Multimorbidity and use of hypnotic and anxiolytic drugs: cross-sectional and follow-up study in primary healthcare in Iceland

In recent years the extensive research on the epidemiology of multimorbidity in several countries has revealed that it is an extremely common problem, not least in primary healthcare where the majority of patients have more than one problem at any given time [1‐3]. The percentage of patients with multiple diagnoses increases as people get older. However, the extent of multimorbidity in young people and even in children is perhaps greater than might be expected, probably because risk factors for multiple diseases are not randomly distributed [1, 4]. Multimorbidity can be measured in several different ways and the range of diseases included varies between studies. A recent study with health data classified according to ICPC-2 chapters (International Classification of Primary Care, Version 2), ICD-10 chapters (International Classification of Diseases, 10th edition) or CIRS domains (Cumulative Illness Rating Scale) found that when multimorbidity is defined as two or more disease entities, prevalence estimates are similar, no matter what diagnostic classification system is used [5].

Multimorbidity has been associated with polypharmacy [6] and lower socioeconomic status [7]. Studies have shown that clustering of diseases or conditions in persons with multimorbidity is complex and does not follow any typology [8‐10]. This could indicate a common root of chronic biological dysregulation in the body, better described as allostatic overload, i.e. a physiological scenario where the organism’s adaptive and restorative capacity is overtaxed [8, 10].

A number of diseases are correlated with insomnia and thus increasing multimorbidity might lead to an increase in sleep disorders. Hypnotic and anxiolytic benzodiazepine derivatives are commonly prescribed to people with a sleep disorder often just labelled with the diagnosis “insomnia” in the medical records. However, in epidemiological studies on prescribing habits it is not always obvious what type of health problems lie behind or are possibly connected with prescribing hypnotic and anxiolytic drugs. This can be of great importance in the primary care setting as a holistic view of a patient’s health problems is usually put in focus.

Since the launching of chlordiazepoxide in 1960 and diazepam a few years later the number of benzodiazepine derivatives and related drugs released on the market has multiplied and their use both as anxiolytics and hypnotics increased dramatically [11]. Now they are some of the most widely prescribed drugs. This has caused much controversy with widely divided opinions, e.g. concerning adverse effects, dependence, abuse liability, efficacy and tolerability [11]. The serious side effects of benzodiazepines have been known for decades [12], and in recent years, concerns have been raised about possible links between the use of these drugs and increased mortality [13‐16]. Nevertheless, an incontestable evidence of a causal relationship has not yet been presented, and findings in other studies have not lent support to a strong association of hypnotic use with excess mortality [17, 18]. A recently published prospective study found that the risk of mortality was not significantly associated with hypnotic use regardless of type and duration [19] while a retrospective cohort study which also corrected for a range of potential confounders found an association with significantly increased risk of mortality [20]. In a newly published study connecting disease burden and use of hypnotics, it was shown that there was a 50 % increase in the use of benzodiazepines 24 months before death. It was concluded that the association between use of these drugs and mortality is at least partially due to confounding [21]. Apart from the well-known risk of falls and fractures, there are other untoward effects such as a possible association with increased suicide risk [15, 22], increased risk of road traffic accidents [23], dementia [24] and Alzheimer’s disease [25]. Besides, the effectiveness of the newer type of hypnotics, the so-called Z drugs (zolpidem, zopiclone) has been questioned [26]. Thus, there are numerous risks associated with the use of these drugs which to some degree might be avoided by using other treatment options for insomnia instead of prescribing hypnotics. Sleep hygiene recommendations could be applied as first-line therapy [27] followed by interventions needing more specialised skills such as cognitive behavioural therapy [28].

The question might be raised whether patients with a heavy disease burden are more prone to becoming chronic users as sleeping problems leading to hypnotics being prescribed, may be a direct consequence of multimorbidity, although multimorbidity as such is not an indication for the prescription of these drugs. In a paper on the medicalisation of sleeplessness it was shown that during the study period insomnia diagnoses increased more than sevenfold while prescriptions for non-benzodiazepine sedative hypnotics (Z-drugs and ramelteon) increased about thirtyfold [29]. It is known that the use of benzodiazepine derivatives and related drugs increases with age [11] and also that the prevalence of multimorbidity increases significantly with age both in men and women [5‐7]. It is also worth noting that in a study comparing measures of multimorbidity to predict outcomes in primary care, the number of drugs prescribed was a powerful predictor of mortality [30]. There have not been many studies on the relationship between selected drug classes and multimorbidity, but it has been shown that comorbidity was a correlate of benzodiazepine use in patients with depressive and/or anxiety disorders [31]. Although several epidemiological studies on multimorbidity have been done, there are still many countries, including Iceland until now, where little is still known about the prevalence of multimorbidity.

The aim of this study was firstly to analyse the prevalence of multimorbidity in primary healthcare, and secondly to analyse the prevalence and incidence of prescriptions of hypnotics and anxiolytics, mostly benzodiazepine derivatives and Z-drugs, in order to find a possible association between the use of these drugs and the prevalence of multimorbidity, based on the assumption that sleeping problems and disease discomfort related to multimorbidity might lead to prescription of these drugs although a clear-cut indication might be lacking. Thirdly, our aim was to find if incident patients took the drugs according to clinical guidelines (2–4 weeks) or if they had become either intermittent or regular users, thus indicating the possibility that multimorbidity might be a contributing factor in the chronic use of these drugs.

A total of 221,822 patients contacted the primary healthcare centres during the period 1 Jan 2009 to 31 Dec 2012 or approximately two-thirds of the Icelandic population, 83 % being local residents. Table 2 shows the subset of patients with multimorbidity compared with the total cohort. Approximately 47 % of the total population did not have any relevant chronic condition as defined in Table 1, nearly 18 % had one chronic condition and 35 % were considered as having multimorbidity. The prevalence increased with age in both sexes up to 79 years of age.

During this four-year period a total of 30,926 patients were prescribed hypnotics/anxiolytics (prevalence: 13.9 %; 10.1 % for men and 17.5 % for women). The prevalence of hypnotics/anxiolytics prescriptions stratified for age and sex is shown in Table 3. It increases steadily with increasing age in both sexes until the highest age group is reached where it drops slightly.

In 2011 there were 4305 patients who had not been prescribed hypnotics/anxiolytics the two previous years who initiated treatment for insomnia or anxiety with a new prescription, hence categorised as incident patients. During a 12 month period approximately 56 % received one prescription only, 33 % received 2–4 prescriptions, and nearly 11 % of the patients were issued 5 prescriptions or more.

Results for the incidence of hypnotics/anxiolytics prescriptions per 1000 persons per year in 2011 according to age and sex are shown in Table 4. The incidence increased with age and was higher in women than in men. Table 5 shows the division into 1 prescription, 2–4 prescriptions and 5+ prescriptions categories, both for all incident patients and the subset of incident multimorbid patients. The proportion of incident patients who received either one, 2–4 or 5+ prescriptions respectively during the 12 month period are plotted in Fig. 1.

Figure 2 illustrates some aspects of the clinical profile among the multimorbid incident patients subdivided into those who are either one time, intermittent or regular users. As shown, the combination of mental disorders and pain related diagnoses dominate the patient profile. Most often the use of these drugs was sporadic.

In the group of patients who were prescribed hypnotics/anxiolytics (30,926), 3 % (1018) had no chronic condition, 12 % (3693) had only one chronic condition while 85 % (26,215) were categorised as being multimorbid. Their distribution according to age, sex and number of conditions can be seen in Table 6 with the ORs for the likelihood of being prescribed a hypnotic and/or an anxiolytic if multimorbid. Figure 3 shows this group according to age and number of conditions and Fig. 4 shows the ORs by age and sex.

In this four-year period the total number of anxiolytics prescribed was 1.92 million defined daily doses (DDDs) [34], thereof 93 % benzodiazepines. During the same time 5.87 million DDDs of hypnotics were prescribed, 11 % benzodiazepines and nearly 88 % Z-drugs, hence in the combined group of hypnotics/anxiolytics, drugs with benzodiazepine structure were 31 % and Z-drugs 66 %.

When the ratio of multimorbid patients prescribed hypnotics/anxiolytics to the multimorbid patients in general is calculated it increases steadily as the number of chronic diseases within one patient increases (Table 7).

Of the 143,662 patients in the PHCCA cohort who did not have multimorbidity, 4711 were prescribed a hypnotic and/or an anxiolytic. On the other hand, 26,215 patients out of 78,160 multimorbid patients were prescribed these drugs. Thus, patients with multimorbidity were much more likely to be prescribed a hypnotic and/or an anxiolytic compared with patients without multimorbidity, OR = 14.9 (95 % CI = 14.4–15.4). Looking at men and women separately, women were more likely than men to be prescribed these drugs, OR = 16.0 (95 % CI = 15.4–16.7) for women versus OR = 12.7 (95 % CI = 12.0–13.3) for men, increasing with age, as shown in Fig. 4.

The mean age of patients who contacted the primary healthcare centres of the capital area was 36.4 years (men 35.7, women 37.0) compared with mean age of 36.8 years (men 36.1, women 37.4) for the total Icelandic population.

This study reveals that in Iceland, as in many other countries, both the prevalence and the incidence of hypnotic/anxiolytic prescriptions are high. A considerable number of incident patients were prescribed benzodiazepines and Z-drugs for several times the recommended duration of therapy which could develop into long-term drug taking by these patients for an extended period, inconsistent with accepted medical practice. Solely explaining this by linear thinking, i.e. that “insomnia” leads to prescription of these drugs is probably simplistic, as the majority of patients prescribed these drugs are multimorbid having several chronic conditions, mainly mental and pain related diseases, which could lead to sleeping problems. There is a variety of concern with long-term use of these drugs as mentioned earlier in this paper, and their long-term efficacy has not been clearly established with some doubt having been cast on their short-term efficacy as well. Doctors should be urged to focus their attention to patients’ problems in the context of their multimorbidity and to be cautious when prescribing hypnotics for patients with multiple chronic conditions, bearing in mind that a holistic person focused approach is needed [4, 8, 41, 42]. They should be encouraged to discuss the duration of therapy with their patients before initiating benzodiazepine therapy in order to enable informed decisions about when these drugs should be discontinued. Multimorbidity per se is not directly an indication for prescribing hypnotics, and sleeping problems caused by the disease discomfort might in many instances be solved with non-pharmacological methods. It would be wise for family physicians to acquire skills in these forms of treatment.

ATC, anatomical therapeutic chemical classification; CI, confidence interval; CIRS, cumulative illness rating scale; DDD, defined daily dose; GP, general practitioner; ICD-10, international classification of diseases, 10^th edition; ICPC-2, international classification of primary care, version 2; ID, unique personal identifier; OR, odds ratio; PHCCA, primary health care of the capital area; Z-drugs, Benzodiazepine-related drugs. ATC-classification: N05CF