First-generation antipsychotic (FGA) long-acting injections or depots were developed in the 1960s and were specifically aimed at promoting treatment adherence in people with chronic schizophrenia, thereby enhancing relapse prevention. ^{Reference Davis, Matalon, Watanabe and Blake1–Reference Haddad, Taylor and Niaz4} Most patients with schizophrenia (about 80%) recover from their first episode of illness. Of these, approximately 80% then go on to relapse within 5 years. ^{Reference Robinson, Woerner, Alvir, Bilder, Goldman and Geisler5,Reference Barnes, Shingleton-Smith and Paton6} Worse still, approximately 5–10% of people with schizophrenia are thought to complete suicide. ^{Reference Miles7,Reference Palmer, Pankratz and Bostwick8}

It is timely to examine the issues around the use of long-acting injections (LAIs) in the maintenance treatment of schizophrenia, as we now have sufficient experience of the first second-generation antipsychotic LAI – risperidone LAI – which has been available since 2002. Another second-generation LAI has been licensed and a third is in process. ^{Reference Fleischhacker9} This supplement contains the first systematic review of the data surrounding the use of injectable risperidone, ^{Reference Fleischhacker9} as well as providing valuable critical updates on diverse LAI-related areas such as pharmacology, attitudes to LAIs, FGA–LAIs compared with oral FGA medications, LAIs and community compulsion, the nursing perspective and a snapshot of English prescribing patterns.

Utilisation rates

Long-acting injection utilisation rates vary greatly internationally, with lower rates seen in France and the USA. ^{Reference Dencker and Axelsson10} In countries such as Australia and the UK, the higher rates of use are currently approximately 30% of patients with schizophrenia, ^{Reference Barnes, Shingleton-Smith and Paton6} and only a minority of these individuals are subjected to compulsory treatment. This is confirmed by recent data from Scotland, which revealed that of 2323 patients with schizophrenia, 34% were prescribed an LAI, of whom 15% were detained (further information available on request). However, even in these countries, the rate of use declined considerably following the introduction of the oral second-generation antipsychotics (SGA–orals) in the 1990s. ^{Reference Johnson and Wright11–Reference Callaly and Trauer15} Subsequently the knowledge base and skills associated with prescribing FGA–LAIs arguably declined, leaving a younger generation of psychiatrists at risk of diminished experience. In turn this may further add to the decline in LAI utilisation. For SGA–LAIs there are other restraints, and perhaps the most common of these is their high purchase cost. The cost differential between FGA–LAIs and SGA–LAIs is large. Financial cost does remain an important concern and in certain areas, use has been restricted by those holding the medication budget. Even if others consider depot utilisation rates in the UK to be at a reasonable level already, it remains true that evidence of regional variation exists.

According to systematic reviews approximately 40–60% of patients with schizophrenia are known to be partially or totally non-adherent to oral antipsychotic medication. ^{Reference Cramer and Rosenheck16–Reference Zygmunt, Olfson, Boyer and Mechanic18} Long-acting injections are indicated where medication adherence is a cause for concern. Thus it is argued by some that it might seem reasonable to consider such injections for approximately half of patients with schizophrenia, and Valenstein et al reported that 61% of patients with schizophrenia had difficulties with adherence at some point over a 4-year period. ^{Reference Valenstein, Ganoczy, McCarthy, Myra Kim, Lee and Blow19} However, the dearth of long-term studies comparing outcomes between oral medication and long-acting injections remains an important cautionary note. The studies that do exist are old, of poor methodological quality, short in follow-up duration and usually underpowered. When an LAI is prescribed, attendance in clinic for a regular injection every 1–6 weeks guarantees delivery of medication and assists in the process of regular review. ^{Reference Wistedt20} Covert non-adherence can be overcome as it is immediately obvious to the clinician when the patient has not been given an injection. ^{Reference Levine, Schooler and Cassano21} This early detection of non-adherence also facilitates the opportunity for early contact with the patient, and discussion about the reasons for non-adherence, including unintentional forgetfulness and disorganisation.

Intentional and overt non-adherence with LAIs does, of course, also occur. However, findings of two systematic reviews on LAIs suggested a non-adherence rate of only 24% (range 0–54%). ^{Reference Young, Zonana and Shepler22,Reference Young, Spitz, Hillbrand and Danieri23} Three subsequent studies investigated patients who remained on LAIs for 12 months. Heyscue et al reported an overall ‘adherence proportion’ of 0.96 (the adherence proportion is the ratio of kept appointments to scheduled appointments for LAI administration). ^{Reference Heyscue, Levin and Merrick24} In a later study 26% were reported to have poor adherence, based on the duration of missed injections. ^{Reference Tattan and Creed25} Most recently, Shi et al calculated the mean medication possession ratio (cumulative number of days covered by depot divided by 365 days) to be 91% for patients receiving FGA–LAIs. ^{Reference Shi, Ascher-Svanum, Zhu, Faries, Montgomery and Marder26} This implies that adherence rates are better with LAIs than with oral formulations. However, it seems that LAIs have lower than expected prescribing rates, i.e. 28–36% of patients with schizophrenia, ^{Reference Paton, Barnes, Cavanagh and Taylor27–Reference Barnes, Paton, Hancock, Cavanagh, Taylor and Lelliott29} but the prescribing practice of an individual psychiatrist is subject to many factors (see Appendix). ^{Reference Patel and David2,Reference Waddell and Taylor30,Reference Gray, Spilling, Burgess and Newey31}

Allowing for the differences between FGA–LAIs and SGA–LAIs, some believe that LAIs as a group are underused. ^{Reference Kane, Aguglia, Altamura, Ayuso Gutierrez, Brunello and Fleischhacker32–Reference Glazer34} It should be noted that LAIs are unable to prevent relapse completely; even in clinical trials there is an irreducible 20–25% of patients who relapse despite receiving an LAI. ^{Reference Adams, Fenton, Quraishi and David35} If patients discontinue LAI treatment their risk of relapse increases, but if a patient relapses despite regular injections then non-adherence can be safely excluded as the cause. ^{Reference Barnes and Curson14} Although the evidence base for examining the potential gains of LAI use over oral preparations remains inadequate,^{294,Reference Fleischhacker9,Reference Adams, Fenton, Quraishi and David35} clinicians can use this as a legitimate reason for not increasing LAI usage. Furthermore, the relative merits of prescribing an SGA–LAI over an FGA–LAI are unknown. Perhaps surprisingly, no prospective head-to-head double-blind randomised controlled study has been conducted to date, although recent evidence and debate have suggested that the efficacy of SGA–orals is not superior to that of FGA–orals. ^{Reference Lieberman, Stroup, McEvoy, Swartz, Rosenheck and Perkins36–Reference Kahn, Fleischhacker, Boter, Davidson, Vergouwe and Keet38} That said, a good study of LAIs would be of long duration since the most salient outcomes go beyond efficacy in terms of symptom control and extend to relapse prevention, rehospitalisation and mortality. Until such studies are done, clinicians are likely to choose one LAI over another according to preference of side-effect profile for each individual patient, and perhaps cost.

Long-acting injections and polypharmacy

A recurring theme in this supplement is antipsychotic polypharmacy and the use of LAIs. ^{Reference Johnson3,Reference Barnes, Shingleton-Smith and Paton6,Reference Taylor39} We note that antipsychotic polypharmacy is more common in those prescribed an LAI. ^{Reference Barnes, Shingleton-Smith and Paton6} Is such polypharmacy always a ‘bad thing’, and are there circumstances in which it makes good pharmacological sense to use this approach? ^{Reference Correll, Rummel-Kluge, Corves, Kane and Leucht40} Certainly it is a part of our prescribing heritage, ^{Reference Johnson and Wright11} and the practice is common in many countries.

There may be a perception that the side-effects of LAIs are worse than those of oral preparations of the same drug, contrary to the available evidence, ^{Reference Patel and David2,Reference Haddad, Taylor and Niaz4,Reference Patel, Nikolaou and David33,Reference Taylor39} and this belief may be fuelled by the use of polypharmacy. It may also be that when clozapine fails, owing to lack of efficacy or non-adherence, prescribers look to LAIs to provide adherence and then add another oral antipsychotic (i.e. polypharmacy) to attempt to compensate for single-agent lack of efficacy in known treatment resistance.

Stigma and LAIs: what's in a name?

It has been noted that LAIs may have an ‘image problem’. ^{Reference Patel, Nikolaou and David33} This is especially true for FGA–LAIs. Many proponents of SGA–LAIs have attempted to dodge this by rejecting the term ‘depot’, which was perceived to be stigmatising, in favour of ‘long-acting injection’, although some have argued that a new word had to be used anyway because of the different nature of the drug, implying that it was not a ‘true’ depot. In 2002 the belief that second-generation antipsychotics were superior in all ways – including efficacy, adherence and side-effect profile – was common. ^{Reference Jones, Barnes, Davies, Dunn, Lloyd and Hayhurst37,Reference Voruganti, Cortese, Oyewumi, Cernovsky and Awad41,Reference Davis, Chen and Glick42} Those with a vested interest seemed to work hard to deter prescribers from associating SGA–LAIs with their older FGA–LAI counterparts. Choosing a collective term for this group of medications was difficult. Underpinning this whole argument was the debate about how to overcome the potential for discrimination associated with the word ‘depot’, as it was felt that it was perceived by some (but not all) clinicians as stigmatising. After debate, the collective term ‘long-acting injection’ was chosen to cover both first- and second-generation antipsychotics in this formulation, with the subcategories of FGA–LAI and SGA–LAI. This was partly an attempt to move away from stigmatising stereotypes, and also to promote therapeutic optimism for a population for whom hope can be all too scarce.

Conclusion

The problems facing psychiatrists, patients and their carers 40 years ago, when FGA–LAIs were first introduced, ^{Reference Johnson3} remain observable today. Non-adherence to antipsychotic medication is still a key concern, particularly with regard to relapse prevention. Approximately 50% of patients with schizophrenia are partially or totally non-adherent to their antipsychotic regimen but only 30% or less are prescribed an LAI, although of course injections only have a chance to work if they are accepted. Heres et al inform us that, as psychiatrists, we grossly underestimate the numbers of our own patients who are non-adherent, ^{Reference Heres, Hamann, Kissling and Leucht43} and anxieties about LAIs are in some cases based on suboptimal knowledge. ^{Reference Waddell and Taylor30,Reference Patel, Nikolaou and David33} Thus, psychiatrists may not even think to consider discussing with their patients the option of prescribing an LAI. This situation is complicated by the fact that LAIs are unable to prevent relapse completely, ^{Reference Adams, Fenton, Quraishi and David35} and there is also a virtual absence of high-quality head-to-head comparisons of oral and LAI antipsychotic formulations. Naturalistic studies of risperidone LAI report attrition rates higher than those seen in randomised controlled trials and the reasons for this are not fully understood. These are gaps in knowledge and behaviour that should concern us.

Long-acting injection prescribing will evolve in the future, and associated clinical guidelines will require regular updates, ^{Reference Kane and Garcia-Ribera44} as seen in the recent update of the National Institute for Health and Clinical Excellence (NICE) guidelines for schizophrenia. ⁴⁵ This guidance states that LAIs should be considered for those who prefer such treatment, and when avoidance of covert non-adherence is a clinical priority. It is conceivable that LAI prescribing will increase in the coming years as more SGA–LAIs will be made available, and the increasing use of compulsory community treatment orders may contribute to this. ^{Reference Lambert, Singh and Patel46} The evidence base supporting (or otherwise) LAIs needs to be substantially updated and improved so that the ‘gaps’ will diminish and more appropriate use of LAIs can occur. Perhaps the most critical issue to be addressed is the need for definitive evidence from randomised controlled trials as to whether or not SGA–LAIs are superior, in terms of clinical outcomes, to their respective oral formulations. This needs to be considered particularly in those who have a history of prior oral non-adherence. Surprisingly, these studies have still not been conducted for risperidone LAI, and we urge industry and other funding bodies to consider this a true priority. Other new studies might address FGA–LAIs v. SGA–LAIs (efficacy and tolerability), the use of ‘real world’ outcomes such as mortality and relapse rates, and the nature of the clinician–patient interaction when LAIs are used.

We also hope that the use of long-acting injections occurs within the context of a therapeutic relationship, where both clinician and patient are working in collaboration, and the option of an LAI is discussed openly and not merely as a last resort. Thus, informed patient choice for LAIs may truly occur and the adverse image and associated stigma of this formulation can be reduced.

Appendix