Built on the 3213 prescriptions from the 308 enrolled older adults, 96.8% cases had at least 1 PIM and lorazepam was the most frequently prescribed drug. Co-administration of 3 or more CNS active drugs was the principal DDI and the most frequent DDI which should be avoided was the co-administration of lorazepam and olanzapine. Constipation was the most common ADR. The co-administration of quetiapine and potassium chloride was an independent risk factor of constipation. Moreover, the co-administration of quetiapine and potassium chloride was an independent risk factor of MCID. While the co-administration of quetiapine and escitalopram was a protective factor of MCID.
The prevalence rate of PIM for older adults with mental illness in this study is much higher than that in other studies. For community-dwelling older adults, the PIM rate ranged from 48.3% to 53%31, 32. For aged patients with cancer, the PIM rate ranged from 47.3% to 73%33, 34. Murphy35 found that 55.8% older adults with mild-to-moderate Alzheimer disease were prescribed a PIM with 30.1% being prescribed 2 and more PIMs in nine European countries. For older adults with diabetes, from 56% to 66.2% cases had at least 1 PIM36, 37. Zahwe38 showed that 80.0% of patients with heart failure were taking at least 1 PIM. While, the result showed in this study is a little higher than that (91.04%) in older patients with depressive disorder12. Differences in the PIM prevalence rates may be related to three reasons. First, three criteria were used as supplements in this study in order to screen all the PIM considering the common drugs used in different countries and the different screening way. Secondly, enrolled patients suffered from diverse diseases so that the drug categories they used were different in different studies. CNS active drugs which have a high proportion based on the three criteria are the primary drugs used for this group patients. Thirdly, study population in studies was in different age ranges and different settings.
Lorazepam was the most commonly prescribed PIM for older adults in this study. It is in line with other studies. Motter15 found that benzodiazepines and antihistamines were the most common medication classes reported followed by tricyclic antidepressants for older people. Huang39 concluded that benzodiazepines, anticholinergics and antipsychotics were the most frequently prescribed classes. Murphy35 found that benzodiazepines, SSRIs and Proton Pump Inhibitors (PPIs) were the most frequent PIMs. Roux31 found that benzodiazepines, PPIs, antipsychotics, antidepressants were the most prevalent PIMs.
Benzodiazepines prescribing to aged people is significantly in excess of what the available evidence suggests is appropriate40. Benzodiazepine use was higher in those with psychiatric disorders as people who had a mood disorder were the most vulnerable41. Psychiatrists had higher rates of benzodiazepine prescription42. That may be explained by the need for benzodiazepines, the lack of effective alternative therapies, the attitude of patients and physicians. As many older adults suffer from insomnia and depression, benzodiazepines are the most frequently used to relieve these symptoms43. Nonbenzodiazepine sedatives have similar risks to those of benzodiazepines and are not suitable as alternative drugs to benzodiazepines according to those PIM criteria. Antihistamines, antipsychotics, melatonin, phytotherapeutics, homeopathy and acupuncture are just not recommended for insomnia treatment, and light therapy and exercise have yet to be evaluated44.
Co-administration of lorazepam and olanzapine is the most common DDI for older adults. Madhusoodanan45 conducted a pilot study and found co-administration of lorazepam and olanzapine caused no adverse consequences. Bergemann46 discovered that the co-administration with lorazepam, the dose-corrected olanzapine plasma concentration was no different from the plasma levels under olanzapine monotherapy. While, a case report showed IM olanzapine and IM lorazepam would lower blood pressure and cause dizziness47. Lee48, 49 revealed that olanzapine and quetiapine inhibited the human ether-a-go-go related gene channels current, and Kowalchuk50 found olanzapine inhibited central ATP-sensitive potassium channel.
Constipation, hepatotoxicity and extrapyramidal symptoms are common ADRs in this analysis which are consistent with other literature. Constipation is a common side effect of antipsychotics and it can lead to serious consequences such as paralytic ileus, bowel occlusion and death51.
Besides, antipsychotics and antidepressants are potential cause of hepatotoxicity, even at therapeutic doses52. Drugs which are substrates of CYP450 enzymes have a higher likelihood of causing drug-induced liver injury (DILI) in a dose independent manner, while drugs which are CYP450 inhibitors have a higher likelihood of generating DILI only when they are administered at high daily doses53. Second-generation antipsychotics could lead to extrapyramidal syndrome with risk factors such as the choice of a particular antipsychotics, high doses, history of previous extrapyramidal symptoms, and comorbidity54. Moreover, all antipsychotics and antidepressants are commonly responsible for ADRs such as dystonia, akathisia and pseudoparkinsonism55. Benzodiazepines and Z-drugs are associated with motor vehicle accidents, falls and fractures, in addition, they may be related to dementia, infections or cancers56.
However, three combinations of drugs except for co-administration of quetiapine and potassium chloride which should be avoided according to the database were neither independent risk factors for each ADR nor showing clinically significant. The probable reasons are following. First, it may be due to the pharmacokinetic multiple pathways of drug metabolism and the DDIs are mainly speculated based on the drugs pharmacokinetics features26. Psychiatric patients are at high risk of pharmacokinetic DDIs and the rate of DDI in patients with cytochrome P450 (CYP) inhibitor or inducer are close to 50%57. Psychotropic drugs are usually metabolized by several enzymes which would reduce the risk of DDI26. Second, the route of administration may affect DDIs and ADRs. DDI may be lower in the oral administration of olanzapine and lorazepam according to the database. In this study, antipsychotics, antidepressants and benzodiazepines were all oral administration and lowered the ADRs. Thirdly, each ADR is the result of a combination of multiple factors and other potential risk factors were neglected. For example, QT interval prolongation and/or torsade de Pointes occur only in the presence of multiple risk factors, such as 65 years old or above, pre-existing cardiovascular disease, bradycardia, female, hypokalemia, hypomagnesemia, a supratherapeutic or toxic serum concentration, or the simultaneous administration of other drugs which delay repolarization or interfere with drug metabolism58. Fourth, the data is biased. For example, there were no ADRs among all patients who used quetiapine in combination with escitalopram.
PIMs are associated with patients’ and doctors’ awareness. For patients, they may not be aware if they are taking PIMs32 or want to try non-pharmacal methods43. As for physicians, they may lack of the necessary vigilance of PIMs in older adults or lack of skills and training on how to regulate benzodiazepines43, 59. Some doctors worry about the doctor-patient relationship and the stress for an older patient with a limited life expectancy if deprescribing benzodiazepines60.
Deprescribing appear to be feasible and safe to solve the problem of PIM61. Rognstad62 conducted an effective educational intervention for general practitioners to reduce the prevalence of PIM.
Deprescribing is suggested to reduce drugs one at a time after careful assessment63. After pharmacist-led deprescribing intervention, patients’ symptoms and healthcare expenditures were reduced34. Furthermore, raising older adults’ awareness about PIMs through education is needed32. Future trials should focus on how to improve the understanding of both doctors and patients on PIM and DDI, the pharmacokinetic and pharmacodynamic mechanism of those CNS active drugs, and the influence of different drug administration routes on DDIs.
Strengths and limitations
There are two strengths in this study. First, we discussed DDIs and ADRs from the perspective of statistical significance and clinical significance. Considering clinical medication safety, we analyzed the relationship between co-administration of drugs and ADRs. Second, we attach great importance to a comprehensive investigation of PIM. Considering the differences in commonly used drugs in diverse countries and the focus of different criteria. We used three criteria as supplements to detect PIM as comprehensively as possible.
Our work consists of two limitations. First of all, we focused on elderly patients with a major diagnosis of psychiatric disorders and central nervous system drugs, and ignored PIMs of other systems and the dosage reduced with varying levels of kidney function based on STOPP version 2. Next, we extracted information from the long-term physicians’ order and ignored the temporary physician’s order.