Validation of the French Version of the Prediction of the Alcohol Withdrawal Severity Scale (PAWSS)

Context Alcohol use disorders (AUDs) are reported in 10-32% of hospitalized medical patients [Nielsen et al. 1994; Smothers et al. 2004; Dolman and Hawkes 2005; Doering-Silveira et al. 2014], in 23 % of hospitalized medical patients and in 25-50% of hospitalized psychiatric patients in France [Paille \& Reynaud 2015]. A hospital admission may result in an abrupt cessation of alcohol consumption for individuals with AUD and thus provide a risk period for alcohol withdrawal syndrome. Even though the majority of patients at risk of AWS will develop only minor or uncomplicated withdrawal symptoms, up to 20% of patients develop symptoms associated with complicated alcohol withdrawal, including withdrawal seizures and delirium tremens [Maldonado et al., 2010]. Complicated alcohol withdrawal syndrome is associated with increased in-hospital morbidity and mortality, increased lengths of stay, inflated costs of care, and worsened cognitive functioning.

Due to lack of any similar previously existing tools, Maldonado et al. developed the Prediction of Alcohol Withdrawal Severity Scale (PAWSS) [Maldonado et al. 2014, 2015]. The PAWSS scale has excellent psychometric characteristics and predictive value of complicated alcohol withdrawal among medically ill hospitalized patients using a PAWSS cutoff of 4, the tool's sensitivity was 93.1% and specificity was 99.5%. The PAWSS scale is recommended by the American Society of Addiction Medicine for alcohol withdrawal management, (ASAM 2020, www.asam.org).

There is no scale available to identify the patients at risk of complicated withdrawal In France. The PAWSS is not a withdrawal severity scale such as the CIWA-AR or Cushman scale, which are scored during withdrawal, but a scale designed to screen beforehand, prior to withdrawal, patients at high risk for the most severe complications (Delirium tremens and seizures), and thus to take appropriate therapeutic measures. It also has the advantage of being able to guide patients towards the most suitable scheduled withdrawal modalities for their condition.

The PAWSS can therefore greatly assist clinicians in directing patients to either outpatient or hospital withdrawal in more or less intensive care, with a minimum of risk. Its use would allow earlier management to limit the sequelae, especially cognitive and those related to seizures.

The PAWSS scale includes an initial screener question and can be used with patients who are not currently exhibiting signs of withdrawal. If a patient does indeed endorse recent intake of alcohol, this must be followed by 10 questions contained in the second part of PAWSS, assessing known risk factors for withdrawal and current clinical status. The PAWSS is heavily based on self-report of alcohol intake and history provided by patients, as the literature suggests that interviews by clinicians can provide the most accurate information on alcohol abuse and relapse, as compared to collateral information or selected laboratory data.

Objectives The primary objective is to validate the French version of the PAWSS scale (PRSAC) in a population of patients with AUD.

The secondary objectives are 1) to confirm the optimal cut-off score of 4 for the PRSAC scale, 2) to evaluate PRSAC scale reproducibility in predicting the risk of complicated alcohol withdrawal syndrome and 3) to identify an optimal strategy for benzodiazepine use following the application of PRSAC scale.

Hypothesis Are the psychometric properties of the French version of the PAWSS scale (PRSAC) comparable to those of the English version in a population of alcohol use disorders subjects?

Methods This study is a multicentric non-interventional prospective cohort study. Subject evaluation included during 3 days.

Day 1: Collection of consent and clinical data, DSM-5 criterion for alcohol use disorders, AUDIT, PRSAC scale (twice by two investigators), and CIWA-AR scale every 8 hours.

Day 2 and Day 3: administration of CIWA-AR scale every 8 hours.

Inclusion criteria Male or female, aged 18 to 60 years Subjects with DSM-5 criteria for moderate to severe alcohol use disorders (AUD) Subjects hospitalized for AUD or another condition with AUD. Subjects who understand French and affiliated to the social security system

Non-inclusion criteria Subject with a neurological disorder, impairing cognition (epilepsy, dementia,…) Subjects unable to understand the questionnaire and to give consent or not volunteering for the study Subject under guardianship or curatorship

Study location

Academic substance abuse treatment and Psychiatry departments of:

Barthélemy Durand Hospital at Etampes, France Paul Brousse hospital in Villejuif, France Rouen university hospital, France Caen university hospital, France Martinique university hospital, France

Results \& Outcome Confirmation of the psychometric qualities of the French version of the PAWSS scale.