Abstract
See Review by Moylan et al., 2012, 46(3): 212–224
Moylan et al. (2012) reviewed the current status of alprazolam use and its role in the treatment of panic disorder. They found evidence that alprazolam use may be increasing and questioned whether clinical practice guidelines are being followed. Of how much concern should this be?
Moylan et al. (2012) estimated the means of 1.81 prescriptions per person meeting the criteria for panic disorder in 1997, and 1.50 prescriptions per person in 2007. This would appear to be a decrease. However, a number of other factors need to be considered. First, as the authors noted, the surveys in 1997 and 2007 used different criteria for panic disorder, and the 2007 survey screened for more symptoms. It has been noted that asking more screening questions about symptoms is likely to yield higher rates of endorsement of symptoms contributing to the criteria, tending to inflate the rate. The second factor identified by the authors, a more lenient method of applying hierarchical rules in 2007, might also have acted to inflate the prevalence. However, the 2007 prevalence rates are closer to those reported in epidemiological surveys in Europe and the USA, suggesting that the 1997 survey may rather have underestimated the prevalence. So it might be reasonable to conclude that the real prevalence of panic disorder in 1997 was similar to that in 2007, that is, about 2.6%. If this was the case, then the rate of alprazolam prescriptions at that time would have been 0.9 per person with panic disorder, suggesting the possibility of a quite marked increase over the subsequent decade.
Alprazolam prescription is reported to have increased from 3.88 to 5.12 defined daily doses (DDD) per 1000 population/day from 1997 to 2007, implying an increase from approximately 4 mg per 1000 population/day to 5 mg per 1000 population/day (Moylan et al., 2012). The DDD of 1 mg is lower than doses of alprazolam used on average in clinical practice and also lower than those used in many studies, including the Cross-National Collaborative Panic Study, Phase Two (flexible dosing to reach 6 mg per day; Cross-National Collaborative Panic Study, Second Phase Investigators, 1992), suggesting that multiples of the DDD are not necessarily cause for concern.
The fact that rates of prescription of other benzodiazepines do not appear to have changed much would suggest that the figures reflect a true increase in alprazolam rather than the possibility that alprazolam has replaced other benzodiazepines whose use has declined. An interpretation of this combined data on prescriptions per person and DDD prescribed might be that persons with panic disorder are now being prescribed alprazolam at slightly higher doses and for a longer period of time than was the case in 1997.
Is this of concern? The pharmacokinetic profile of alprazolam allows it to produce fairly rapid anxiolysis without significant sedation. In behavioural terms, relief of anxiety that occurs rapidly and closely contingent on the action taken (such as avoidance or ingestion of alprazolam) is likely to strongly reinforce repeated use of that strategy, perhaps more so than use of a benzodiazepine which produces more gradual anxiolysis. The short half-life of alprazolam results in a commensurately quick offset of action, when fear of another panic attack, if not symptoms of anxiety, may recur or, with chronic use, rebound. Withdrawal symptoms occurring even between the last dose of one day and the first dose of the next day are not uncommon. Thus the classic scenario in which a substance is taken to ameliorate the symptoms of its withdrawal occurs. This may be a particular problem with alprazolam compared to longer acting benzodiazepines.
There are pharmacological alternatives to benzodiazepines. Antidepressants have established efficacy in panic disorder (Baldwin et al., 2005) whereas beta blockers, although frequently prescribed, do not have robust evidence of efficacy. Evidence continues to accrue that relapse rates are high following discontinuation of pharmacotherapy, even after long periods of active treatment (Choy et al., 2007), but that the maintenance of gains is more likely following cognitive behavioural therapy (CBT). Although meta-analysis has encountered significant problems due to numerous sources of heterogeneity and methodological limitations in the research literature, there is reasonable evidence to suggest that antidepressant pharmacotherapy alone is less effective than a combination of antidepressant and CBT, and that antidepressants alone are likely to be less effective than CBT alone (Bandelow et al., 2007; Watanabe et al., 2007). Due to the paucity of studies comparing benzodiazepines alone and in combination with CBT, no reliable conclusions can be drawn. It seems likely that current clinical practice guidelines for panic disorder, in their preference for psychological therapy or the combination of psychological and antidepressant therapy in the treatment of panic disorder, have probably taken into account the findings from antidepressant studies and the problems of dependence and cognitive impairment with benzodiazepines. CBT has also been reported to be a more cost-effective treatment than antidepressants (Heuzenroeder et al., 2004).
What are the barriers to patients receiving CBT for panic disorder? Available data suggest two salient factors: failure of individuals to seek help and failure of the health professionals consulted to apply appropriate treatments. Data from the first Australian National Survey of Mental Health and Wellbeing (NSMHWB) indicated that 68% of those with panic disorder, and 42% of those with agoraphobia, sought help for their disorder (Issakidis and Andrews, 2002). It was estimated that 76% of those with panic and 20% of those with agoraphobia who did seek help reported receiving treatment considered to be of potential effectiveness such as medication or CBT, although unfortunately rates of each were not reported, and the medication was not specified. So one could estimate that the prescriptions being written were being distributed among 76% (and probably fewer) of those with panic disorder and a much lower proportion of those with agoraphobia. It could thus be argued that rates of alprazolam prescription are even higher than estimated by Moylan et al. (2012).
Panic disorder without agoraphobia is associated with the highest likelihood of recovery among the anxiety disorders. The Harvard/Brown Anxiety Research Project (HARP) is a naturalistic longitudinal study that has reported on patients receiving treatment as usual. At intake in 1989–1991, over 80% of participants were receiving a benzodiazepine, although use of selective serotonin reuptake inhibitors (SSRIs) gradually increased (Yonkers et al., 2003). Remission was defined with a low threshold as being a period of 8 consecutive weeks with occasional or no symptoms, and was most likely to occur within the first 2 years. After 8 years of follow-up, 76% of women and 69% of men with panic disorder without agoraphobia had achieved a period of remission, but only 39% of women and 35% of men with panic disorder with agoraphobia. Relapse was common, especially in women. The HARP study is consistent with the NSMHWB findings that individuals with panic disorder, and especially agoraphobia, are not receiving effective treatment.
In Australia, symptom severity and attitudinal variables appear to be of greater importance than structural variables in determining the rate and timing of help seeking (Issakidis and Andrews, 2002; Thompson et al., 2004), suggesting a role for strategies to improve mental health literacy. With the advent of Internet-based therapies of demonstrated efficacy, the potential barrier of limited accessibility is greatly diminished as a reason for not prescribing psychological therapy.
In summary, there is evidence to suggest that alprazolam is possibly being prescribed at higher doses for longer periods of time in panic disorder. This is of concern if it is being used as the primary treatment instead of psychological therapies with a greater duration of effect, or pharmacological therapies with low potential for dependence and cognitive impairment. Better education of consumers regarding treatment options and effectiveness, and support for health professionals in adhering to guidelines, may be helpful.