Clinical & Policy Updates:
SMMGP Clinical Update August-September 2013
Compiled by Dr Euan Lawson
|Download the PDF version of this Update here! (PDF*, 83K)|
Efficacy of as-needed nalmefene in alcohol-dependent patients with at least a high drinking risk level: results from a subgroup analysis of two randomized controlled 6-month studies
van den Brink W, Aubin H-J, Bladström A, et al. Alcohol Alcohol 2013;48:570-8
This aim of this post-hoc analysis was to investigate the effectiveness and the safety of 18mg nalmefene versus placebo in reducing alcohol consumption - in those who did not reduce their consumption after initial assessment. It was a subgroup analysis of two randomised controlled studies (ESENSE 1 and ESENSE 2) that were both six months long. These two studies were run across several European countries and have had their primary findings published in 2012 and 2013. They had exactly the same design and involved recruiting alcohol dependent people who were randomised to 24 weeks of as-needed treatment with nalmefene or placebo. Patients were told to take one tablet per day on days where they perceived they were at risk of drinking alcohol. They all received psychosocial interventions. No treatment goals were set and information wasn't collected on the participants' initial aims - both abstinence and reduced-risk drinking were accepted. The primary outcomes were the change in baseline in number of heavy drinking days (HDDs) and the total alcohol consumption per day at month six (measured in grams of pure alcohol).
The results included 667 patients with 335 people in the nalmefene group and 332 people in the placebo group. There were no significant differences between the nalmefene and placebo groups at baseline in terms of HDDs with a range of 21.6 to 23.1 days/month. At six months the mean number of HDDs was reduced in the nalmefene group by an average of 3.2 days (95% CI: 1.6 to 4.8; p<0.0001). The average total alcohol consumption at the start of the study was in the range 99 to 114g/day. The total alcohol consumption was reduced by 51.4g/day in the placebo group but it was 65.7g in the nalmefene group - a treatment effect of 14.3g and which is highly statistically significant. [A unit of alcohol in the UK is defined as 7.9g of alcohol.] There were more adverse events in the nalmefene group and these did lead to more dropouts as well. During the six months of treatment 77% of patients on nalmefene reported one or more adverse effects with the most common being dizziness, nausea and insomnia - compared to 67% of placebo. However, improvements in clinical status and liver function tests were better in the nalmefene groups compared with placebo.
Commentary: Over the summer a new drug became available for use in the treatment of alcohol dependence. Nalmefene (Selincro®) epitomises what seems to be a definite shift in the aims of some pharmaceutical companies with an increasing emphasis on a harm reduction approach. The important factor about this sub-group analysis is that it pulled out data for those with a higher drinking risk. It has been recognised that people who seek help after they make the decision to tackle their drinking will often reduce their consumption with minimal intervention. This analysis looks a sub-group who didn't reduce their drinking in that period immediately after first contact. It represents a hard-to-treat group. It is worth noting that this study showed a big reduction in drinking in the placebo group yet nalmefene still showed an effect over and above this large effect.
The authors conclude that nalmefene on an as-needed basis should be offered to all alcohol-dependent patients in primary care who are not able to reduce their alcohol consumption following an initial assessment or brief intervention. If forthcoming guidelines in the UK support the authors' conclusion then this heralds the significant arrival of harm reduction in alcohol treatment.
Cost-effectiveness of injectable opioid treatment v. oral methadone for chronic heroin addiction
Byford S, Barrett B, Metrebian N, et al. British Journal of Psychiatry. Published Online First: 12 September 2013. doi:10.1192/bjp.bp.112.111583
This cost-effectiveness study was run off the back of the Randomised Injectable Opiate Treatment Trial (RIOTT) that was reported in the Lancet in 2010. In RIOTT the two interventions consisted of twice daily injected heroin doses and once daily injectable methadone. These were self-administered under direct nursing supervision in special clinics. They were treated for 26 weeks and a control group on optimised oral methadone was also included. In the original study the primary clinical outcome was the proportion of participants negative for street heroin in at least 50% of their urine tests after 14 weeks. The economic outcome used in this study was quality-adjusted life years (QALYs). They took into account a wide range of economic resources that included: detailed information on the medication, equipment, urine testing and staff costs. It was all calculated using 2007/08 costings and hasn't been adjusted for inflation to 2013 prices.
The results showed that the total intervention costs were higher for the injectable heroin group - £8995 compared with £4674 for the injectable methadone and £2596 for the oral methadone group. There were no significant differences in other resources categories. When other costs such as other services and crime were considered then the oral methadone group was most expensive (£15,805), followed by injectable heroin (£13,410), and methadone (£10,945). However, the cost effectiveness analysis showed that injectable heroin was most effective and oral methadone was dominated by both the injectable options (i.e. it was more expensive and less effective).
Commentary: RIOTT confirmed other similar studies in other countries that there was clear evidence of the effectiveness of injectable heroin in those who were regarded as treatment failures with oral methadone. Injectable Opioid Treatment (IOT) is now evidenced as a clinically effective second line treatment for a small group of people who have repeatedly failed to respond either to standard methadone treatment or to residential rehabilitation. A three-year pilot programme was set up in 2012 (reporting in March 2015) with sites in Brighton, Durham and London exploring the appropriate referral pathways into and out of injectable opioid treatment and its cost-effectiveness, with a view to informing and securing future commissioning arrangements.
I wrote in June 2010 in the Clinical Update: Cost-effectiveness is crucial and the immediate future of IV heroin in the UK may hinge around this issue. It's not one that will be easily ducked in an austere climate of treatment rationalisation. Decent, therefore expensive, clinical facilities and staffing will be required over and above current practice in many areas. Those that might be tempted to drift back toward injectable methadone as an interim measure will have to consider the lack of a significant difference between injectable and oral methadone that this study highlighted. The cost issue remains a significant one given the level of clinical supervision and support needed.
We've had to wait over three years for some further information on the costs. This paper now points towards the injectable preparations being cost-effective when compared with oral methadone for chronic refractory heroin dependence. The problem is that the savings are made in the criminal justice sector and when the health service perspective is considered on its own then injectables are no longer cost-effective. Assuming the prejudice about the provision of injectables is laid aside one wonders if commissioners will be prepared to countenance the costs when the societal benefits are spread so widely.
Buprenorphine Implants for Treatment of Opioid Dependence: Randomized Comparison to Placebo and Sublingual Buprenorphine/Naloxone
Rosenthal RN, Ling W, Casadonte P, et al. Addiction. Published Online First: 7 August 2013. doi:10.1111/add.12315
This study reported on a randomised controlled trial looking at the safety and efficacy of buprenorphine in those having implants versus placebo implants. The study was carried out in 20 addiction treatment centres in the United States. The primary aim of this study was to compare buprenorphine implants to placebo, with a secondary aim to assess if implants were non-inferior to sublingual buprenorphine. Initially, all the participants were inducted onto sublingual buprenorphine. Subsequently, depending on randomisation, the participants went into one of three arms of the study - buprenorphine implants (80mg/implant), placebo implants, or sub-lingual buprenorphine/naloxone (12-16mg/day). The sublingual buprenorphine wasn't blinded at all.
The implants were made by Titan Pharmaceuticals who were the sponsor for this study. They involved the insertion of the implant into the subdermal space (2-3mm below the skin) in the inner, upper arm. All the implants were removed at six months. The subjects and study staff were all blinded with the exception of the implanting clinicians. During the course of the 24 weeks all the participants could receive extra buprenorphine/naloxone.
The primary outcome was the percentage of urines negative for opioids collected weeks 1 to 24. There were 114 in the buprenorphine implant group, 54 in the placebo implant group, and 119 receiving sublingual buprenorphine. The results showed that buprenorphine implants were significantly better than placebo. As part of the secondary analysis the Clinical Opiate Withdrawal Scale (COWS) and Subjective Opiate Withdrawal Scale (SOWS) were both recorded. This showed that those in the implant group experienced statistically significant levels of withdrawal symptoms compared with those getting sublingual buprenorphine - but this didn't translate into more drug use when compared with placebo as the implants were non-inferior to sublingual buprenorphine when it came to percentage of negative urines.
Commentary: If you're feeling some mild deja vu with this study it bears considerable similarity to a randomised controlled study of buprenorphine implants published in JAMA three years ago. We reported on it in the October 2010 Clinical Update. A number of weaknesses were highlighted and I commented then that we needed studies comparing buprenorphine implants to sublingual treatment.
This study partially answers that but the primary outcomes are still focused on comparing implants to placebo. I have some reservations about whether or not it's really helpful for clinicians to compare a buprenorphine implant to placebo. Most right thinking clinicians will want to know how an implant compares to the standard set by sublingual buprenorphine. This was a Phase 3 study so placebo-controlled studies are often necessary to achieve the licensing but there is a risk that it inflates the results in the eyes of the clinician. In the end it shows that implants were no better but no worse either. In many ways that is useful to know - after all it is possible to imagine patients expressing a personal preference for implants. However, imagination is all we have, not ideal for a clinician, as we don't have any actual research on whether people would want them.
Sheron N, Moore M, O'Brien W, et al. Feasibility of detection and intervention for alcohol-related liver disease in the community
Br J Gen Pract 2013; DOI: 10.3399/bjgp/13X673711
This study randomly selected individuals from nine general practices across Hampshire, Wiltshire and Dorset. They were offered the WHO Alcohol Use Disorders Identification Test (AUDIT) and those who fell into the high risk category, scoring 8 or more, were offered the Southampton Traffic Light test (STL). They were given the results and then followed up with a repeat AUDIT after one year. The STL is a blood test that combines two markers of liver fibrosis - hyaluronic acid and collagen P3NP and puts them together with the platelet count as a measure of portal hypertension. In time honoured fashion for traffic light tests, red signified trouble and likely liver fibrosis. The results were communicated to the patient in a standard fashion with those in the red and amber groups recommended for some kind of follow up assessment. The follow up AUDIT scores were obtained for 303/393 (77%) and half of the 153 people with evidence of liver damage had reduced their drinking by at least one AUDIT category compared to the 52/150 without evidence of damage.
Alcohol consumption screening of newly-registered patients in primary care
Khadjesari Z, Marston L, Petersen I, et al. Br J Gen Pract 2013; DOI: 10.3399/bjgp13X673720
This paper took the Health Improvement Network (THIN) primary care database (covering around 6% of the UK population) and extracted data relating to alcohol consumption, types of screening test used and units of alcohol per week drunk. In total there were 382,609 patients registered and 292,376 (76%) had an entry for alcohol consumption with the data recorded in units of alcohol per week in 124,348 (43%) and as a Read Code for units per week in 115,010 (39%). Only 9% of patients were recorded as having completed a validated screening test - of which AUDIT and AUDIT-C were the most common.
Commentary: Perhaps the most useful point in the first study is to highlight that ways are being developed to give early indicators of liver disease. Given liver disease is rising and much of it is alcohol-related then measures are going to be developed before hospitals are swamped. Inevitably that means there will be a raft of initiatives and efforts directed at primary care. This paper reminds me of one related to smoking where folk had their spirometry checked and it was expressed to them in terms of lung age. It seems that the simple measure of giving feedback in a quantifiable way has some effect on people's behaviour.
The second paper emphasises a problem for GPs and clinical practice but reflects a wider malaise. Why are clinicians so reluctant to use validated screening tools and questionnaires? Some clinicians may feel that they are crude tools that gatecrash the consultation and that they are too mechanistic or too blunt a tool. It makes many GPs quite uncomfortable, yet the evidence suggests that clinical skill and judgement is distinctly flaky at picking up alcohol use disorders when compared to AUDIT.
Clinical effectiveness of baclofen for the treatment of alcohol dependence: a review
Brennan JL, Leung JG, Gagliardi JP, et al. Clin Pharmacol 2013;5:99-107
This open access review summarises some of the evidence around the clinical effectiveness of baclofen. There have only been four randomised controlled studies thus far and they were included in this review. The authors noted that the primary outcomes have differed between these studies and this hampers any systematic review. In a couple of the studies patients randomised to baclofen did achieve higher rates of absence than in those taking placebo. The secondary analysis is important in that it has suggested baclofen is safe in those with alcohol dependence - including those with moderate or severe cirrhosis. There may also be some anxiolytic effects. However, the largest RCT couldn't find any benefit over placebo. The authors' summary was that baclofen does seem to have some beneficial effects but, as one might expect, the conclusions are limited by the small number of studies.
Commentary: There are few new drugs coming through for the treatment of alcohol dependence when it comes maintaining abstinence. As has already been discussed in this update it seems that the pharmaceutical industry is moving toward the exploration of agents that will reduce harm. That's not necessarily a bad thing - at present there is something of an all-or-nothing paradigm with alcohol dependence and its treatment. Many clinicians will have experienced the horror of watching, alongside friends and family, as an individual is destroyed by alcohol dependence and detox upon detox doesn't hold. Baclofen holds promise but the evidence base is limited. It's worth being aware that patients on any other psychotropic agents were excluded from the studies and there's no evidence about dose adjustment in renal impairment. If you want to acquaint yourself with the current situation then this open access paper from Clinical Pharmacology: Advances and Applications is a concise summary.
QTc Interval Screening in an Opioid Treatment Program
Katz DF, Sun J, Khatri V, et al. Am J Cardiol 2013;112:1013-8
This paper reported on the experience of an ECG-based safety program in people on methadone in a clinic in the United States. The results showed that out of 531 new entrants, 436 (82%) had at least one ECG taken. Out of these, 186 (35%) had a pre-treatment ECG. A total of 21 patients had a QTc interval that was longer than 500ms. The authors describe 32 unique interventions to address the prolonged QT interval in those 21 patients. These included: 9 patients who had methadone dose reductions; 9 patients who had dose increases halted; other QTc prolonging drugs were stopped in 2 patients; counselling to avoid other illicit substances such as cocaine and alcohol on 8 occasions; and 4 people were converted to buprenorphine. Subsequently, 12 out of the 21 had a QTc interval that had reduced to <500ms.
QTc prolongation in methadone maintenance - the role of HCV infection
Gholami N, Boesch L, Falcato L, et al. Swiss Med Wkly 2013;143:w13852
This was a short report that gave details on QTc interval in people receiving methadone maintenance treatment in Switzerland. They used a convenience sample of 210 patients. A total of 96 patients (45.7%) had one or more prescribed medications with a QT-prolonging potential and just over a quarter of them were HCV positive. The mean daily methadone dosage was 113.6mg per day. The multiple linear regression showed that longer QTc interval was associated with methadone dosage (p=0.012) and HCV positive status (p=0.005). Commentary: In the first study the authors conclude that clinical characteristics alone were inadequate to identify patients who need ECG screening. They also suggested that the implementation of an "ECG-based intervention" in methadone maintenance can decrease the QTc interval. The reactions of the clinic to the prolonged QTc is the most interesting part of the report. Is reducing methadone more likely to harm the patient than to benefit them? This study can't answer that but it does highlight some other strategies to address prolonged QTc intervals. In particular, the number of patients who were also on other QT prolonging drugs in these studies was startling.
Cost-effectiveness of HCV case-finding for people who inject drugs via dried blood spot testing in specialist addiction services and prisons
Martin NK, Hickman M, Miners A, et al. BMJ Open 2013;3. doi:10.1136/bmjopen-2013-003153
This study was a cost-utility analysis looking at increasing hepatitis C case-finding in people who inject drugs through the use of dried blood spot testing. They used a mathematical model to capture the potential prevention benefits of treatment. The results suggested that dried blood spot testing for HCV is likely to be cost-effective when using the typical thresholds considered in the UK - generally around £20,000-£30,000 per QALY gained. However, while this holds for addiction services it was only the case in prison if a minimum level of continuity between the prison and the community was maintained. Much of the cost-effectiveness stems from the prevention benefit as treating people for HCV reduces the pool of infected people and ultimately will reduce the incidence of infections.
Commentary: I accept that cost-utility studies aren't going to set most clinician's hearts alight with excitement but the issue of testing for HCV is a key one and this study highlights some important issues. Working in clinics, the challenges of venepuncture will be appreciated by anyone who has spent a few minutes gazing, tapping, and stabbing blindly at the mottled arm of a person with a long history of injecting drug use. Of course, case-finding in itself isn't cost-effective as such - all the benefit comes from getting people into treatment.