The aim of this study was to investigate the utilization of pharmacological treatment for AUD in specialist care among patients with comorbid alcohol-attributable medical diagnoses using a cohort of the total population in Sweden.

Our findings reveal that nearly one in four patients was diagnosed with at least one alcohol-attributable medical disorder within one year of their AUD diagnosis. The most prevalent comorbid medical diagnoses were cardiovascular and gastrointestinal diseases, followed by neurological diseases and diabetes mellitus.

Across all categories of concurrent alcohol-attributable medical comorbidities, the odds for filling any AUD pharmacotherapy prescription were consistently lower compared to AUD diagnoses without medical comorbidities. The prescription receipt rate was particularly low among AUD diagnoses with comorbid cardiovascular diseases, followed by comorbid cancer, neurological and gastrointestinal diseases — a novel finding not previously reported in the literature.

There were only small differences in the odds of filling various types of pharmacotherapies. Comorbid diagnoses which were contraindications to AUD medication were associated with lower odds for filling a prescription but did not fully explain the observed prescription gap. In particular, diagnoses with contraindications for one type of medication (e.g., disulfiram) were also associated with lower odds for receiving prescriptions for acamprosate and naltrexone. This underscores the complexity of prescribing patterns and suggests that the barriers to pharmacotherapy extend beyond contraindications.

In 10% of all AUD diagnoses, a concurrent gastrointestinal disease was identified. Among this group, approximately 16% filled a prescription for AUD pharmacotherapy. This represents a notably higher prescription receipt rate compared to studies from the United States, where only 0.5–2.4% of patients with comorbid AUD and liver diseases received pharmacological AUD treatment [11, 17]. This disparity could be explained by variations in study design, or differences in health care systems.

Alcohol-related liver diseases, including conditions such as liver fibrosis, cirrhosis, alcoholic hepatitis and pancreatitis, constitute some of the most prominent adverse health consequences of alcohol consumption, with about half of all liver-related diseases attributed to alcohol [18]. Furthermore, effective management of alcohol consumption, particularly achieving and maintaining abstinence, plays a crucial role in increasing survival rates in liver diseases [19], whereas the persistence of alcohol consumption contributes to complications and progression [20]. AUD pharmacotherapy is considered cost-effective for patients with alcohol-related liver cirrhosis. A recent study demonstrated that medication-assisted AUD therapy in patients with alcohol-related liver cirrhosis provided greater benefits at lower costs compared to no intervention [21]. Existing evidence supports the effectiveness of AUD treatment in individuals with liver diseases and suggests treatment with acamprosate for this group [20, 22]. Similarly, interventions to reduce alcohol consumption have been shown to reduce episodes of acute pancreatitis [18, 23]. The implementation of a multidisciplinary management approach for chronic pancreatitis, involving pancreatologists and addiction specialists, has also demonstrated positive effects on patient’s drinking behaviour [24].

Cardiovascular diseases constituted the second largest category of medical diseases in the present study. In 9% of all AUD diagnoses, a concurrent cardiovascular diagnosis was present. Despite their high prevalence, cardiovascular diseases showed the lowest prescription receipt rates among the medical disease categories. Specifically, odds for an individual with both cardiovascular disease and AUD filling a prescription were 59% lower in comparison to those with only AUD.

Cardiovascular diseases remain one of the leading causes of death in Europe [25]. Reduced alcohol use has been linked to improved blood pressure, with the greatest benefits observed among heavier drinkers [26]. In line with this, improvements in blood pressure have been found following treatment for alcohol dependence [27]. In a general population sample, a reduction of at least 150 g of alcohol per week among adults with heavy alcohol use was associated with lower odds for a range of self-reported cardiovascular diseases such as arteriosclerosis, angina, tachycardia, or myocardial infarction [28]. These findings emphasize the importance from a public health perspective for addressing the low prescription rate for AUD pharmacotherapy among individuals with cardiovascular diseases.

Among AUD diagnoses and either comorbid neurological disorder, diabetes mellitus, infectious diseases, or cancer, the rates of prescription receipt were similar, ranging from 15 to 18%. However, the prevalence of these comorbid diagnoses varied; approximately 4–5% of all AUD diagnoses had a comorbid neurological diagnosis or diabetes mellitus, while only 0.5-1% had a concurrent infectious disease or cancer diagnosis.

Alcohol consumption can impact health outcomes of pre-existing epilepsy by interfering with anti-epileptic drugs [29]. Moreover, alcohol consumption has been associated with reduced diabetes self-care behaviours as well as lower engagement with diabetes-related care [30, 31]. Malignancy stands out as one of the leading causes of premature death among AUD patients [32]. Additionally, in HIV/AIDS patients, alcohol use has been associated with an increased risk of transmission through sexual risk behaviour, non-compliance with antiretroviral treatment and disease progression, ultimately contributing to increased HIV/AIDS mortality [33,34,35,36]. However, similar to our findings, a recent study among American veterans identified low initiation and retention in AUD treatment among people living with HIV [37].

The observed treatment gap in the present study may be attributed to factors at multiple levels, including health care systems, clinical practices, and the individual patient [8].

At the health care level, improved integration of specialist addiction consultation teams into the secondary level of medical care has been shown to increase initiation of pharmacotherapy [38]. For example, the implementation of an inpatient addiction medicine consultation service in a general hospital has been shown to effectively reach patients with medical diseases and AUD [39]. Interventions targeting cardiovascular and gastrointestinal settings could be particularly impactful, given they were the two largest groups of medical disorders co-occurring with AUD in this study. Additionally, mapping the current prescription practices of AUD pharmacotherapy in different health care settings, is crucial for developing healthcare services.

Moreover, development of new pharmacotherapies or repurposing of existing medications could offer additional treatment alternatives. For instance, baclofen can be prescribed to individuals with AUD and concurrent liver disease. A recent meta-analysis showed promising results for baclofen in reducing heavy drinking and increasing abstinence compared to placebo [40]. Another potential agent is varenicline [41]. However, in our study the lower odds for pharmacotherapies were not fully explained by the presence of contraindications, suggesting additional barriers beyond limited pharmacotherapy options.

On the clinician level, known barriers to prescribing AUD pharmacotherapy include perceived lack of effectiveness, time constrains and inadequate training [42,43,44,45]. Furthermore, stigmatizing attitudes among healthcare professionals towards individuals with substance use disorders are increasingly recognized as barrier to AUD treatment engagement [46, 47].

Barriers on the patient level are low knowledge of AUD pharmacotherapies [48], and the perception that AUD treatment is not effective [49]. Studies specifically focusing on individuals with AUD and medical comorbidities, highlight barriers such as the desire to handle alcohol-related problems independently, reluctance to abstinence-only-treatments, a perceived lack of integration between addiction care and medical care, and fear of stigmatisation [50, 51].

This study used register-based data of the total population of Sweden. The registers show high level of completeness, with a low risk of selection bias. The registers have high internal validity, for example, the National Patient Register shows a positive predictive value of 85–95%, suggesting a high overlap between registered diagnoses and medical records [52]. Another strength is the use of a recognized measure of prescription rates the year following AUD diagnosis [53], contributing to a high clinical relevance of the results.

One important limitation is the absence of data on kidney diseases, which are contraindicators for acamprosate and naltrexone treatments. Furthermore, data on prescriptions other than AUD pharmacotherapy, potentially encompassing contraindications (e.g., opioid treatment), were not included. Additional methodological limitations include the measurement of medical diagnoses only following, not preceding, the AUD diagnosis, and that it was not recorded whether the medical diagnosis occurred before or after the receipt of the AUD prescription.

Also, the study solely considered filled prescriptions of AUD pharmacotherapy, capturing both the prescriber and patient behaviours. Previous Swedish research indicated an overlap of 83% between issued and filled prescriptions [54], suggesting that the presented results are largely attributed to behaviours among prescribers rather than patients. The absence of data on AUD pharmacotherapies dispensed directly at the clinic, especially for disulfiram, may lead to an underestimation of the prescriptions; however, this limitation most probably does not change the general conclusion.

Finally, register data relying on diagnoses from specialist and inpatient care represents a conservative measure of AUD in the general population, primarily capturing the more severe continuum of AUD. The dataset did not include primary care data, where up to half of all AUD and a large part of medical diagnoses are made [55].

The data for this study was collected between 2007 and 2015, which may be considered a limitation. However, during this period, there were no significant changes in Swedish policy or health care organization that would substantially impact the study results.