Alcohol is a psychoactive substance with dependence-producing properties,¹ with alcohol use disorder (AUD) being the most prevalent mental health disorder and one of the leading risk factors for morbidity, disability and mortality in the world.² The World Health Organization (WHO) 2011 Global Status Report ranked South Africa as a country with one of the riskiest patterns of alcohol consumption and with the highest reported alcohol consumption levels in Africa.² A study from South Africa showed that alcohol use was associated with an increased risk of psychiatric disorders³ and has huge economic, social and health costs.⁴ In South Africa, the Western Cape has the highest prevalence of lifetime alcohol use (45%) and risky drinking behaviour (15%),⁵ where alcohol has been identified as the third most common primary substance of abuse for which treatment is sought, after methamphetamines and cannabis, respectively.⁶

Alcohol’s mechanism of action includes effects on multiple neurotransmitter systems and receptors, and its acute intake causes depression in the central nervous system (CNS) as a result of increased gamma-aminobutyric acid (GABA) neurotransmission and antagonist action at glutamate’s N-Methyl-D-Aspartate (NMDA) receptor.⁷ Chronic use of alcohol causes adaptive downregulation of GABA receptors and leads to an increased requirement of alcohol to produce the same effects, a phenomenon called tolerance.⁸ However, subsequent abrupt reduction or cessation of alcohol intake in the chronic user results in a reduction in the alcohol-mediated CNS inhibition and the glutamate-mediated CNS excitation being left unopposed, resulting in the clinical features of alcohol withdrawal.^8,9 Alcohol withdrawal syndrome (AWS) starts within hours to a few days after cessation or reduction of heavy alcohol use and may present with: (1) sweating, (2) rapid pulse, (3) tremors, (4) insomnia, (5) nausea, (6) vomiting, (7) hallucinations, (8) agitation, (9) anxiety and (10) seizures.¹⁰ This disorder is potentially a life-threatening condition, especially in the case of complications including seizures or delirium tremens (DT).⁹

The effective management of AWS includes a combination of supportive and pharmacological measures that aims to make this period as comfortable and safe as possible.^11,12 Benzodiazepines are currently recommended as the first drug of choice in the management of AWS¹³ but have substantial side effects in addition to being potentially highly addictive and therefore need to be used with caution in patients with a lifetime history of substance abuse.¹⁴ There are currently two regimens commonly used for prescribing benzodiazepines for alcohol withdrawal: the fixed-dose regimen (FDR) and the symptom-triggered (STR) regimens.⁸ The use of STR has been proven to effectively treat uncomplicated alcohol withdrawal and reduces the risk of over- and under-medication in patients.^15,16 Previous studies have also found STR to be more advantageous compared with FDR because of shorter treatment duration, the lower total amount of benzodiazepines used (TABU) and reduced frequency of benzodiazepines used during the course of AWS in general medical settings and alcohol detoxification units.^15,16,17,18

The Revised Clinical Institute Withdrawal Assessment for Alcohol Scale (CIWA-Ar) is a 10-item rating scale with a maximum score of 67, which assesses the clinically pertinent features of alcohol withdrawal.¹⁹ It is quick to administer, has high inter-rater reliability and allows for hourly repeated scoring. It is also not copyrighted and may be reproduced freely. The CIWA-Ar is a widely accepted and validated assessment tool used to monitor the severity of AWS and in titrating pharmacotherapy.¹⁹ The use of the CIWA-Ar rating scale during alcohol withdrawal was implemented at Stikland Hospital in 2016, and so the current study aims to evaluate its effects on the dose of benzodiazepines prescribed during alcohol detoxification and to assess its impact on withdrawal-related outcomes.

No differences in any of the sociodemographic variables between the two groups was noted (Table 1). The clinical characteristics were also comparable across the two groups except for medical comorbidities, where 67.1% of the 2015 group had at least one medical comorbidity compared with 43.1% in the 2017 group (p = 0.005). The medical comorbidities included: (1) hypertension (25.9%), (2) dyspepsia (18.5%), (3) previous tuberculosis (11.9%), (4) other chronic lung diseases (7.4%), (5) human immunodeficiency virus (HIV) (5.9%), (6) diabetes mellitus (3.0%), (7) dyslipidaemia (2.2%) and (8) other medical comorbidities (other gastrointestinal diseases, cardiovascular diseases, previous traumatic brain injuries, previous cerebrovascular accidents, dermatological and musculoskeletal diseases, 19.3%) across the two groups. Depressive disorders were found to be the most common psychiatric comorbidity (2015: 20%, 2017: 23.1%) and other comorbid substance use disorders were noted in approximately 15% of the study sample with cannabis use disorder being the most common at 14%.

The alcohol-associated variables such as: (1) age at first alcohol intake, (2) duration of problematic drinking, (3) time since last alcohol intake at admission, (4) predominant pattern of drinking and (5) previous attendance to alcohol rehabilitation were similar across the two groups (Table 2). Moreover, approximately two-thirds of the patients reported drinking on a constant or daily basis (2015: 65.7%, 2017: 66.2%) compared with those who drank less frequently. More than a third reported having attended an alcohol rehabilitation programme previously (2015: 38.6%, 2017: 39.1%).

Regarding withdrawal-related outcomes (Table 2), there was no statistical difference in the duration of stay (p = 0.46) and outcome of admission (p = 0.09) between the two groups, and no complicated alcohol withdrawals were reported in either group. The percentage of patients requiring benzodiazepines during alcohol withdrawal decreased from 51.4% in 2015 to 33.8% in 2017 (p = 0.039). Among those who received benzodiazepines in the 2015 cohort (n = 36), all 36 (100%) had alcohol withdrawal as an indication and five (13.9%) had high blood pressure also marked as an indication, compared with those in the 2017 cohort (n = 22) where 15 (68.2%) had alcohol withdrawal marked as an indication while 13 (59.1%) had high blood pressure marked as an indication. Notably, from the 2017 cohort, 20 patients (90.9%) received benzodiazepines even though their maximum CIWA-Ar scores were less than 10 and 17 patients (77.3%) had a maximum CIWA-Ar score of less than nine (data not shown). Furthermore, seven (31.8%) patients in the 2017 group received benzodiazepines with high blood pressure as the only indication, all with maximum CIWA-Ar scores of less than 10. The median (IQR) TABU decreased from 5 (15) mg in 2015 to 0 (5) mg in 2017 (p = 0.01), and it was also noted to be higher (p = 0.031) in those who reported drinking constantly compared with those who drank 4–6 times per week and less frequently (Table 3).

The purpose of the present study was to perform a chart review of admissions over a six month period before and after the implementation of the CIWA-Ar rating scale during alcohol withdrawal and to evaluate whether it reduced benzodiazepine use and improved withdrawal-related outcomes in patients admitted to an alcohol rehabilitation programme.

As evidenced by previous studies evaluating STR against FDR,^15,16,17,18 the implementation of systematic and objective rating of withdrawal symptoms during alcohol withdrawal in the current study showed a statistically significant reduction in the TABU in the 2017 group when compared with benzodiazepines being prescribed pro re nata in the 2015 group. It was also consistent with a previous study in a naturalistic and low-risk population evaluating STR versus treatment as usual (TAU).²² However, the 2017 group also had a lower proportion of patients who reported having at least one medical comorbidity compared with the 2015 group, which might have been because of under-reporting from the patients or less strict documentation on admission. Notably, the patients in the current study all had stable chronic medical comorbidities, which are thought to contribute very little to the severity of AWS and are systematically underreported in this patient population.²³ Both groups had similar drinking patterns and previous rehabilitation episodes; this supports the assumption that the groups had similar disease severity. The current study thus supports the evidence that using an STR protocol over pro re nata prescription is more advantageous in terms of lowering the TABU during alcohol withdrawal despite the increase in the most favourable outcome (alcohol withdrawal not requiring treatment) not attaining statistical significance.

Despite fewer patients in the post-CIWA-Ar implementation group requiring treatment during withdrawal, it was noted that among those who did receive benzodiazepines, the majority had their maximum CIWA-Ar score below the cutoff of 10. While some studies have recommended prescribing benzodiazepines during withdrawal even for CIWA-Ar scores between 8 and 10,^9,24,25 it was also noted that almost a third of the patients received benzodiazepines for high blood pressure as the only indication with their CIWA-Ar scores being < 10. One explanation for the decreased prescription of benzodiazepines in the 2017 group could be that the nursing staff felt more comfortable with an objective measure of withdrawal symptoms and prescribed less benzodiazepines, with the notable exception of high blood pressure. The above implies that with stricter implementation of the CIWA-Ar during alcohol withdrawal and improved management of high blood pressure such as pre-admission screening and timely resumption of anti-hypertensives in those who had defaulted on their treatments, the TABU could have been minimised further.

There were no recorded complicated alcohol withdrawals across both groups, with most patients included in the programme completing at least 10 days post-admission. Only six patients prematurely discontinued the programme, with one patient from each group being diagnosed with active tuberculosis and being redirected to a specialised tuberculosis hospital while the remaining were self-discharges. The high completion rate of the programme and the absence of complicated alcohol withdrawals in the specialised rehabilitation programme are likely a result of the use of benzodiazepines during alcohol withdrawal, the enrolment of mostly low-risk patients through the self-booking process, thus implying higher motivation levels and the likelihood of patients lowering their alcohol consumption prior to admission, and the medical examination on admission where patients with unstable medical comorbidities are redirected to specialised medical facilities.

The majority of patients across both groups reported drinking daily or constantly compared with a less frequent predominant pattern of drinking, where severe alcohol dependence with higher frequency and higher daily units of alcohol consumption were associated with a higher risk of complications and higher amounts of benzodiazepine used during withdrawal.^8,26 In the present study, patients who reported a higher frequency of alcohol consumption were noted to require more benzodiazepines during alcohol withdrawal across both groups, with the caveat that the data collected included only the self-reported predominant pattern of drinking without any specific timeline prior to admission and no quantification of their alcohol consumption. The above makes an argument for more detailed documentation about the pattern and amount of alcohol consumption in units per week in the weeks prior to admission or measurement of the blood alcohol levels upon admission to have better insight into the severity of alcohol dependence and its effects on alcohol withdrawal symptom severity. Furthermore, the current study also highlighted the high incidence of dual diagnosis and other comorbid substance use disorders among the patients, emphasising the need for screening and intervention for the above as part of holistic biopsychosocial rehabilitation in the target population.

The current study is the only one to our knowledge that evaluated the use of an objective withdrawal rating scale in a South African setting, where we showed that the use of the CIWA-Ar is a safe and effective alternative to the previously used pro re nata prescription of benzodiazepines, which relied heavily on the nursing staff’s level of experience with alcohol withdrawal. Furthermore, in comparison with the 2020 South Africa National Standard Treatment Guidelines for uncomplicated alcohol withdrawal at the primary health care level, which recommend FDR, totalling 100 mg of diazepam over a period of 1 week,²⁷ the current study demonstrates that the implementation an objective withdrawal rating scale and STR during alcohol withdrawal could be beneficial in terms of reduction of benzodiazepine use while remaining safe and without any increase in negative outcomes if implemented at primary health care levels. The study also highlighted the need for stricter implementation of the CIWA-Ar scoring and optimisation of management of medical conditions during alcohol withdrawal to further limit the use of benzodiazepines and further supported the need for screening and intervention for other comorbid substance use disorders and dual diagnoses because of their high occurrence in the target population. However, there are a few potential limitations that need to be highlighted. First, it was a retrospective cohort analysis of standardised patient charts in one specialised ARU, which included a mostly low-risk and highly motivated population and excluded medically unstable patients, thus limiting its applicability to other general medical settings. Also, the current study did not evaluate other important factors in AUD rehabilitation like the patients’ comfort, satisfaction, duration of detoxification and relapse-associated factors in repeat admissions, which should be the focus of future studies.

The present study demonstrated that the use of a rating scale linked STR during alcohol withdrawal in a specialised ARU is a safe and effective alternative and led to a decrease in the number of patients requiring benzodiazepines as well as a decrease in the TABU, without increasing the incidence of complicated alcohol withdrawals.