6 March 2009

Close examination finds flaws with Annals article on QT effects from methadone.

Krantz MJ, Martin J, Stimmel B, Mehta D, Haigney MCP. QTc Interval Screening in Methadone Treatment. Ann Intern Med 2009 (March 17 issue)

Dear Reader,

The title of this article has changed since its original publication and the connection with CSAT has been omitted. I found problems with the methods, processes of ‘consensus’, deductions, conclusions and references.

It is evident (but not stated specifically that I could find) that the main issue being addressed is the occurrence of cardiac arrhythmia in patients being prescribed methadone. Yet the title seems to imply that QT prolongation of itself is a problem, despite 40 years of experience showing it is common (up to 40% of subjects) and yet of unknown clinical significance. Torsade tachycardia has very largely been reported in complex medical cases and those taking extremely high doses of methadone rather than those on standard maintenance treatment.

In my view these authors do not make a logical case for their title: “QTc Interval Screening in Methadone Treatment” but ask readers to accept that there is a problem and that their recommendations form a solution to diagnosing and addressing it. Recommendation 1 involves disclosing the cardiac toxicity to all patients; Recommendation 2 advises a history and physical exam. Serial cardiographs are promoted in Recommendation 3. Recommendations 4 and 5 are a cook-book way of dealing with this difficult and largely uncharted clinical territory. Few clinicians have looked after more than one or two such patients and so a rational approach has not yet been arrived at and it is hard to imagine this is the last word on the subject.

Their first paragraph contains a circular argument since they use the existence of a drug black box warning and FDA safety warning on methadone as support for the case for methadone being dangerous. Yet these measures are a result of the same concerns as Krantz claims to be responding to, so as commercial or regulatory decisions, they are not scientific sources, depending as they do on a variety of factors beyond clinical medicine and public health.

Krantz and colleagues are initially at pains to point out the factors which lead to increased risk of torsade de pointes in relation to prolonged QT interval, sex, heart rate and other factors. The rest of the article lacks clarity and the concise scientific discussion that one normally expects in Annals.

Regular practice would start by describing a clinical or public health problem such as a series of case reports, approximate incidence and evidence of the existence of a recognisable syndrome and a possible causation. They appear unwilling or unable to define the problem and its scope. In proposing these clinical recommendations, some of which are ’motherhood statements’ while others appear arbitrary and untested. In this way Krantz and colleagues deny readers a proposed rationale to demonstrate how known reported cases could have been avoided as a result of their newfound wisdom. They circumvent their subject in numerous ways, drawing quite tenuous conclusions from circumstantial reports with no actual cases of torsade arrhythmias despite being cited as important studies demonstrating its importance (eg. Chugh’s study from Portland, Fanoe from Denmark, Wedam from Baltimore).

Despite torsade de pointes being the complication they are addressing, the article spends most text discussing QT prolongation, something we know happens commonly in methadone patients (up to 40%), and which we know, in methadone clinic patients on ‘normal’ doses, is of little if any clinical significance. Torsade can occur in those with normal QT intervals (Ehret) and in those not taking methadone (Smith). While taking pains to be conservative and conceding the many weak links individually in documenting this subject, these authors still conclude that their advice is based on good science.

As above, it is hard to understand how, from a knowledge of the case reports, such a strategy as proposed by Krantz and colleagues would or could prevent torsade cases. The QT interval is regularly normal before and after the triggering events (Sticherling). I have written to Krantz, Haigney, Stimmel and Martin individually to ask how their strategy could apply to the case reports in the literature. I have been sent no attempts to explain this rather large gap in the logic. My understanding is that few if any reported cases would have been prevented by these measures in Annals.

The authors state that of Pearson’s 59 FDA reported cases there was an 8% mortality (Paragraph 14). They omit to say that only one of the 5 deaths was a torsade case (the others QT abnormalities reported but no torsade). Further, the single death was in a 47 year old female patient who also had a myocardial infarction as well as prescription of azithromycin and droperidol. Both the latter drugs are known to be cardio-toxic. The mean dose of the 59 cases was over 400mg daily. None of the other 4 deaths in Pearson’s FDA series had torsade from their prolonged QT intervals and we are not told any further details of the causes of death. Two of the four had been given methadone intravenously (off-label) at extremely high dose levels (360 and 1680mg daily). Another was a 78 year old woman who had been prescribed cisapride, a drug which is no longer available in some countries. The only patient in this group of five deaths who might have been a standard methadone patient also died from un-stated causes, aged 40 on the unusual dose of 29mg daily (and there was no torsade in her case).

By comparison, Krantz’s series of 17 cases had no deaths (0%), Sticherling’s 5 cases all survived (0% deaths) and Justo’s compilations (including some of the above cases) reported no deaths (0%).

Hence the suggestion that any group of methadone patients had a mortality of 8% is almost meaningless without a denominator. Considering the age and other details of reported cases, these would have little relevance to young people with addiction problems who may be started on opioid maintenance therapy. Few if any of those reported torsade cases come from newly started addiction clinic patients, despite the most worrying trial of QT prolongation (Wedam) finding 12% in the high risk group within 4 months of starting treatment. Even if there were a small incidence of significant QT problems, these would still be outweighed by benefits to patients. Krantz himself proposed that putting a drug injector onto methadone had the scope to reduce rates of endocarditis in the community as a “common sense notion” (2001). Endocarditis is probably more common than torsade de pointes arrhythmia.

Krantz and colleagues argue (paragraph 2) in favor of routine cardiographs by taking examples of findings with two antiarrhythmic drugs (sotalol and dovetailed), “highlighting the importance of pretreatment ECG screening for identifying susceptible patients”. One wonders at this comparison when these cases clearly already had heart disease by definition, in contrast to young people attending for addiction treatment. While it would obviously be inappropriate to treat arrhythmias without a baseline and on-going cardiographs, there can be no parallel here with methadone as the authors attempt. A fairer comparison might be prescribing erythromycin, haloperidol or other such agents to young people without cardiac histories.

The 17% mortality of torsade is based on two old references from the French literature relating to hospitalized torsade cases (Paragraph 16). This rate may now be lower in view of better communications, wider availability of ECG and defibrillators as well as improved specialist care. On the other hand, torsades may have become more readily diagnosed, due in part to the advent of automated digital machinery with QTc print-outs.

In paragraph 21 Krantz and colleagues combine 8 references as supporting a correlation between prescribed methadone dose level and QT interval. In fact, Peles’ trial from Israel (which probably had the highest average doses and largest range of any such report) found no significant correlation between their patients’ dose levels and corresponding QT interval. A sub-group of cocaine users were examined separately and a (significant) correlation was found which may or may not support Krantz and colleagues’ thesis. Further, they quote Martell as supporting the correlation but fail to add (as Cruciani states:) “Martell and co-workers studied heroin addicts during the first two months of induction therapy with methadone and observed a higher increment in the duration of the QTc in those patients receiving 110-150 mg/24 h. The clinical significance of this change is questionable, however, because the increment was only 13.2 ms.”

Further, in paragraph 21 these authors state, or rather understate: “Methadone dosages exceeding 100 mg/d have frequently been noted in published cases of torsade de pointes, and some case reports (43, 47, 55) highlight QTc-interval normalization after methadone discontinuation or dose reduction.” In fact methadone dosages exceeding 200mg, 300mg and 400mg have frequently been noted in reports of Pearson and Krantz (2002). Some of the highest were 1100mg, 1680mg, 1000mg in Pearson’s paper. Further, when QTc interval was available after the torsade event and the triggering factor has been removed, QTc intervals nearly always returned to normal or near normal. Krantz omits this common and important finding while stating “some case reports (43, 47, 55) highlight QTc-interval normalization after methadone discontinuation or dose reduction.” To this one should add the several reports where normalisation of the QT interval was reported after addressing triggering factors (eg. all 5 cases of Sticherling, De Bels’ two cases, one reverting to normal while the other’s QT interval dropped from 736ms to 502ms in 4 days).

This style of writing is much closer to advocacy than careful scientific discourse. While there are caveats and alternatives mentioned at various points, the overall feeling is that there is a case already made and this text is there to support it. The choice of references is another example of a lack of balance. Justo’s prominent literature review from the Addiction journal is omitted. Krook’s item which addresses their exact subject is also surprisingly left out (Krook AL, Waal H, Hansteen V. Routine ECG in methadone-assisted rehabilitation is wrong prioritization. Tidsskr Nor Laegeforen 2004 124;22:2940-1).

The authors also unfortunately omitted two highly relevant recent items (i) Athanasos P, Farquharson AL, Compton P, Psaltis P, Hay J. Electrocardiogram characteristics of methadone and buprenorphine maintained subjects. J Addict Dis. 2008 27(3):31-5 (ii) Cruciani R. Methadone: To ECG or Not to ECG…That Is Still the Question. Journal of Pain and Symptom Management 2008 36;5:545-552. These two address Krantz’s issues directly and each makes enlightening and balanced reading, contributing substantially to the field, yet they are ignored by Krantz and colleagues. Cruciani was available in April 2008 while Athanasos on 12th June 2008. Several of the other 95 references were accessed as late as November 12 2008 according to the text.

With almost 100 other references, some of only tenuous relation to the subject, it is a flaw to have missed other such relevant and contributory sources. In this small field, such documents are usually publicised on the internet, professional list-servers and news-wire services long before they reach formal publication date (as in the case of this very item in Annals which appeared in a previous version in early December 2008). The reader may understand cut-off dates for recent references, but to omit Krook and Justo would seem to show a lack of thoroughness unbefitting a panel which proposes to develop guidelines for physicians who work in this important field.

Derivative internet summaries:
http://www.ncbi.nlm.nih.gov/pubmed/19047020
http://www.tripdatabase.com/spider.html?itemid=801110

Comments by Andrew Byrne ..

Clinic web page: http://www.redfernclinic.com/#news

References:

Justo D, Gal-Oz A, Paran Y, Goldin Y, Zeltser D. Methadone-associated Torsades de Pointes (polymorphic ventricular tachycardia) in opioid-dependent patients. Addiction. 2006 101:1333-1338

Smith WM. Cardiac repolarisation: the long and short of it. MJA 2008 188;12:688-689

Ehret GB, Voide C, Gex-Fabry M, Chabert J et al. Drug-Induced Long QT Syndrome in Injection Drug Users Receiving Methadone: High Frequency in Hospitalized Patients and Risk Factors. Arch Intern Med 2006 166:1280-1287

Wedam EF, Bigelow GE, Johnson RE, Nuzzo PA, Haigney MCP. QT-Interval Effects of Methadone, Levomethadyl, and Buprenorphine in a Randomized Trial. Arch Intern Med 2007 167;22:2469-2473

Pearson EC, Woosley RL. QT prolongation and torsades de pointes among methadone users: reports to the FDA spontaneous reporting system. Pharmcoepidemiol Drug Saf. 2005 14;11:747-753

Krantz MJ, Lewkowiez L, Hays H, Woodroffe MA, D. Robertson AD, Mehler PS. Torsade de Pointes Associated with Very-High-Dose Methadone. Ann Intern Med. 2002 137:501-504

Krantz MJ. Clinical Concepts- Cardiovascular Health in MMT Patients. Addiction Treatment Forum 2001 No 4

Peles E, Bodner G, Kreek MJ, Rados V, Adelson M. Corrected-QT intervals as related to methadone dose and serum level in methadone maintenance treatment (MMT) patients - a cross-sectional study. Addiction 2007 102;2:289-300

Cruciani R. Methadone: To ECG or Not to ECG…That Is Still the Question. Journal of Pain and Symptom Management 2008 36;5:545-552

Martell BA, Arnsten JH, Krantz MJ, Gourevitch MN. Impact of methadone treatment on cardiac repolarization and conduction in opioid users. Am J Cardiol. 2005;95:915-8

Other references on request.