Trajectories of Use in Elderly Incident Benzodiazepine Users of Norway

Background: The use of benzodiazepines (BZD) is associated with risk of addiction and benzodiazepine users frequently co-medicate with opioids. Trajectories of use and how these is associated with opioid use has rarely been studied before.

Objectives: To characterize trajectories of use amongst elderly new benzodiazepine users and to describe how the use of opioids change between the different trajectories.

Methods: The study population consists of all elderly (aged 65 and older) in Norway who received at least one prescription of benzodiazepines in 2013.

The group based trajectory model (GBTM) uses 360 days of dispensed prescriptions after the initial prescription. The data was divided into twelve 30 days periods and the number of DDD dispensed in each period was calculated. These numbers were normalised by dividing each period with the number of DDDs of the first period. Three groups were created using quadratic terms.

Exposure to BZD is defined as any prescription dispensed in ATC class N05A. Incident use is defined as no use in the previous 12 months. Exposure to opioids is defined as any prescription dispensed in ATC class N02A. Statin exposure was used as a control and is defined as dispensed prescriptions in ATC class C10AA.

Change in use is measured as the number of DDD dispensed in the three months preceding the incident benzodiazepine dispensation compared with the the number of DDD dispensed in the final three months of the follow-up window.

Results: 54.8% of incident BZD users (33293 persons) in 2013 had only one prescription dispensed. Of the elderly with more than one prescription GBTM created three groups. The three groups represent increasing, continuous and decreasing use of BZD. 3.5, 16.4 and 80.1 percent of users belonged to the three groups respectively. Increased use led to a 15% increase in use of opioids (from 15.8 to 18.1 DDD) For the continuous group opioids increased by 6% from 14.7 to 15.6 DDD while for the decreasing and one prescription group there was a decrease in opioid use. By contrast statin use decreased for all groups.

Conclusions: A small but clinically relevant proportion of new BZD users has a trajectory of use corresponding to increasing use. Among these users a large proportion also use opioids and this use increased during the first 12 months of BZD use. These users should be targeted by interventions to avoid misuse and abuse.