmful use of alcohol, representing 5.3% of all deaths. Beyond health consequences, the harmful use of alcohol brings significant social and economic losses to individuals and society. Nevertheless, the availability and coverage of alcohol withdrawal treatment is limited in many parts of the world [1].

Chronic alcohol consumption of more than 40 g per day induces cytochrome P450 2E1 (CYP2E1), which enhances its own oxidation to acetaldehyde [2]. Structural cellular and mitochondrial alterations by acetaldehyde cause decreased adenosine triphosphate generation and production of reactive oxygen species (ROS). These cause damage to macromolecules, including lipid peroxidation and protein oxidation, with generation of oxidation products such as malondialdehyde (MDA) and dityrosine, and production of proinflammatory cytokines that activate neutrophils and induce liver damage [2]. In Germany, alcohol withdrawal treatment (AWT) is administered mainly as inpatient therapy. National guidelines generally recommend use of symptom-based protocols, which tend to have shorter treatment durations [3, 4]. Clomethiazole (CMZ) is the only drug approved for AWT in Germany.1 Nevertheless, many healthcare providers use benzodiazepines such as diazepam (DIAZ), although a network meta-analysis by Bahji et al. has shown that CMZ appears to be superior to DZP and some other benzodiazepines in preventing seizures and reducing the severity of alcohol withdrawal [5]. The finding that CMZ inhibits CYP2E1 [6], suggests that AWT with CMZ might result in less oxidative stress and inflammation, and hence quicker normalization of liver parameters compared to treatment with benzodiazepines [7, 8].

Our naturalistic prospective study aimed to assess the benefits of either CMZ or DIAZ in AWT, by applying longitudinal measurements of biomarkers for oxidative stress, inflammation and liver injury in patients.

Methods

Study population

From October 2017 to August 2020, patients (18–65 years-old) routinely admitted for AWT were screened for participation in this prospective study. Ethics approval was obtained from the Otto-von-Guericke-University Magdeburg Institutional Review Board (reference number 80/17). We applied ICD-10 criteria to diagnose alcohol dependence and the Alcohol Use Disorders Identification Test (AUDIT) was used as a supporting test [9]. Alcohol craving was measured with the Obsessive Compulsive Drinking Scale (OCDS) [10]. Exclusion criteria encompassed concomitant immunological diseases, severe infections, trauma, terminal illnesses affecting the brain (e.g. severe liver, kidney, heart and lung diseases, and tumours), treatment with cortisone or other immunosuppressive or immunomodulating substances, co-use of psychotropic substances, and incapacity or refusal for additional blood tests. In this study, the choice of drug was naturalistic and the study was prospective, with CMZ and DIAZ being administered using symptom-based protocols. A study sketch was created which was approved by the ethics committee and the patients who came to our department for AWT gave written informed consent to be willing to allow longitudinal blood takes, willingness to answer clinical questionnaires and for analysis of data from rating scales. Two treatment groups of 25 patients were established, matched for sex, body mass index (BMI), age, smoking, chronic obstructive pulmonary disease (COPD), and age at first AWT (Table 1; [4, 11]).2 Control samples (n = 25) were obtained during the same period from healthy donors, including hospital staff, students and their relatives.

Table 1.

Demographic, clinical and routine laboratory data at baseline

Variables	CMZ (n = 25)	DIAZ (n = 25)	Controls (n = 25)	Test	Test value	p-value	Post-hoc tests (p-value, FDR)
							DIAZ-CMZ	CMZ-Controls	DIAZ-Controls
Age (y)	51.0 (42.0;57.0)	51.0 (38.0;56.5)	53.0 (43.0;56.5)	H-test	KW = 0.26	0.878	0.838	0.838	0.838
Sex (female/male)	5/20	4/21	4/21	Fisher exact	–	1.000	1.000	1.000	1.000
BMI (kg/m2)	24.6 (22.0;27.0)	24.0 (19.6;26.7)	25.7 (24.0;28.3)	H-test	KW = 4.43	0.109	0.453	0.204	0.168
Smoking (yes/no)	19/6	22/3	2/23	Fisher exact	–	<0.001	0.463	<0.001	<0.001
COPD (yes/no)	Yes: 3/No: 22	Yes: 4/No: 21	Yes: 0/No: 0	Fisher exact	–	1.000	1.000	1.000	1.000
Age at first alcohol withdrawal treatment (y)	41.0 (29.3;44.5)	32.0 (24.5;49.5)	–	U-Test	W = 102.00	0.948
AUDIT (sum score)	27.0 (22.0;31.0)	29.0 (23.5;33rr.0)	2.0 (0.0;3.0)	H-test	KW = 46.17	<0.001	0.517	<0.001	<0.001
OCDS (sum score)	19.0 (15.0;23.0)	21.0 (14.8;27.0)	1.0 (0.0;1.0)	H-test	KW = 47.40	<0.001	0.562	<0.001	<0.001
Leukocyte count (× 109/L)	7.60 (6.12;8.74)	7.00 (6.00;9.00)	6.77 (5.94;7.40)	H-test	KW = 3.71	0.157	0.749	0.134	0.355
CRP (mg/L)	0.70 (0.25;4.01)	2.32 (0.60;5.85)	0.95 (0.63;2.45)	H-test	KW = 3.93	0.140	0.181	0.467	0.181
Biomarkers of alcohol consumption and liver damage
BAC (neg/pos)	7/18	6/19	25/0	Fisher exact	-	<0.001	1.000	<0.001	<0.001
BAC (o/oo, only for BAC pos cases)	1.9 (0.9;2.6)	2.1 (1.4;2.5)	–	U-test	W = 144.50	0.429
EtG in whole blood (neg/pos)	0/25	1/24	19/6	Fisher exact	-	<0.001	1.000	<0.001	<0.001
EtG in whole blood (ng/mL)	1736 (590;4158)	3355 (1371;4820)	5 (2;23)	H-test	KW = 14.67	<0.001	0.298	<0.001	<0.001
PEth in whole blood (neg/pos)	0/25	0/25	12/11	Fisher exact	-	<0.001	1.000	<0.001	<0.001
PEth in whole blood (µmol/L)	4.69 (3.44;6.14)	4.28 (3.12;6.24)	0.09 (0.07;0.14)	H-test	KW = 26.53	<0.001	0.882	<0.001	<0.001
CDT (neg/pos)	6/18	6/19	23/0	Fisher exact	-	<0.001	1.000	<0.001	<0.001
CDT (%)	5.1 (3.0;9.9)	6.4 (3.7;10.1)	–	U-test	W = 167.00	0.915
GGT (µmol/L)	2.44 (1.25;7.91)	1.96 (0.73;3.01)	0.38 (0.29;0.56)	H-test	KW = 31.01	<0.001	0.187	<0.001	<0.001
ALAT (µmol/L)	0.84 (0.55;1.31)	0.84 (0.53;0.96)	0.36 (0.30;0.46)	H-test	KW = 25.30	<0.001	0.438	<0.001	<0.001
ASAT (µmol/L)	1.08 (0.77;1.64)	1.10 (0.63;2.16)	0.41 (0.37;0.47)	H-test	KW = 41.89	<0.001	0.915	<0.001	<0.001
ASAT/ALAT ratio	1.29 (1.05;1.92)	1.43 (1.15;2.01)	1.08 (0.90;1.28)	H-test	KW = 11.31	0.004	0.528	0.027	0.003
MCV (fl)	93.7 (90.1;97.5)	91.0 (87.8;95.9)	87.8 (83.8;91.4)	H-test	KW = 18.47	<0.001	0.114	<0.001	0.013

Data are presented as median (quartile 1; quartile 3) or number of cases

p-values in bold are statistically significant

FDR false discovery rate, neg negative, pos positive

Samples and analyses

EDTA-blood, serum and urine samples were obtained upon admission before medication initiation (baseline, BL) and on days 3 (D3) and 5 (D5). All participants underwent a urine drug screen. A complete differential blood count was determined using an XN-3000 automated counter (Sysmex Corporation; Kobe, Japan). Measurements of C-reactive protein (CRP), creatinine, glucose, lipids, and biomarkers of liver damage [alanine-aminotransferase (ALAT), aspartate-aminotransferase (ASAT), gamma-glutamyltransferase (GGT)] were determined on a random access nalyser (cobas c701/c501, Roche Diagnostics; Mannheim, Germany).

Markers of alcohol consumption at admission included blood alcohol concentration (BAC: cut-off 0.1°/_oo, enzymatic alcohol dehydrogenase/ADH oxidation), ethyl glucuronide (EtG: cut-off 1 ng/mL) [12], phosphatidylethanol (Peth: cut-off 0.050 µmol/L) [13], and carbohydrate-deficient transferrin (CDT: cut-off 2.0%, high performance liquid chromatography/HPLC, ClinRep®, RECIPE Chemicals + Instruments; GmbH, Germany).

Blood CMZ concentrations were determined on an Acquity Ultra Performance Liquid Chromatograph connected to a Xevo-XS tandem mass spectrometer using the Chromsystems [Waters; Munich, Germany; limit of quantification (LOQ) 15 ng/mL]. Diazepam and the active metabolite Nordiazepam were quantified using the same system (LOQ 1 ng/mL).

As indicator of oxidative stress, total MDA in plasma samples was quantified through derivatization with 2–4-dinitrophenylhydrazine, and urinary dityrosine levels was measured using HPLC with fluorimetric detection following solid-phase extraction (Macherey & Nagel, Germany), as described [14]. Serum concentrations of 17 inflammatory cytokines (Table 2) were assessed using the LEGENDplex™ Human Inflammation Panel 1 multiplex immunoassay (BioLegend, San Diego, USA).

Table 2.

Linear Mixed Model (LMM) analyses of liver enzymes, MDA, dityrosine or cytokines as the respective dependent variables, drug type (CMZ or DIAZ) and time point (baseline, day 3, day 5) as fixed factors, and blood alcohol concentration (BAC) on admission as a covariate

Variables	Influence of drug type (p-value)	Influence of time (p-value)	Influence of BAC on admission (p-value)	Interaction drug type X time (p-value)
GGT (µmol/L)	0.616	0.050	0.647	0.368
ALAT (µmol/L)	0.841	0.113	0.482	0.069
ASAT (µmol/L)	0.709	0.567	0.092	0.129
ASAT/ALAT ratio	0.792	<0.001	0.040	0.481
MDA (nmol/mL)	0.617	<0.001	0.011	0.505
Dityrosine (ng/mL)	0.495	0.002	0.407	0.099
Cytokines
IFN-α2 (pg/mL)	0.379	0.136	0.142	0.891
IFN-γ (pg/mL)	0.042	0.594	0.721	0.909
IL-1β (pg/mL)	0.209	0.151	0.567	0.631
IL-6 (pg/mL)	0.607	0.003	0.035	0.326
IL-8 (pg/mL)	0.308	0.659	0.867	0.384
IL-10 (pg/mL)	0.220	0.018	0.019	0.418
IL-12p70 (pg/mL)	0.895	0.348	0.404	0.587
IL-17A (pg/mL)	0.816	0.079	0.760	0.714
IL-18 (pg/mL)	0.646	<0.001	0.261	0.995
IL-23 (pg/mL)	0.241	0.303	0.927	0.712
IL-33 (pg/mL)	0.343	0.369	0.200	0.700
MCP-1 (pg/mL)	0.575	<0.001	0.982	0.202
TNF-α (pg/mL)	0.214	0.791	0.822	0.403

p-values in bold are statistically significant

Statistics

Data analyses were conducted using the open-source statistical software package R. Group differences between categorical data were determined through Fisher’s exact tests. Non-parametric Kruskal–Wallis H-tests and Mann–Whitney U-tests were employed to calculate group differences. Friedman’s test, along with post-hoc U-tests for repeated measures data, was applied to analyse the time course of oxidative stress, inflammatory markers and liver enzyme biomarkers within the CMZ and DIAZ treatment groups. To control for multiple hypothesis testing, Benjamini–Hochberg false discovery rate (FDR) p-value adjustment was utilized. A linear mixed-effects model (LMM) was employed to assess interactions between drug type and time, with liver enzymes, MDA, dityrosine, or inflammatory markers as dependent variables, drug type (CMZ or DIAZ) and time point (BL, D3, D5) as fixed factors, and blood alcohol concentration on admission serving as a covariate. CMZ was converted to DIAZ equivalents for statistical purposes (192 mg CMZ ≙ 5 mg DIAZ) [15]. All statistical tests were two-tailed with p < 0.05 considered significant.

Results

Baseline demographics, clinical and routine laboratory data

No significant differences were observed between the CMZ, DIAZ, and control groups in terms of age, sex, BMI, COPD, leukocyte count and CRP levels at baseline. The CMZ and DIAZ groups had significantly higher AUDIT and OCDS scores, and significantly more smokers compared to controls (p < 0.001, Table 1). Significant differences were observed among all biomarkers of alcohol consumption and liver damage, including BAC, EtG, Peth, CDT, GGT, ALAT, ASAT, ASAT/ALAT ratio (i.e. the de Ritis ratio) and MCV, indicating higher values in both CMZ and DIAZ groups compared to controls (p < 0.05, Table 1), without significant differences between CMD and DIAZ groups at baseline.

Time course in CMZ and DIAZ treatment groups

(a) Duration of AWT, required drug dose and complications

Both treatments were comparable in terms of mean cumulative drug dose required [CMZ group: 42.5 (26.3;87.5) vs. DIAZ group: 45.0 (35.0;107.5) DIAZ equivalents (mg), p = 0.476] and complication rates [CMZ group: 1 of 25 vs. DIAZ group: 2 of 25 patients with delirium (haloperidol treatment), p = 1.000; CMZ group: 0 of 25 vs. DIAZ group: 2 of 25 patients with hypertensive crisis (clonidine or nitrendipine treatment), p = 0.490; CMZ group: 0 of 25 vs. DIAZ group: 1 of 25 with epileptic seizure, p = 1.000]. Mean AWT duration tended to be shorter in CMZ versus DIAZ group, but this did not reach statistical significance [74.3 (51.8;106.0) vs. 101.8 (74.0;133.7) h, p = 0.065]. No serious adverse events were observed for either treatment.

(b) Biomarkers of liver damage, oxidative stress and inflammation

Only the CMZ group showed significant changes over time in the biomarkers for liver damage GGT and ASAT, while the ASAT/ALAT ratio decreased over time in both the CMZ and DIAZ groups (Supplementary Table). However, no significant differences over the time were found between CMZ and DIAZ groups (Table 2, drug type X time interaction; Supplementary Figure).

MDA increased significantly over time on D3 and D5 compared to BL in the DIAZ group. However, MDA and dityrosine showed no significant differences among CMZ and DIAZ groups over time (drug type X time interaction).

IL-6, IL-10 and MCP-1 decreased significantly in DIAZ, and IL-18 was significantly decreased in both CMZ and DIAZ over time, with no significant differences between treatment groups at any time point.

Discussion

Although not significant, our study showed that the CMZ group took approximately 24 h less to complete the AWT than the DIAZ group. This could be advantageous because patients without physical withdrawal symptoms are better able to participate in psychotherapeutic interventions. However, our main hypothesis that AWT with CMZ would lead to a quicker normalisation of oxidative stress, inflammatory cytokines and liver enzymes compared to DIAZ during the 5-day preservation period was not proven.

This finding may contrast in part to two studies by Hohmann et al. covering the same clinical trial, which demonstrated an advantage of CMZ over the benzodiazepine clorazepate regarding regression of CYP2E1 and ALAT [16, 17]. However, it should be noted that our naturalistic study covered the clinical course and laboratory parameters for only 3–5 days, while Hohmann et al. addressed a longer observation period of 7–10 days. Consequently, we cannot say whether a longer period of observation might have resulted in significant differences between medication groups. Furthermore, the previous findings with clorazepate might not be completely transferrable to DIAZ even though both drugs are classed as benzodiazepines.

The finding that the ASAT/ALAT ratio decreased over time under CMZ and DIAZ therapy may indicate a reduction in liver damage with each of these treatments [18]. Although some other markers of oxidative damage (MDA) and inflammation (IL-6, IL-10 and MCP-1) showed potential differences over the abbreviated time course, a daily and more extensive time course would be required to confirm these differences between the two treatment groups.

Considering these inconclusive results, we suggest the need for larger clinical studies comparing the effects of CMZ with those of multiple benzodiazepine compounds. In addition to assessment of potential adverse events, these studies should include measurements of CYP2E1 activity, as well as biomarkers of oxidation–reduction (REDOX), inflammation and liver damage. This will be important in the search for CYP2E1 inhibitors that are effective and safe to use in AWT over longer time periods.

More fundamentally, however, the question arises as to whether it would actually have a relevant influence on the medium to long-term clinical outcome if a drug is superior in terms of oxidative stress, inflammation and liver health during a few days of AWT, given the long-term accumulated liver damage caused by alcohol dependence.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

ADH

Alcohol dehydrogenase

ALAT

Alanine-aminotransferase

ASAT

Aspartate-aminotransferase

AUDIT

Alcohol use disorders identification test

AWT

Alcohol withdrawal treatment

BAC

Blood alcohol concentration

BL

Baseline

BMI

Body mass index

CDT

Carbohydrate-deficient transferrin

CMZ

Clomethiazole treatment group

COPD

Chronic obstructive pulmonary disease

CRP

C-reactive protein

CYP2E1

Cytochrome P450 2E1

D3

Day 3

D5

Day 5

DIAZ

Diazepam treatment group

EtG

Ethyl glucuronide

FDR

False discovery rate (corrected p-value)

GGT

Gamma-glutamyltransferase

HPLC

High performance liquid chromatography

IFN

Interferon

IL

Interleukin

LOQ

Limit of quantification

LMM

Linear mixed-effects model

MCP-1

Monocyte chemotactic protein 1

MCV

Mean corpuscular volume

MDA

Malondialdehyde

OCDS

Obsessive compulsive drinking scale

PEth

Phosphatidylethanol

REDOX

Oxidation–reduction

ROS

Reactive oxygen species

TNF

Tumour necrosis factor

Author contributions

JS, BR and MB designed the study. BR performed the cytokine analyses. MB determined EtG, PEth, CMZ and benzodiazepine blood levels. WJ provided the oxidative stress data and KB performed the routine laboratory and CDT measures. DS, AK, GML and CJS recruited and characterised patients and controls under supervision of JS. HD was responsible for data collection, database management and statistical analyses under supervision of JS and PCG. DS, AK and JS wrote the first manuscript draft. PCG, WJ, UJM, KB, MB and BR helped to interpret the data, reviewed the manuscript and made contributions for improvement. All authors read and approved the final manuscript.

Funding

Open Access funding enabled and organized by Projekt DEAL. This study was funded by CHEPLAPHARM Arzneimittel GmbH (Greifswald, Germany) under the title "Biomarker Analysis - Oxidative Stress in Alcohol Withdrawal" (CMZ study).

Availability of data and materials

All data are either presented in the manuscript or supplementary materials. Additional information can be made available upon reasonable request.

Declarations

Conflict of interest

None of the authors declare any competing interests. The study design, data collection, interpretation of results, and manuscript preparation were not influenced by the funding source.

Ethics approval

Ethics Committee of the Otto von Guericke University in Magdeburg (reference number 80/17).