Magic bullets for insomnia? Patients' use and experiences of newer (Z drugs) versus older (benzodiazepine) hypnotics for sleep problems in primary care

INTRODUCTION

Insomnia is a common, often chronic condition that increases with age and has a reported prevalence rate in Europe ranging from 4% to 37%.1^,2 About half of those with sleep problems seek medical help,3 which often involves a prescription of hypnotic drugs including benzodiazepines like temazepam, or Z drugs such as zopiclone, zolpidem, or zaleplon. Most hypnotic prescribing takes place in primary care, and drug treatments may be inappropriately prescribed for 4 weeks or longer in up to 50% of new prescriptions.4

Over the past decade, a gradual reduction has occurred in the prescribing of older benzodiazepine hypnotics, while the use and cost of hypnotic drugs continues to rise overall.5 This has been due to fears over benzodiazepine use and abuse and an increase in prescribing of Z drugs, which were marketed as safer and less liable to dependence compared with benzodiazepines.6 Zopiclone is now the most frequently prescribed hypnotic in the UK at 4 million items (39% of items) costing £10m (43% of total hypnotic cost). Temazepam is the next most commonly prescribed with 3.5 million items (35%) at £4m (19%).7 As a result, over £22m is spent every year in primary care on 10 million items of hypnotic drugs, and these figures have shown no decline since the early 1990s. This is partly explained by an increasing population, with an average growth per year of 0.4–0.5% since 1991 and an increasing proportion of older people.8

Previous research has documented the attitudes of patients and doctors to benzodiazepines,9^–15 but little has so far been published about their perceptions or experiences of Z drugs, either alone or compared with benzodiazepine hypnotics. Studies of benzodiazepine hypnotics have shown that patients believe that they are more effective and safer than do doctors.14^,15 A recent study of GPs showed that, compared with benzodiazepines, they believed Z drugs to be safer, more effective, less liable to cause side effects, and the drugs of choice for a range of indications,16 despite a lack of evidence to support such notions.17

The clinical benefits of hypnotics have actually been shown to be small, with significant risks of complications arising from adverse cognitive or psychomotor effects, and daytime sleepiness that may persist for several months after stopping the drug.18 Such unintended reactions are reported most frequently by older patients, who are also the most likely to be prescribed the newer drugs. Complications, such as falls, fractures, and road traffic collisions, have been linked to these drugs,19 which have considerable potential for tolerance and addiction.

The aim of this community survey was to investigate and compare patients' perceptions of benefits and risks of benzodiazepines and Z drugs. The study formed part of a larger investigation into methods for reducing hypnotic prescribing in primary care, and also aimed to define potential interventions for better management of sleep appropriate to the primary care setting.

METHOD

West Lincolnshire Primary Care Trust (PCT) comprised 40 general practices (now part of Lincolnshire Teaching Primary Care Trust), serving 214 000 patients. Previous attempts to lower relatively high rates of hypnotic prescribing encountered resistance to change. A survey instrument was designed to collect data using a previously published questionnaire,14 incorporating elements from literature searches, and discussion points raised within the project steering group, as well as advice received from experts in the field.

In 2005, forms were posted to a sample of patients on the lists of GP principals in West Lincolnshire PCT. Selection was made by each participating practice being asked to submit a list of patients who had been prescribed a Z drug or benzodiazepine taken in a single night-time dose to induce sleep in the previous 6 months. From this list of names, a random sample of 50 patients from each practice received the questionnaire for completion. The questionnaire focused on patients' attributes, views about indications for drug treatments (for example, insomnia or anxiety), and their assessment of outcomes.

Returned questionnaires were entered into a spreadsheet according to a predetermined coding frame. SPSS (version 12.1) was used for data analysis. A χ² test was used for initial group correlations and logistic regression (both backward conditional and forward conditional) to analyse differences in responses for Z drugs and hypnotic benzodiazepines correcting for age, sex, and duration of drug use. Analysis was restricted to patients who stated that they were taking these drugs for insomnia rather than for other indications, such as musculoskeletal pain or epilepsy. Missing data were not included in the comparisons.

RESULTS

Altogether, 935 of 1600 (58.4%) analysable responses were received from responders who had been prescribed a benzodiazepine or Z drug in the previous 6 months. Thirty-two of the 40 practices surveyed contributed patients to the survey, and the response rate from each varied considerably (mean 44.5%, standard deviation [SD] = 20%). Of the responders, most (705 of 935; 75.4%) were taking their drugs for insomnia rather than other indications, such as anxiety, epilepsy, or drug addiction (Table 1). Further analyses were carried out only for those confirming that they had taken drugs for insomnia. Z drugs were taken more commonly (370 of 705; 52.5%) than benzodiazepines (268 of 705; 38.0%), reflecting current prescribing trends. Hypnotics prescribed included zopiclone (328 responders; 46.5%), zolpidem (39; 5.5%), zaleplon (3; 0.4%), temazepam (161; 22.8%), nitrazepam (48; 6.8%), diazepam (46; 6.5%), or other benzodiazepines (13; 1.8%). Just over half (50.4%; 355 of 705) of responders were aged 65 years of age or older.

The majority of participants, 87.9% (n = 705), were first prescribed a hypnotic by their GP, rather than a psychiatrist or other health professional. Many had been advised to continue treatment for longer than recommended under the licence: 45.4% (n = 320) indicated they were advised to continue treatment for a month or more, and a further 42.3% (n = 298) reported not being advised on treatment duration. Most responders, 92.1% (n = 649), were on repeat prescriptions of hypnotics, and over two-thirds were taking hypnotics daily rather than intermittently (Table 1).

In a number of important respects those prescribed Z drugs differed from those taking benzodiazepines (Table 2); they were more likely to be younger, to have received their first prescription from a psychiatrist, stated a wish to come off hypnotics, and made at least one attempt to come off medication. These differences were significant at P<0.05 (but not at P<0.01). Over two-fifths of responders (41.8%; n = 295) reported at least one side-effect from hypnotic use. Withdrawal effects were commonly reported among those who had tried to stop their treatment, even if they were on intermittent rather than daily therapy. Responders who felt they may be at risk of becoming drug dependent were more likely to want to stop medication than those who already acknowledged dependence or felt that they were not dependent on hypnotics (P<0.001).

No significant differences between Z drugs and benzodiazepines were found in respect of perceived benefits or adverse effects, including withdrawal or dependence (Table 3). Almost one in five patients (18.6% overall) expressed a desire to come off hypnotic medication. Patients on Z drugs were more likely to wish to stop (22.7% versus 12.3%; odds ratio [OR] = 1.67; 95% confidence interval [CI] = 1.13 to 2.49), and to have attempted to come off medication than those on benzodiazepines (52.4% versus 41.0%; OR = 1.54, 95% CI = 1.12 to 2.12).

Most patients using hypnotics stated they took less time to get to sleep (Z drugs versus benzodiazepines, 76.2% versus 79.1%), and woke less during the night (60.8% versus 55.6%), but half the users or fewer agreed that they slept longer (48.6% versus 47.8%), felt rested on waking (47.0% versus 45.1%), were more active (37.0% versus 35.8%), or felt better overall (48.1% versus 50.0%); these differences were not significant. Side-effects including daytime drowsiness (20.5% versus 19.0%), headache (18.1% versus 14.9%), dizziness (13.5% versus 17.2%), difficulty concentrating (16.7% versus 14.2%), or difficulty thinking (14.9% versus 14.6%), confusion (11.6 versus 7.8%), shaking (9.2% versus 7.5%), and falls (5.9% versus 7.8%) were not significantly different between those on Z drugs or benzodiazepines (Appendix 1).

DISCUSSION