Anticholinergic Burden and Safety Outcomes in Older Patients with Chronic Hepatitis C: A Retrospective Cohort Study

Aim: Older patients with chronic hepatitis C infection starting direct-acting antivirals (DAAs) are frequently prescribed multiple medications that may be categorized as inappropriate. Anticholinergic burden has been shown to be a predictor of adverse health and functional outcomes. Different scales are available to calculate anticholinergic burden. The aim of this study was to determine the prevalence of anticholinergic medication among older patients treated with DAAs and the risk factors associated using the Anticholinergic Cognitive Burden (ACB) scale, the Anticholinergic Risk Scale (ARS) and the Anticholinergic Drug Scale (ADS) and analyze the resulting safety consequences. Methods: Observational, retrospective cohort study of consecutive patients ≥65 years old receiving DAAs and taking concomitant medication. This study was conducted in accordance with the Strengthening the Reporting of observational studies in Epidemiology Statement. Results: 236 patients were included. The average age was 71.7 years, 73.3% cirrhotic, and 47% patients took ≥5 medicines. According to the ACB, ARS and ADS scales, 35.2% (n = 83), 10.6% (n = 25) and 34.3% (n = 81) of the patients were treated with anticholinergic medication. Two hundred-and-six (86%) patients presented any adverse events (AEs) during therapy. ARS scale showed a significant relationship between presence of anticholinergic medication and AEs. A large number of patients suffered anticholinergic events, with more events per patient in patients taking anticholinergic drugs. Conclusions: Older hepatitis C chronic patients are exposed to potentially inappropriate polypharmacy and anticholinergic risk, according to the ACB, ARS and ADS scales. The three scales showed different results. Only the ARS scale was associated with AEs, but the rate of anticholinergic effects per patient was significantly higher in patients with anticholinergic drugs, regardless of the scale used. Consider quality of pharmacotherapy when starting DAA with a multidisciplinary approach could improve health outcomes.


Statistical Analysis
Quantitative variables were compared by Student's t-test. Results were analyzed using the χ 2 -test for comparison between qualitative variables. In addition, different groups of patients according to anticholinergic burden were considered: 1-2 points (low anticholinergic burden) and ≥3 points (high anticholinergic burden) according to ADS, ACB and ARS scales. If data of anticholinergic risk was missing for one drug it was considered as no risk. If no anticholinergic risk data were available for a drug, it was considered non-risk (0 points). p < 0.05 was considered statistically significant.
No differences were observed among patients taking and not taking anticholinergic medication for the remaining demographic or clinical variables, including potential DDIs. Similarly, no differences were observed when comparing the groups with a high anticholinergic burden with those with a lower anticholinergic burden for any variable.
Drugs classified as nervous system drugs (ATC = N) were the most commonly considered inappropriate by the three scales and taken by the majority of patients. Figure 1 presents the proportion of patients with drug exposure for the principal pharmacologic classes of medications for each score. In addition to the HCV treatment, patients took an average number of 5 concomitant medication (SD 2.8 drugs), with 47% patients with moderate or excessive polypharmacy. The most used drugs were diuretics (53%), psycholeptics (37%), drugs for acid related disorders (36%) and agents acting on the rennin-angiotensin system (36%). Potential DDIs were present in 156 patients (66%).
No differences were observed among patients taking and not taking anticholinergic medication for the remaining demographic or clinical variables, including potential DDIs. Similarly, no differences were observed when comparing the groups with a high anticholinergic burden with those with a lower anticholinergic burden for any variable.
Drugs classified as nervous system drugs (ATC = N) were the most commonly considered inappropriate by the three scales and taken by the majority of patients. Figure 1 presents the proportion of patients with drug exposure for the principal pharmacologic classes of medications for each score.

Safety Outcomes
AEs were reported in 206 patients (87.3%). The most reported AEs were fatigue (97 patients, 41.1%), gastrointestinal symptoms (63 patients, 26.7%), skin complaints (53 patients, 22.5%) and anemia (48 patients, 20.3%). Nine patients required hospital admission during study period, and two of them died. Three patients discontinued treatment. Two patients died out-of-hospital and the cause of death was unrelated to the DAA therapy according to physician criteria. Only one patient discontinued treatment due to adverse events, after a 3D overdose that caused him malaise, edema and gastrointestinal complaints.
Only the ARS scale showed a significant relationship between the presence of anticholinergic medication and AEs, and high anticholinergic burden and number of AEs per patient (Tables 2-4). However, by using this scale, no differences were observed between taking and not taking anticholinergic medication, or when comparing the groups with a high anticholinergic burden with those with a lower anticholinergic burden, for hospital admission or death.
Regarding anticholinergic events, no differences were observed between taking and not taking anticholinergic medication, but patients with anticholinergic burden suffered more anticholinergic events, regardless of the scale used. The ADS scale also showed a significant relationship between high anticholinergic burden and rate of anticholinergic events per patient (Tables 2-4).
A total of 151 were treated with ribavirin, and suffered more AEs compared to patients without ribavirin, but the difference did not reach statistical significance (89.4% vs. 83.5%, p = 0.19). Dose reduction or discontinuation of the RBV daily dose was observed in 71 of them (47%).

Discussion
This real world population study shows the prevalence of anticholinergic medicines in older chronic hepatitis C patients with advanced liver disease and high comorbidity starting DAAs, by using three validated scales-the ADS, ACB and ARS [32][33][34] and analyses the results of these scales on adverse outcomes.
We found a high prevalence of patients taking anticholinergic drugs using ACB and ADS scales (35.2% and 34.3%). We also found a relationship between comorbidity and drugs with a greater anticholinergic effect. Patients with anticholinergic drugs had more comorbidity (CRG ≥ 6/05) [39]. The presence of moderate and excessive polypharmacy (≥5 and ≥10 drugs) has been found to be a risk factor for the presence of anticholinergic medication; data in concordance with those observed by Sevilla et al. [30]. Only the ARS scale showed significant relationship between anticholinergic burden and AEs, also in the rate of AEs per patient. The rate of anticholinergic events per patient was significantly higher in patients using anticholinergic drugs according to the three scales. Prevalence of patients taking anticholinergic drugs was statistically significantly higher (p < 0.0001) on the ACB and ADS scales (35.2% and 34.3%) than on the ARS scale (10.6%). Anticholinergic medications are considered inadequate for geriatric patients [39]. Mechanisms potentially explaining the association between anticholinergic medicines and adverse events include age-related changes in pharmacokinetics and pharmacodynamics, as well as increased permeability of blood-brain barriers [40]. Additionally, in chronic hepatitis C patients, advanced liver impairment might predispose to an increased risk of adverse effects from anticholinergic drugs.
Our data show similar results using ACB and ADS scales but different with ARS, in accordance with previous studies [30,31,41]. According to the ACB and ADS scales, 8.9% of patients had a high anticholinergic risk, statistically significantly higher than the 4.2% obtained using the ARS scale (p = 0.04). ACB scale includes high-ceiling diuretics (such as furosemide) in the group of anticholinergics, and our population has advanced hepatic disease, with extensive use of this kind of diuretics, resulting in a high use of anticholinergic medication. In addition, both ACB and ADS scale include paroxetine and quetiapine as moderate or high anticholinergic burden. Both drugs were taken by a large number of our study patients mainly because antipsychotics are widely used in geriatrics for different disorders and quetiapine and paroxetine, specifically, had been shown to be efficacious in generalized anxiety disorder, which is one of the most common psychiatric disorders in primary care [42,43]. In addition, estimates from the literature suggest that psychiatric disorders were common comorbidities among HCV-infected patients [44]. Moreover, anxiolytics and antidepressants have been associated with adverse outcomes in the elderly, such as falling and hospital admissions [45].
ARS is one of the tools associated with the highest number of patient-related outcomes, being associated with hospitalization, mortality, falls and functional decline [4]. By using the ARS scale, we found a relationship between the presence of anticholinergic drugs and the development of AEs in patients receiving DAAs. Although no significant differences were found, there also was a higher presence of hospital admission in patients with higher anticholinergic burden, according to ACB and ARS scales. These results are in accordance with Wan-Hsuan et al., who found an association between the ARS scale and all-cause admission to hospital. They also consider polypharmacy and anticholinergic burden as quality indicators of polypharmacy in older adults [46]. Hsu et al. also showed good response relationships between anticholinergic burden using the ACB scale and a variety of adverse outcomes in older adults. This study shows that anticholinergic scales tend to be an indicator of prescriptions with a risk for AEs, also described by Jean-Bart et al [45]. The prolonged and cumulative administration of these drugs makes them especially vulnerable to anticholinergic adverse effects because of the advanced age and frailty [47]. Anticholinergic burden has also shown to be a good predictor of adverse health and functional outcomes [48][49][50][51][52]. Hence, anticholinergics are generally categorized as potentially inappropriate medications for use in older adults and the estimation of the anticholinergic drug burden has been suggested as a way of reducing the risk of secondary cognitive decline of drug therapy and of optimizing polypharmacy in the elderly [53,54]. In our study, mortality was slightly higher in patients taking medication with anticholinergic burden according to ADS scale. A recent meta-analysis indicated an association between anticholinergic exposure and higher risk of mortality using the ACB, ARS and ADS scales [12]. Recently, Lozano-Ortega et al. described that the ACB and ADS scales were well suited for implementation in observational studies where anticholinergic exposure needs to be quantified [10].
Antiviral treatment was generally well tolerated, with only one patient discontinuing treatment due to adverse events. Nevertheless, two patients died during treatment even though the cause of death was unrelated to the DAA therapy. Fatigue was the top reported AE, in concordance with the data observed by Huang et al. [55]. In their study, DAA were found to have higher reporting rates in a few AEs-e.g., fatigue and abdominal pain-using data from the U.S. Food and Drug Administration Adverse Event Reporting System, and diarrhea, abdominal pain, pain and pyrexia using data of Electronic Health Records [55]. In our study population, the prevalence of skin complaints and gastrointestinal effects was higher than that found by Villani et al. in a recent meta-analysis [24]. Anemia was frequent in our population, which is typically associated with RBV use, similar to other studies [14,24,56]. In 47% of patients, RBV dose reduction or discontinuation was needed, which is a higher proportion of cases than reported by Conti et al. [15].
Although the rate of DDIs in our study was quite high (66%), there were no clinically significant interactions related to AEs, hospital admission nor death. It was in part because a meticulous DDIs assessment before treatment initiation and careful monitoring were realized by the multidisciplinary team to avoid DDIs.
Possible anticholinergic effects were observed in a large number of patients, in almost half of the patients who suffered any adverse event. Effects probably related to dry skin, confusion and restlessness were the most observed. It should be noted that effects related to dry skin may also be as a result of the antiviral treatment described above as skin complaints. Although no relationship was seen between anticholinergic effects and anticholinergic burden according to any of the scales, a significant relationship was observed between the number of anticholinergic effects and the anticholinergic burden with all scales.
The main strengths of our study are the analysis of polypharmacy, its appropriateness using anticholinergic scales, and safety-related outcomes in a real-world population of older patients receiving DAA therapy including a large number of cirrhotic patients (n = 236 patients) with high comorbidity, polypharmacy and, therefore, risk of DDIs. To our knowledge, this is the first study assessing the appropriateness of concomitant therapy and its clinical safety outcomes in older chronic hepatitis C patients receiving DAAs. No studies are available on HCV older patients receiving DAAs to compare the results on the use of anticholinergic agents. Variability between scales in the proportion of patients identified as taking anticholinergic drugs and the grade of anticholinergic burden was found. As other authors have suggested, the use of multiple anticholinergic burden measures on the same population reduces risk from heterogenicity and increases confidence in making comparisons [12,41]. Nevertheless, the study has some limitations. Firstly, its retrospective nature may imply some selection bias. Since it includes out-hospital patients, even though all medicines dispensed during the study period were included in the analyses, it could not be ascertained whether the dispensed medicines were actually consumed. Secondly, our findings only indicate association, and were susceptible to residual confounding. Casual relation between anticholinergic burden and AEs was not proven. Finally, electronic clinical records data are collected based on routine medical practice, not for pharmacovigilance research, and some AEs, such as fatigue, can be caused by HCV infection itself, or other long-term AEs may not be observed due to the relatively short follow-up time.
In concordance with Park et al., our study shows differences in the total prevalence of anticholinergic use evaluated by the different scales, and there is no standardized rating scale for the measurement of anticholinergic burden; therefore, further research is necessary to develop a useful and comprehensive tool identifying medications with anticholinergic properties [54]. However, as suggested by Hanlon et al., anticholinergic risk scales are easy and useful for identifying patients at risk of adverse effects, regardless of the scale used [41]. Sessa et al., found a proportion of preventable adverse drug reactions involving DAAs, suggesting that it would be a target for improvement [56]. Anticholinergic medications are a potentially modifiable risk factor for the prevention of adverse events and one may hypothesize that recognizing the use of anticholinergic drugs-and therefore potentially inappropriate polypharmacy-by means of these scales, could help in identifying older patients with comorbidities at risk of adverse events when starting antiviral therapy [57]. As suggested by Merle et al., limiting drug prescription to essential medications and periodically re-evaluating all use of drugs in the elderly could reduce the prevalence of AEs [58]. The study highlights the need to revise concomitant conditions and the treatment of hepatitis C chronic patients, and therapy initiation presents a window of opportunity where a multidisciplinary could make patient-centered decisions. Anticholinergic burden tools-probably ARS as the most clinically relevant-might be recommended as a complementary procedure to comprehensive geriatric assessment, with a multidisciplinary approach in patients starting any treatment, mainly chronic hepatitis C patients with advanced liver disease, high comorbidity, polypharmacy and risk of DDIs, when starting DAAs [4].
In fact, HCV therapy initiation presents a window of opportunity for overall treatment review-particularly for those prescribed multiple medicines, or taking combinations of medicines with a higher risk of adverse effects, including enhanced coordination of care between hepatologists, clinical pharmacists and other subspecialists.
In conclusion, older hepatitis C chronic patients commonly had multiple comorbidities and used co-medications with potential anticholinergic effects, therefore they are exposed to inappropriate polypharmacy. The presence of anticholinergic drugs was associated with AEs using ARS scale. The rate of anticholinergic effects per patient was significantly higher in patients with anticholinergic drugs, regardless of the scale used. Clinicians treating older adults starting DAA should be aware of the risk associated with comorbidity and comedications that may increase the risk of AEs. To provide optimal antiviral treatment, a coordination of care between hepatologists and clinical pharmacists supported by a multidisciplinary team is needed.