Volume 189 Issue 8

The humanities offer tools for wise application of biomedical knowledge and promotion of humane medical care Writing in the New York Times, columnist David Brooks recently described a “distinct brand of social misfits” in “fields like law, medicine or politics, where a person’s identity is defined by career rank”.1 He fears that their childhoods may have been spent in domestic “achievatrons” that ensured their academic success but compromised their interpersonal skills. Brooks believes that American society produces a disproportionate number of people with a “rank-link imbalance”, which he described as “the social skills required to improve their social rank, but none of the social skills that lead to genuine bonding”. These people have opinions about everything and “treat their conversational partners the way the Nazis treated Poland. They crush initial resistance, and the onslaught of accumulated narcissism is finally too much to bear”.1 It is hard to know whether Australia has a disproportionate number of “misfits” in law, medicine or politics, but if this is the case, Brooks suggests that they are the people most likely to force their way to the top of their career ladder and make life miserable for the rest of us. One way of producing doctors (or lawyers or politicians) with a capacity for genuine bonding might be to broaden their education.2 However, physician and writer Rafael Campo argues that “no one has proven that injecting the humanities in any form into medical settings translates to more humane physicians or better cared-for patients”.3 Campo’s use of the word “injecting” is telling; it conveys a sense of the humanities as something foreign to medicine. To appreciate whether the humanities are indeed foreign to medicine, try to imagine a health care facility in which no ethical issues are explored, no lessons have been learnt from the past, no cultural awareness is displayed, no written words (other than technical communications) appear, and no books, films, television programs, plays or concerts are discussed by patients or staff. Imagine that there are no artworks, no music and no other aesthetically pleasing elements. Although some of our hospitals are admittedly run down, the products of the arts and humanities are nevertheless all around us. Two recent Australian examples illustrate why we need to draw on the humanities in health care. The first is the front cover of the 19 May 2008 issue of this Journal, which depicts the phrase “Sorry, the first step” spelt out in candles in front of Parliament House.4 Many doctors are indeed sorry that biomedical solutions to Indigenous health problems have been confounded by ignorance concerning Indigenous history and culture.5 Fortunately, the Indigenous Health Curriculum Framework prepared for the Committee of Deans of Australian Medical Schools gives priority to topics such as culture, self and diversity, Indigenous history and society, and models of health service delivery.6 All of these areas draw on knowledge and insights from the humanities. The arts also offer teaching resources that provide for better cultural understanding. Recent examples include the film Ten canoes, which imaginatively recreates the world of the Yolngu people; Kate Grenville’s novel The secret river, and Doris (Garimara) Pilkington’s book Follow the rabbit-proof fence and its subsequent film adaptation; and plays Murras, Coordah and The keepers, which explore the impact of government policies of forced removal. Such resources communicate and educate by being emotionally engaging. The second example relates to quality and safety in health care. Among the competencies outlined in the National Patient Safety Education Framework are communication skills, teamwork, leadership, honesty and respect.7 The intellectual foundations for these competencies lie in the medical humanities, in particular psychology, sociology, philosophy and ethics as applied to medical practice. Biomedicine puts at our disposal the tools for safe, effective health care; the humanities explore their wise application in practice. Certain educational approaches accommodate the humanities better than others.8 Problem-based learning and its variants engage students’ emotions by giving each patient a story. However, problem-based learning is easily subverted by ignoring or parodying the human, experiential features of clinical problems. Privileging biomedical subjects over the humanities quickly alerts students to what counts as knowledge.3 A good medical curriculum provides time and resources for emotional engagement, reflection, and independent, self-directed learning — qualities that characterise what is best about the study of the humanities.9 The human experience of illness is most powerfully conveyed to students by those who have first-hand knowledge. It can be supplemented by poems, novels and films that faithfully represent that experience: Iris, A beautiful mind and The sea inside (which explore dementia, schizophrenia and quadriplegia, respectively) are recent examples. The Medical Humanities website of New York University provides an extensive database of resources on literature, arts and medicine.10 In the United States, the Accreditation Council for Graduate Medical Education has identified compassionate patient care and professionalism among six required competencies for residents, which training programs must assess.11 It has been suggested that the humanities, and specifically bioethics, could contribute to resident education.12 However, it has been argued that time and effort would be better spent in humanising the US health care system itself.13 The humanities cannot make people behave well. The late John Eisenberg, Director of the Agency for Healthcare Research and Quality in the US, has shown that doctors respond to many different influences.14 Well intentioned educational interventions will not produce more humane doctors if their role models’ behaviour suggests that it is better to do well than to do good. Medical facilities are moral worlds15 in which humane behaviour is elicited by being treated humanely,16 both in medical schools and in clinical settings.17 The humanities provide insight into why people (including patients, doctors and politicians) behave as they do and have done in the past. Equipping students with such insight is a necessary but not a sufficient strategy in the never-ending battle with the rank-link misfits.

J Jill Gordon MPsychMed, PhD, FRACGP

Combined with appropriate lifestyle modification and statin therapy, raising HDL levels may be an important strategy to reduce cardiovascular risk In the past two decades, we have made significant advances in the management of cardiovascular disease (CVD), with a dramatic reduction in cardiovascular events occurring in parallel with the widespread use of statins to lower low-density lipoprotein (LDL) cholesterol levels. Nevertheless, despite intensive use of statin therapy, a significant patient cohort remains at high risk of cardiovascular events, paving the way for new strategies to reduce their residual cardiovascular risk. One attractive target for such strategies is high-density lipoprotein (HDL) cholesterol. There is strong epidemiological evidence demonstrating the inverse relationship between the incidence of cardiovascular events in normal populations and serum HDL levels. Based on data from the Framingham Heart Study, the risk of myocardial infarction increases about 25% per 0.13 mmol/L decrement in serum HDL below median values.1 HDL levels are also predictive of coronary events in patients with known CVD across a range of LDL levels, as demonstrated in the Treating to New Targets trial, in which nearly 10 000 patients with established CVD were treated with statins.2 HDL levels were inversely predictive of time to first major cardiovascular event across the spectrum of LDL levels, including patients with treated LDL levels below 1.8 mmol/L, highlighting the predictive value of HDL levels independent of LDL levels. The atheroprotective effects of HDL have been attributed to its ability to mediate “reverse cholesterol transport”, where cholesterol in peripheral tissues is transferred via plasma to the liver for either recycling or excretion. More recently, the antioxidant and anti-inflammatory properties of HDL have been explored. In particular, intravenous preparations of HDL have been shown to dramatically reduce acute vascular inflammation in animal models, improve endothelial function (an important surrogate of cardiovascular risk), and promote atheroma regression and stabilisation in human studies.3 Moreover, because of the heterogeneity of human HDL, an appreciation of function as well as absolute concentration is becoming increasingly important. For example, HDL from subjects with diabetes has been shown to be less effective in its cholesterol-effluxing and anti-inflammatory capacity.4 Lifestyle modifications in the form of regular exercise, smoking cessation, weight loss and moderate alcohol consumption have each been shown to individually raise HDL levels by 5%–10%.5 Although statins are currently the cornerstone of lipid-modifying therapy, they raise HDL levels by only 5%–10%.6 On the other hand, fibrates have been shown to raise HDL levels by 10%–15%.7 Fibrates regulate HDL metabolism as ligands and activators of the nuclear transcription factor peroxisome proliferator-activated receptor-α.8 The benefits of using fibrates to raise HDL levels have been suggested by a number of randomised trials. For example, the Veterans Affairs High-Density Lipoprotein Intervention Trial (VA-HIT) included 2531 patients with CVD, with an LDL level ≤ 3.6 mmol/L, HDL level ≤ 1.0 mmol/L, and triglycerides ≤ 3.4 mmol/L; patients were randomly assigned to receive treatment with gemfibrozil or placebo. At 5 years, the combined primary endpoint of cardiac death and non-fatal myocardial infarction occurred less often in the gemfibrozil-treated group, and the reduction in this endpoint correlated strongly with both serum HDL levels and degree of weight loss, but was independent of changes in LDL cholesterol or triglycerides concentration.7,9 Nicotinic acid is another effective HDL-raising drug, with an ability to raise levels by up to 30%. The HDL-Atherosclerosis Treatment Study (HATS) reported the effects of combined therapy with a statin and niacin on 160 patients with CVD, with an HDL level < 0.9 mmol/L and LDL level < 3.75 mmol/L.10 Compared with placebo, patients receiving simvastatin plus niacin were significantly less likely to experience a cardiovascular event. Furthermore, the magnitude of the reduction of clinical events with drug therapy was greater than that observed in studies of statins alone, suggesting that raising HDL levels provides additional protection beyond that attributable to simply lowering LDL levels.10 The side-effect profile remains problematic with this drug class, though development of extended-release preparations may overcome these issues. Most recently, a novel class of HDL-raising medications, the cholesteryl ester transfer protein (CETP) inhibitors, which prevent the transfer of cholesteryl ester from HDL to triglyceride-rich lipoproteins in exchange for triglyceride, have been tested in large clinical trials. In particular, the CETP inhibitor torcetrapib was evaluated in the multicentre randomised Investigation of Lipid Level Management to Understand its Impact in Atherosclerotic Events (ILLUMINATE) trial, which compared torcetrapib with placebo in more than 15 000 patients receiving atorvastatin.11 A significant increase in HDL levels (72%) and an additional decline in LDL levels (25%) below baseline after 12 months of torcetrapib therapy were seen. However, the trial was prematurely terminated due to a significant increase in cardiovascular events in the treatment arm; an “off target” effect on the aldosterone receptor leading to blood pressure elevations was postulated as a potential mechanism for this.11 Despite the setback experienced with torcetrapib, further studies are underway to develop more target-specific CETP inhibitors. Intravenous reconstituted HDL preparations and apolipoprotein A-I mimetic compounds are also in development. In the future, these compounds may be used to regress atheroma or suppress the arterial inflammation that is the hallmark of the acute coronary syndromes. Thus, raising HDL levels, in combination with optimising LDL cholesterol levels, blood pressure and glycaemic control, as well as appropriate lifestyle modification, represent important strategies for reducing residual cardiovascular risk. Such measures should see further improvements in clinical outcomes for patients with CVD.

Sanjay Patel MB BS, FRACP

Objective: To examine short- and medium-term outcomes of percutaneous coronary interventions (PCIs), with a focus on comparing drug-eluting stents (DESs) with bare-metal stents (BMSs).Design, setting and participants: Retrospective analysis of data from the Melbourne Interventional Group (MIG) registry, a large multicentre Australian registry. The study cohort consisted of 6364 consecutive patients undergoing 7167 PCIs between April 2004 and August 2007.Main outcome measures: Clinical events including death, myocardial infarction (MI), target lesion revascularisation (TLR), target vessel revascularisation (TVR) and major adverse cardiac events (MACE) (a composite of death, MI and TVR), at 30 days and at 12 months.Results: The cohort was predominantly male (74%), with a mean age of 64.7 years (SD, 12.0 years). DESs were used in 3482 (51.4%) of PCIs. In the overall cohort, rates of clinical events were low at 30 days: mortality (1.9%), MI (2.4%), TLR (2.0%), TVR (2.4%) and MACE (5.7%). At 12 months, event rates were: mortality (5.2%), MI (6.0%), TLR (5.8%), TVR (8.2%) and MACE (16.2%). Patients receiving DESs had similar mortality rates to those receiving BMSs (4.0% v 6.0%; P = 0.62 [propensity score-adjusted]); late thrombosis rates were also similar in the two groups (0.8% v 1.1%; P = 0.38). The proportion of patients receiving DESs fell significantly over time, from 54.9% in the first 24 months to 44.7% in the last 15 months of the study period (P < 0.01). Independent predictors of 12-month mortality included diabetes, renal failure, ST-segment-elevation MI and cardiogenic shock.Conclusion: Our clinical event rates were comparable with international registry outcomes. Rates of mortality and stent thrombosis were no higher in patients with DESs than those with BMSs. Although DESs were used in about half the procedures (preferentially for higher-risk lesions), recent trends suggest their use is in decline.

Andrew E Ajani MD, FRACP, FJFICM · Christopher M Reid BA, MSc, PhD · Stephen J Duffy FRACP, MRCP, PhD · Nick Andrianopoulos MB BS, MBiostat · Jeffrey Lefkovits MB BS, FRACP · Alexander Black MB BS, FRACP · Gishel New FRACP, FACC, PhD · Robert Lew MB BS, FRACP, PhD · James A Shaw MB BS, FRACP, PhD · Bryan P Yan MB BS, FRACP · Ronen Gurvitch MB BS · Ali Al-Fiadh MB BS · Angela L Brennan RN, CCRN · David J Clark MB BS, FRACP

Objective: To assess the effectiveness of the PAST (Pre-hospital Acute Stroke Triage) protocol in reducing pre-hospital and emergency department (ED) delays to patients receiving organised acute stroke care, thereby increasing access to thrombolytic therapy.Design: Prospective cohort study using historical controls.Setting: Hunter Region of New South Wales, September 2005 to March 2006 (pre-intervention) and September 2006 to March 2007 (post-intervention).Participants: Consecutive patients presenting with acute stroke to a regional, tertiary referral hospital.Intervention: PAST protocol, comprising a pre-hospital stroke assessment tool for ambulance officers, an ambulance protocol for hospital bypass for potentially thrombolysis-eligible patients, and pre-hospital notification of the acute stroke team.Main outcome measures: Proportion of patients who received intravenous tissue plasminogen activator (tPA), process of care time points (symptom onset to ED arrival, ED arrival to tPA treatment, and ED transit time), and clinical outcomes of patients treated with tPA.Results: The proportion of ischaemic stroke patients treated with tPA increased from 4.7% (pre-intervention) to 21.4% (post-intervention) (P < 0.001). Time point outcomes also improved, with a reduction in median times from symptom onset to ED arrival from 150 to 90.5 min (P = 0.004) and from ED arrival to stroke unit admission from 361 to 232.5 minutes (P < 0.001). Of those treated with tPA, 43% had minimal or no disability at 3 months.Conclusions: Organised pre-hospital and ED acute stroke care increases patient access to tPA treatment, which is proven to reduce stroke-related disability.

Debbie A Quain BA(Nursing) · Mark W Parsons PhD, FRACP · Allan R Loudfoot MBA · Neil J Spratt PhD, FRACP · Malcolm K Evans RN, BA(HealthManagement) · Michelle L Russell RN, CM · Angela T Royan BNursing · Andrea G Moore BNursing · Ferdinand Miteff MB ChB · Carolyn J Hullick FACEM · John Attia MD, PhD, FRCPC, FRACP · Patrick McElduff BMath, PhD · Christopher R Levi BMedSci, FRACP

Objectives: To describe the epidemiology of mumps and examine potential factors underlying the recent increase in the incidence of mumps in Australia.Design, setting and participants: Analytical descriptive study, for all Australian states and territories, of mumps notifications (1994–2007); hospitalisations for mumps (1994–2005); and mumps seroprevalence in a nationally representative sample of 2787 subjects (1997).Main outcome measures: Incidence of notifications and hospitalisations for mumps; seropositivity by birth cohort.Results: Notified mumps cases increased from 60 in 2002 to 231 in 2005 and 512 in 2007. Between 1994 and 2005, there were 605 hospitalisations for mumps. Mumps seropositivity in all states and territories in 1997 was high (range, 87.1%–94.3%). The predominant age group affected by mumps shifted to adults over time: between 2005 and 2007, 41% of cases occurred among people aged 20–29 years. Cases were concentrated among the birth cohort of 1978 to 1982, who had higher rates of notifications and hospitalisations for mumps and a lower seropositivity rate (92% [95% CI, 89%–94%]) than other birth cohorts.Conclusions: The birth cohort of 1978 to 1982 was too old to reliably receive a second dose of measles–mumps–rubella (MMR) vaccine in the 1998 Australian Measles Control Campaign and too young to have had mumps infection. Renewed efforts to maximise two-dose MMR coverage are important for prevention of mumps and measles in young adults.

Padmasiri E Aratchige MB BS, MSc, MD · Peter B McIntyre FRACP, FAFPHM, PhD · Helen E Quinn BSc(Hons), MAppEpi, PhD · Gwendolyn L Gilbert FRACP, FRCPA, MD

Objective: To determine the prevalence and associated epidemiological features of Helicobacter pylori infection in child refugees in Western Australia.Design and participants: Cross-sectional study of 193 eligible African refugee children (aged < 16 years) at their initial health assessment after resettlement in Australia between 1 February and 30 November 2006.Main outcome measures: (i) Prevalence of H. pylori infection determined by monoclonal faecal antigen enzyme immunoassay testing (MFAT); (ii) associations of H. pylori infection with epidemiological factors (age, sex, transit through refugee camps, comorbidities and treatment interventions).Results: MFAT was performed in 182 of the 193 children; 149 of these 182 (82%) had H. pylori infection. Age was an independent predictor of H. pylori infection (odds ratio [OR], 1.18; 95% CI, 1.07–1.31). No sex differences were observed. Premigration antimalarial therapy (with sulfadoxine–pyrimethamine and artesunate) significantly reduced the prevalence of H. pylori infection (age-adjusted OR, 0.33; 95% CI, 0.15–0.75).Conclusion: African refugee children have a high prevalence of H. pylori infection. Increasing age is a strong predictor of infection and antimalarial treatment may have a protective effect.

Sarah Cherian MB BS(Hons), FRACP · David Forbes MB BS, FRACP · Frank Sanfilippo BPharm, PhD · Angus Cook MB ChB, PhD · David Burgner MB ChB, FRACP, PhD

Objective: To assess the impact of a long-term comprehensive control program for sexually transmitted infections (STIs) in remote Aboriginal communities in Central Australia, and to investigate a recent rise in gonorrhoea prevalence.Design: STI prevalence was determined from annual, cross-sectional, population-wide, age-based screening, 1996–2006. During 2006, gonococcal isolates were obtained by on-site culture and tested for antimicrobial susceptibility.Setting: Six remote clinics on the Anangu Pitjantjatjara Yankunytjatjara (APY) Lands, South Australia, which are served by Nganampa Health Council, an Aboriginal community-controlled health service.Participants: All resident Aboriginal people aged 14–40 years at the commencement date of each annual population-wide screen.Main outcome measures: Multivariable logistic regression models were used to compare prevalence of chlamydial infection, gonorrhoea and syphilis measured during each annual population-wide screen; antimicrobial susceptibility of gonococcal isolates obtained in 2006.Results: Between 1996 and 2003, there was a significant reduction in prevalence of gonorrhoea and chlamydial infection, by 67% and 58%, respectively. Subsequently, chlamydia prevalence rate plateaued, but there was a rapid rise in prevalence of gonorrhoea. Syphilis prevalence decreased linearly over the study period (odds ratio, 0.81; P < 0.001). During the first 6 months of 2006, 89 gonococcal isolates were obtained, 39 through on-site culture during the 6-week screening period, and all were sensitive to penicillin (in the less-sensitive category).Conclusions: The decrease in STI prevalence asssociated with the program was maintained until 2006 for chlamydial infection and syphilis, but not for gonorrhoea, which rose in prevalence after 2003. There was no change in antimicrobial resistance to explain this rise, and gonorrhoea transmission dynamics and travel of core transmitters to regions without STI control programs might be responsible.

Rae-Lin Huang MB BS(Hons), MPH, FRACGP · Paul J Torzillo MB BS, FRACP, FJFICM · Vivien A Hammond RN, RM, GradDipNursing · Stephanie T Coulter BLabMed · Adrienne C Kirby BSc(Hons), MSc

To the Editor: In the preceding article, we report on a substantial rise in prevalence rates of gonorrhoea in a population in remote Central Australia.1 This rise occurred in the context of a sustained major reduction in sexually transmitted infections (STIs) in the region, achieved by a comprehensive program of STI control, described in the article1 and previously.2 We found that the gonorrhoea outbreak was not due to penicillin resistance of the causative organism, and we hypothesise that it was due to the introduction and dominance of a more infectious clone.3,4 This rise in gonorrhoea in a region widely acknowledged to have the most successful STI control program in the country prompted several commentators to argue that both this program, and screening as a measure for STI control in remote Indigenous communities, had failed, and to advocate a range of other approaches.5 We recently completed the analysis of the 2008 annual population-wide STI screen, which achieved a 78% participation rate among eligible participants. These data strongly suggest that the gonorrhoea outbreak seen over the previous 4 years has been controlled (Box). Furthermore, the current prevalence rates are among the lowest seen in the past decade. These findings suggest that a comprehensive STI control program, such as that delivered by the Nganampa Health Council, can not only reduce STI rates, but also control outbreaks, provided the program is sustained. During most of the past decade, the prevalence of syphilis remained below 1%, of chlamydial infection below 6%, and of screening test-positive gonorrhoea below 8%, as measured during the annual population-wide screens. This program should be replicable in other regions, if appropriate resources and expertise are applied, thus providing an opportunity to improve an important area of Indigenous health using current public health knowledge. Age-adjusted prevalence rates of chlamydial infection, gonorrhoea and syphilis among 14–40-year-olds on the APY Lands, 1996–2008 APY = Anangu Pitjantjatjara Yankunytjatjara.

Rae-Lin Huang · Paul J Torzillo · Adrienne C Kirby

Objective: To examine the association between long-term health conditions and being out of the labour force among older Australians.Design, setting and participants: Retrospective analysis of cross-sectional data from the Australian Bureau of Statistics 2003 Survey of Disability, Ageing and Carers for people aged 45–64 years.Main outcome measures: Rates of premature retirement associated with ill health; odds ratios of being out of the labour force associated with each long-term health condition and number of conditions; weighted population estimates; estimates of gross domestic product lost as a result.Results: 9198 people surveyed were aged 45–64 years, 3010 of whom were not in the labour force. Of these, 1373 (45.6%) had retired because of a chronic health condition, most commonly a back problem (10.4%), or arthritis and related disorders (8.6%). When adjusted for age and sex, all conditions studied except diseases of the ear and mastoid process, other endocrine/nutritional and metabolic disorders, noise-induced deafness or hearing loss, and high cholesterol were significantly associated with being out of the labour force. Extrapolating from these results, an estimated 663 235 older Australians were not working because of ill health, reducing Australia’s gross domestic product by around $14.7 billion per annum.Conclusion: Prevention of long-term health conditions may help older Australians remain in the labour force longer, thereby increasing revenue to fund health care for the ageing population.

Deborah J Schofield BSpPath, GradDipComp · Rupendra N Shrestha BSc, MSc(Statistics) · Megan E Passey BMed(Hons), MPH, MSc · Arul Earnest DLSHTM, MSc · Susan L Fletcher BAppSc(Psych), PGDipPsych

Objective: To determine if a robust clinical review process can influence an organisation’s response to adverse patient outcomes.Design and setting: Retrospective analysis of the activity and outputs of the Clinical Review Committee (CRC) of a university-affiliated tertiary hospital from 1 September 2002 to 30 June 2006.Main outcome measures: Engagement of clinicians (number on CRC, number interviewed for the clinical review process, number of specific referrals from clinicians); and numbers of cases reviewed, system issues identified, recommendations made to the hospital board, and ensuing actions.Results: A multidisciplinary CRC with 34 members established a robust clinical review process and identified 5925 cases for initial case review. Of these, 2776 (46.8%) fulfilled one or more of the specified criteria for adverse events and progressed to detailed review; 342 of these (12.3%) were classed as serious or major. A total of 317 staff (11%) were interviewed, and 881 system issues were identified, resulting in 98 specific recommendations being made to the Clinical Board and implementation of 81 practice changes (including seven hospital-wide projects) to improve patient care.Conclusion: A robust, multidisciplinary clinical review process with strong links to managers and policymakers can influence an organisation’s response to adverse patient outcomes and underpin a clinical governance framework.

Imogen A Mitchell FRACP, FJFICM · Bobby Antoniou RN, GradDipCritCare · Judith L Gosper RN, RM · John Mollett · Mark D Hurwitz FRACP, FCP(SA) · Tracey L Bessell BPharm, MPH, PhD

Mumps has re-emerged as an infection in the developed world. Its epidemiology has changed, with the majority of cases now primarily affecting adolescents and adults. While mumps is easily suspected if parotitis is present, parotitis is absent in 10%–30% of symptomatic cases. Mumps is a systemic infection with a variety of extra-parotid complications. In Australia, mumps diagnosis is confirmed by antibody testing and reverse transcriptase-polymerase chain reaction techniques. Suitable specimens for testing are serum, saliva, urine and cerebrospinal fluid. Treatment is generally supportive, although intravenous immunoglobulin therapy may have a future role in mumps management. Interferon alpha-2b treatment may be considered specifically for mumps epididymo-orchitis. Mumps vaccine is included in the measles–mumps–rubella (MMR) vaccine. In Australia, this vaccine is routinely administered at the ages of 1 and 4 years. Serious reactions to the mumps components of the MMR vaccine are rare.

Sanjaya N Senanayake BSc(Med), MAppEpid, FRACP

Australia is an attractive workplace for overseas-trained specialist (OTS) anaesthetists. The path to recognition of the qualifications and experience of OTS anaesthetists is, in my opinion, bogged down in an overzealous assessment procedure. The Australian and New Zealand College of Anaesthetists (ANZCA) is a self-proclaimed professional body that is not subject to regulation by the federal government. Medical authorities such as the Australian Medical Council and state medical boards have no influence on ANZCA’s assessment criteria and procedures. In my opinion, the current state of affairs with regard to assessment of OTS anaesthetists can not be justified.

Soren Sondergaard DMSc, MD

To the Editor: Thank you for the opportunity to reply to Sondergaard’s article.1 The Australian health care system is highly complex, involving both federal and state jurisdictions. The Council of Australian Governments, consisting of representatives of national, state and territory governments, makes decisions that have an impact on the respective health ministers and their health departments, state/territory medical boards, and the Australian Medical Council (AMC). The Australian and New Zealand College of Anaesthetists (ANZCA), one of 12 medical colleges accredited by the AMC, is acknowledged by all of these entities as the body responsible for education, training, and standard-setting in anaesthesia. ANZCA’s policies and processes relating to overseas-trained specialists (OTSs) have been scrutinised by the AMC, which requires annual reports from the College; by the Australian Health Workforce Officials Committee (AHWOC); and by the Australian Competition and Consumer Commission (ACCC), in their review of all colleges. ANZCA is progressively altering its OTS processes, as agreed with these authorities. In accordance with AMC guidelines, OTSs are assessed against published criteria, based on their training, qualifications, skills and experience. Anaesthetists from Scandinavia, like those from many other developed countries, are usually assessed as being “partially comparable” to Australian-trained anaesthetists and require 12–24 months’ supervised practice and a performance assessment under Australian conditions before being eligible for Fellowship of ANZCA. It is possible for applicants to be considered “substantially comparable”, providing they are judged to have similar qualifications by training and examination (suitable applicants from the United Kingdom would come under this category). It is worth noting that there are no reciprocal arrangements in the field of anaesthesia between Australia and any other countries. All applicants must be assessed under the same agreed processes as governed by the AMC and the federal government. In relation to Sondergaard’s article, there are some errors that need to be corrected. The AMC is not responsible for “full registration” of doctors — that is the prerogative of the state and territory medical boards, each of which has its own legislative requirements. It is worth noting that, in South Australia, only Fellows of ANZCA can practise as specialist anaesthetists. “Physician assistants”, who in any event are fully supervised by specialists, are currently performing pilot roles in some hospitals to determine the feasibility of assistant anaesthetist roles. This does not have anything to do with “circumventing” the College’s regulations. The quotes in Sondergaard’s Box 1 are extracts from the AMC’s 2002 accreditation report on ANZCA and relate to more flexible methods of recognising the qualifications and training of OTSs. ANZCA has reported annually to the AMC on progress in this matter. With guidance from the AMC and regulatory authorities, changes are continually being made to bring all medical colleges into line with each other. The quote in Box 2 is an extract from the 2005 ACCC/AHWOC review of Australian specialist medical colleges. It refers to all the colleges, and all are working with federal, state and territory regulatory authorities and the AMC to address the issues raised. As the registering authorities, state and territory medical boards act independently (as in the Bundaberg Hospital case2) and do not necessarily accept the advice of any college. The Editor of the Medical Journal of Australia has published a number of editorials that have dealt very fairly with a range of relevant difficult issues, including OTSs,3 task transfer4 and the need for change.5 In conclusion, Sondergaard requested and was granted an interview with ANZCA’s OTS Interview Panel, and his application was dealt with according to the College’s usual processes, following guidelines laid down by the AMC and medical boards. He has lodged an appeal with ANZCA.

Leona F Wilson

Tumour necrosis factor inhibitors have revolutionised the management of Crohn’s disease, but reports of a possible association between concomitant infliximab and immunomodulator therapy and hepatosplenic T-cell lymphoma (a rare form of aggressive non-Hodgkin’s lymphoma) have emerged. We describe the first case in Australia of hepatosplenic T-cell lymphoma in a patient who had been treated with infliximab and immunomodulators for Crohn’s disease. Clinical recordIn January 2006, a 39-year-old man of European background presented with severe sepsis of unidentified source. This was complicated by hypotension and multiorgan failure, including renal impairment, abnormal liver function and myocardial injury, and he required intensive care. He had a 13-year history of active perianal and ileal Crohn’s disease, which had been treated with prednisolone (varying doses) continuously from 1993, and with azathioprine (2–2.5 mg/kg) from 1993 to 1994 and then continuously from August 1999 (after drainage of a perianal abscess). He had also received three doses of infliximab (5 mg/kg) between November 1999 and January 2000, which achieved a partial response. Azathioprine and prednisolone therapy were continued after infliximab therapy, with relatively good control of symptoms. Results of full blood examinations during azathioprine therapy were within normal ranges. On admission, azathioprine therapy was discontinued, but corticosteroids were continued. Blood cultures grew methicillin-sensitive Staphylococcus aureus. Full blood examination revealed mild anaemia (haemoglobin level, 91 g/L; reference range [RR], 130–170 g/L) and mild lymphocytopenia (white blood cell count, 0.6 × 109/L; RR, 4.0–11.0 × 109/L), but other parameters were initially normal. Biochemical analysis revealed acute renal failure (potassium, 6.5 mmol/L [RR, 3.5–5.0 mmol/L]; bicarbonate, 21 mmol/L [RR, 22–31 mmol/L]; creatinine, 580 μmol/L [RR, 60–110 μmol/L]; and urea, 19.9 mmol/L [RR, 2.5–8.3 mmol/L]). Liver function tests revealed transaminitis (alanine aminotransferase, 432 U/L [RR, < 55 U/L]; aspartate aminotransferase, 2246 U/L [RR, < 50 U/L]; and bilirubin, 53 μmol/L [RR, < 19 μmol/L]). There was also evidence of coagulopathy (international normalised ratio [INR], 2.8 [RR, 0.8–1.3]), grossly abnormal levels of inflammatory markers (C-reactive protein, 218 mg/L [RR, < 8 mg/L]; erythrocyte sedimentation rate, 140 mm/h [RR, 2–14 mm/h]), and elevated troponin I levels (5.24 μg/L; RR, < 0.10 μg/L). Nine days after admission, the patient developed pancytopenia (haemoglobin level, 78 g/L; white blood cell count, 0.9 × 109/L; and neutrophil count, 0.3 × 109/L [RR, 2.0–8.0 × 109/L]). This was attributed to drug-induced suppression of bone marrow, and granulocyte-colony stimulating factor (G-CSF) was administered. On discharge, leukocyte and neutrophil levels were normal, G-CSF therapy was discontinued, and follow-up was arranged. Ten days after the patient was discharged, he re-presented with a 2-day history of fever, malaise and non-specific abdominal pain. Physical examination revealed a temperature of 38.5°C, tachycardia (heart rate, 110 beats/min), and hypotension (blood pressure, 100/60 mmHg); cardiovascular and respiratory systems were otherwise unremarkable. Abdominal examination revealed new hepatosplenomegaly, with no stigmata of chronic liver disease. There was no clinical evidence of lymphadenopathy. Repeat haematological testing revealed pancytopenia: haemoglobin level, 100 g/L; white blood cell count, 0.7 × 109/L; platelet count, 48 × 109/L (RR, 140–400 × 109/L); neutrophil count, 0.0 × 109/L; and lymphocyte count, 0.5 × 109/L (RR, 1.2–4.0 × 109/L). A blood film showed atypical lymphocytes and blast cells. Serum lactate dehydrogenase level was grossly elevated at 5351 U/L (RR, 210–420 U/L), and liver function tests showed abnormal results. Abdominal computed tomography confirmed gross splenomegaly, with the spleen being 24 cm long on its major axis, but no lymphadenopathy. Microscopic examination of a liver core biopsy specimen revealed an atypical lymphoid infiltrate in the sinusoids, especially around central veins (Box), with immunophenotype bcl-2+, CD3+, CD43+, Ki67+, ALK1 −, bcl-6 − , CD5 −, CD10 −, CD20 −, CD30 −, CD79 − and cyclin D1 − . Subsequent genetic studies of a bone marrow biopsy specimen revealed rearrangement of the T-cell receptor γ-chain gene, consistent with hepatosplenic T-cell lymphoma (HSTCL). The lymphoma was treated with one cycle of cyclophosphamide, mesna, dexamethasone, doxorubicin and vincristine, which achieved a partial response, but subsequent salvage chemotherapy with ifosfamide, carboplatin and etoposide did not arrest disease progression. In June 2006, the patient received a sibling allogeneic bone marrow transplant, after conditioning with etoposide and total body irradiation.1 The patient remained largely free of disease after transplantation, with monitoring on an outpatient basis and adjustment of immunosuppression as clinically indicated. However, in April 2007, he was hospitalised for investigation of diarrhoea and worsening liver function. He underwent colonoscopy and biopsy of a caecal polypoid mass; results of histological examination were consistent with recrudescence of HSTCL. Positron emission tomography showed multiple sites of relapse. A palliative approach was adopted, and the patient died in June 2007. DiscussionWe describe the first case, to our knowledge, in Australia of HSTCL in a patient with Crohn’s disease who had been treated with infliximab and immunomodulators. Inhibitors of tumour necrosis factor (TNF) such as infliximab have shown great efficacy in the treatment of luminal and fistulising Crohn’s disease, as well as ulcerative colitis.2 Infliximab is a chimeric (human/murine) monoclonal antibody that binds to human TNF-α. It was approved by the United States Food and Drug Administration (FDA) in 1998 for the treatment of moderate-to-severe Crohn’s disease when response to immunomodulator therapy is inadequate. More recently, the FDA has expanded the indication to include ulcerative colitis refractory to conventional treatment. In Australia, the Pharmaceutical Benefits Advisory Committee recently approved TNF inhibitors for the treatment of Crohn’s disease. However, the safety of such biological therapy is a concern: a recent meta-analysis reported a threefold increase in the risk of malignancy (solid and haematological) with the use of anti-TNF therapy in rheumatoid arthritis patients.3 This increase may be partly due to severity of disease or to other aspects of disease management. Interestingly, to date there have been no reports of HSTCL in rheumatoid arthritis patients. The TREAT (Crohn’s Therapy Resource, Evaluation and Assessment Tool) registry, which monitors over 3000 patients with Crohn’s disease who have been treated with infliximab, has not reported an increased incidence of lymphoma.4 However, it is estimated that a substantially larger number of patients would need to be monitored to detect a significant increase in a rare adverse event, such as lymphoma.5 Between 2000 and 2006, the Adverse Drug Reactions Advisory Committee received 319 reports involving anti-TNF therapy, including five cases of lymphoma.6 HSTCL is a rare form of aggressive non-Hodgkin’s lymphoma that comprises 5% of peripheral T-cell lymphomas. Reports of approximately 120 cases have been published worldwide. Eight cases of HSTCL have been identified from the post-marketing infliximab safety database run by the FDA, which seems more than expected, all in young men with a history of Crohn’s disease and concomitant use of azathioprine or mercaptopurine.7 In addition, there have been 15 reports of ‘‘T-cell lymphoma’’ in patients treated with infliximab, but data from the Adverse Event Reporting System were limited.7 Furthermore, there have been no reports of HSTCL associated with other anti-TNF therapies (eg, etanercept and adalimumab) used for any indication.7 Of note, there have been four reports of HSTCL in patients who received azathioprine or mercaptopurine alone for 4–6 years.8 The case we describe differs substantially from previously reported cases of HSTCL associated with concomitant infliximab and immunomodulator treatment in Crohn’s disease. To our knowledge, it involves the oldest patient and longest lead time (72 months) reported to date. Patients in most other cases have been younger than 22 years, with a lead time of less than 58 months7 (unpublished data, Centocor, Horsham, Pa, USA). Recent data suggest an association between lymphoma — especially HSTCL — and concomitant use of infliximab and immunomodulators in Crohn’s disease. However, the mechanism of this possible association remains unclear. The use of biological therapy as a “bridge” to stabilise the disease, while waiting for immunomodulators to become clinically effective, may need to be considered with increased caution. Immunomodulators might need to be avoided after infliximab therapy, and alternative treatments considered. These may include maintenance with ongoing biological therapy alone, use of newer anti-TNF therapy, or earlier surgery (especially in young men). Until further evidence emerges, long-term surveillance of patients who have used biological therapy is warranted. In particular, biological therapy should be used cautiously in the management of refractory inflammatory bowel disease. The decision to use infliximab should be tempered by observations of an association with HSTCL, and strategies for long-term maintenance therapy need to be developed. Liver core biopsy specimen of a patient with Crohn’s disease, following infliximab and immunomodulator therapy Low magnification (A: haematoxylin and eosin stain; original magnification, × 10) and high magnification (B: CD3 stain; original magnification, × 40) views of a liver core biopsy specimen, showing prominent atypical lymphoid infiltrate in the sinusoids and around central veins, and occasional red cell extravasation. The high magnification shows strong CD3 staining (arrows), findings consistent with hepatosplenic T-cell lymphoma.

Musa Drini MB BS · Peter J Prichard MD, FRACP · Gregor J E Brown PhD, FRACP · Finlay A Macrae MD, FRACP, FRCP

A 51-year-old woman presented with a 2-day history of left iliac fossa pain. Axial and coronal computed tomography images (Figures) revealed a linear intraluminal foreign body (arrows) impacted in the wall of a blind-ended loop of small bowel over the midline, consistent with a fishbone in a Meckel diverticulum. Small bowel dilatation due to ileus (D) and inflammatory stranding in the peritoneal fat (S) were also present. Surgery revealed pinpoint bowel perforation caused by a 2 cm long fishbone, from Pampus argenteus (silver pomfret), in a 5 cm long Meckel diverticulum.

To the Editor: We report the case of a 41-year-old woman who sought medical opinion about an unpleasant body odour, first noticed when she was 7 years old. After experiencing ridicule, distress, shame, anxiety and low self-esteem during her school years, she first consulted a doctor about the problem at the age of 17 years, then again 2 years later, followed by a further four doctors over the next 20 years. All dismissed her concerns, and she was repeatedly told that she had a hygiene neurosis. Investigations and treatments during this time included being “sniffed”, vaginal swabs and vaginal cauterisation. Finally, a general practitioner referred her to a dermatologist, who consulted a microbiologist, and the diagnosis of trimethylaminuria (TMAU), or fish malodour syndrome, was confirmed by urinalysis. Now having a name for her condition, she found an Internet-based support foundation and referred herself for genetic counselling. TMAU is caused by an enzyme deficiency due to mutations in the flavin-containing mono-oxygenase 3 (FMO3) gene,1 resulting in excess excretion of trimethylamine in urine, sweat and breath. It is diagnosed by clinical symptoms and urine analysis.2 The characteristic body odour resembling rotting fish can be intermittent, variable and influenced by diet, hormones and medications. Restriction of choline- and carnitine-rich dietary precursors (eg, fish, eggs, soybeans, peas) is difficult to maintain and effective in only 25% of patients.2 Acid soaps and body lotions can often reduce the odour.3 The metabolic and clinical manifestations of TMAU are generally regarded as benign, as there is no associated organ dysfunction. This designation, and the fact that the condition is often unrecognised by doctors, can have important ramifications including missed or delayed diagnosis.4 Affected individuals experience shame and embarrassment, fail to maintain relationships, avoid contact with people who comment on their condition, and are obsessive about masking the odour with hygiene products and even smoking. The malodorous aspect can have serious and destructive effects on schooling, personal life, career and relationships, resulting in social isolation, low self-esteem, depression, paranoid behaviour, and suicide.4, Psychosocial problems resulting from delayed diagnosis, body odour and the lack of cure are considerable, making this a far from “benign” disorder. Recognition of TMAU as a significant clinical entity and increased understanding of the issues patients face are needed. Awareness of the typical patient history would facilitate prompt metabolic diagnosis and pre-empt some of the associated psychosocial sequelae. Referral of patients for genetic counselling enables short-term psychosocial support and family cascade genetic testing. Consultation with a metabolic clinic for dietary management may also be beneficial.

Helen Mountain · Joanna M Brisbane · Amanda J Hooper · John R Burnett · Jack Goldblatt

To the Editor: In their recent case series, Lintzeris and colleagues state that the symptoms leading to hospital presentation were “associated” with naltrexone implants.1 In half of the 12 cases, the symptoms were related to the induction of withdrawal rather than the presence of naltrexone — an important distinction that may not be apparent to all, but of which the authors would be aware. Three of the remaining cases reflect the complexities of pain management in patients being treated with naltrexone, which are not restricted to those with implants. This dilemma is also encountered in patients taking buprenorphine. Patient 8 had an anxiety disorder, and his symptoms probably related to an absence of opiate; and Patient 9 had symptoms that were probably due to cocaine use. Abstinence is a patient’s choice, and naltrexone can be effective in minimising the risks associated with the consequent lowered tolerance to opioids.2 I am concerned that undue emphasis appears to have been placed on the association of adverse outcomes with implants, with such statements as: “This case series identifies severe adverse events associated with the use of naltrexone implants”. In fact, the causal condition is usually the induction of opiate withdrawal rather than the presence of a naltrexone implant.3 The statement that “Most of these cases (8/12) can be attributed to the naltrexone implant or implantation procedure” is true, but it would be more appropriate to acknowledge that the procedure appeared to be the cause in most cases and that only in one case (Patient 7) could the implant be conclusively implicated. It is not appropriate to refer to difficulties in pain management as “severe adverse events”, as the blockade of the μ opioid receptor is the reason naltrexone is used. In addition, it is important to note that a trial of injectable naltrexone has shown promising results.4 While these facts are acknowledged by the authors, their emphasis on the association of adverse events with the presence of a naltrexone implant, rather than demonstrating causality, is concerning. Similarly, this approach appeared in Gibson and colleagues’ earlier commentary on overdose deaths in patients with naltrexone implants.5 Nonetheless, I commend the authors for raising the issues relating to the need to examine the role of naltrexone implants in treating opioid dependence and exploring alternatives to the use of rapid detoxification. I also endorse the need for thorough assessments, treatment planning and evaluation in patients undergoing such therapies.

D Martyn Lloyd-Jones

To the Editor: We read with interest the report of 12 hospital presentations related to the use of naltrexone implants.1 The accompanying editorial highlights how the rigorous scrutiny required to evaluate the efficacy and safety of this procedure is lacking.2 Regrettably, this study is likely to distort rather than inform the debate. Lintzeris and colleagues1 only identified patients with naltrexone implants who were referred to the Drug and Alcohol Consultation–Liaison services, not all patients presenting to the study hospitals. Additionally, the authors did not follow the methodology of chart reviews, as recommended by Gilbert and colleagues.3 Four of the 12 patients clearly had problems unrelated to their naltrexone implants. There was no attempt to identify the number of naltrexone implantations performed (ie, the denominator), nor was there any attempt to compare the naltrexone group with others being treated with agents such as methadone or buprenorphine. In 2003, we published our experience of naltrexone-accelerated detoxification in the emergency department (ED) of Sir Charles Gairdner Hospital.4 The hospital’s clinical toxicology service is based in the ED, and the hospital is located 2 km from the only private clinic in Perth that was using naltrexone during the study period in 2001. Working collaboratively with this clinic, patients developing complications from detoxification were referred to our service. In 6 months, 42 patients (7% of all those receiving naltrexone treatment) presented to the ED — 17 within 24 hours of treatment and 31 within 48 hours. Gastrointestinal symptoms of withdrawal were present in 18 patients, and central nervous system symptoms of withdrawal (predominantly agitation) in 14. Two patients required intubation for airway compromise secondary to a combination of agitation and chemical sedation. In 23 patients receiving naltrexone implants (rather than oral therapy), three developed infections and three complained of local pain. The mean length of stay for all patients was 18 hours (compared with 2.3 days in Lintzeris et al’s study), with the longest stay being 92 hours for one of the patients admitted to the intensive care unit. During the study period and the subsequent 6 months, four deaths of individuals who had undergone naltrexone-accelerated detoxification were reported to the Coroner, all being classified as probable drug overdose, probably opioid. None of these patients had presented to the hospital’s ED during the study period. It is important that good data are made available to inform the debate on naltrexone implants in the management of opioid dependence. Better communication and collaboration between clinics using naltrexone, EDs, and alcohol and drug services will improve the care of these patients.

Mark Little · Lindsay M Murray

To the Editor: The study by Lintzeris and colleagues1 and the associated editorial2 criticise naltrexone implants. However, the study is replete with errors regarding the 12 cases reported. A key to successful rapid opioid detoxification (ROD) is octreotide. An agonist of non-μ gut receptors, it prevents gastrointestinal symptoms of withdrawal.3,4 Earlier aggressive dosing is more effective. Only two of the six patients reported as having precipitate opiate withdrawal received octreotide, administered in both cases by their local doctor. Presumably this doctor performed the ROD. Why was the octreotide administration not repeated? Patient 7 was reported as having a localised abscess at the implant site. These may arise a variable period after implant insertion. Magnesium-monostearate, used as a binder in naltrexone and many other implants, may liquefy, resembling pus. Results of repeated microbiological examination are always negative. An actual infection at the implant site is rare. Patient 8 demonstrates that many patients use opiates and other drugs as self-medication for their psychiatric symptoms. It is regrettable that this patient’s psychiatrist was not involved in the decision making. Was the doctor doing the ROD aware that there was a psychiatrist involved? Transient psychosis may occur following ROD; use of antipsychotics is appropriate. Arrhythmias (as diagnosed in Patient 9) are a textbook in themselves. Pre-ROD preparation requires assessment for comorbidities including bacterial endocarditis and arrhythmia syndromes. Substance-misusing patients often use arrhythmogenic agents such as amphetamine. No mention was made of precipitate withdrawal symptoms in this patient, so an electrolyte abnormality is unlikely. Patients 10–12 had problems that were unrelated to the naltrexone implants and poorly managed by the receiving hospital. Naltrexone has 2 log the affinity for opiate receptors to that of morphine and most other opiates.5 This means that 1 mg of naltrexone will block 1 g of another opiate. Using opiates to treat patients in these circumstances therefore makes no sense. Ketamine, local anaesthetic blocks, paracetamol and non-steroidal anti-inflammatory drugs are alternative options for treatment. To compare naltrexone implants with thalidomide2 is emotive; all the components of the implants are approved agents. Similarly, to say they have not been fully tested is to condemn all products supplied by compounding pharmacists. Testing of serum naltrexone levels (to assess whether an implant is still working) is not rebatable under Medicare, so is expensive to perform. Research is a necessary component of naltrexone implants, and I would happily cooperate with any of the study authors in coordinating and performing such research.

Michael P W Kozminsky

To the Editor: A recent article by Lintzeris and colleagues,1 ostensibly about naltrexone implants, actually has little to do with implant treatment. Only one of the 12 reported cases involved problems linked specifically to naltrexone (antagonist) treatment being administered in implanted rather than oral form. This case involved infection at the implant site — undesirable, certainly, but about as noteworthy as the occasional infections that occur after abdominal surgery or breast implantation, despite antibiotic prophylaxis and careful technique. The remaining 11 admissions reflected not implant use but either the procedure of rapid antagonist induction (RAI) or the desired pharmacological effects for which naltrexone was prescribed in the first place (and one admission for unrelated pneumonia). RAI is needed because conventional antagonist induction requires complete opiate withdrawal before starting naltrexone. As true conventional withdrawal completion rates, even with inpatients, are typically below 30%,2 conventional techniques will typically achieve only derisory naltrexone induction rates. Various forms of RAI or accelerated induction, with induction rates typically around 100%, are more effective and more cost-effective.3,4 Similarly, complaining that pain management is difficult in naltrexone patients is like complaining that patients being treated with anticoagulants bleed or bruise more after surgery or injury. Fortunately, effective non-opiates (notably subanaesthetic ketamine) exist. The acute post-RAI problems seen in this case series would have been identical had every patient undergone RAI to oral rather than implanted naltrexone. Indeed, while 150 mg orally could block opiate analgesia for up to 72 hours, implants produce low but consistently effective naltrexone levels, disappearing within hours of implant removal, should that be unavoidable. I agree, however, that RAI clinics should optimise immediate post-RAI management. A recent conference featured the first presentations of the first four randomised controlled trials of implanted naltrexone. Three showed statistically and (more importantly) clinically significant advantages over “as usual” post-withdrawal treatment,5 or over oral naltrexone and placebo implants.6,7 The fourth,8 comparing implants with methadone maintenance in pre-release prisoners, had only modest statistical power (n = 21), but implant patients had less dropout and used about 50% less heroin. We already know that long-acting implants can largely prevent the opioid overdoses that are otherwise an intrinsic and sometimes lethal hazard of all abstinence-based programs;9 and that all implants prevent the otherwise high relapse rates typical of the first 4–6 weeks after detoxification.10 Surgeons, I am shocked to discover, have been using anaesthesia for 162 years without a comparable placebo-controlled evidence base.

Colin L Brewer

In reply: Our case series1 aimed to alert health practitioners to the types of hospital presentations associated with naltrexone implants and issues in managing such patients, particularly as these have not been well documented in the medical literature. These letters appear critical of any such communication. We maintain that our article1 identifies important clinical issues — such as the difficulties of providing analgesia in patients with implants of uncertain duration of effect (whereby the treating hospital staff have no way of knowing whether the implant is providing “active” plasma levels of naltrexone, given the lack of a licensed product). Ultimately, our study could not hope to provide the type of data that can only be accurately determined by independent clinical trials — such as the incidence rate of serious adverse events (as suggested by Little and Murray) or how these are most effectively managed (eg, the role of octreotide compared with ondansetron for managing protracted vomiting, as suggested by Kozminsky). A key complaint from the correspondents is whether the serious adverse events of opiate withdrawal and dehydration were caused by the naltrexone implant or by the rapid opioid detoxification process that accompanied the implant’s insertion. We acknowledged the difficulty of attributing causality in our article, but, unlike the correspondents, we cannot dismiss the possibility that the naltrexone implant may have contributed to the severity or duration of the opiate withdrawal syndrome. Clearly, there is a pharmacological basis as to how the presence of naltrexone from an implant may contribute to opiate withdrawal in a recently detoxified individual. Research comparing rapid opioid detoxification plus a naltrexone implant with conventional detoxification plus a naltrexone implant is required, to allow an assessment of whether naltrexone implants themselves contribute to the incidence, severity or duration of any opiate withdrawal. Until such data exist, it is appropriate to associate opiate withdrawal with naltrexone from the implants. We concur with several of the correspondents that further research is required to address many of the issues raised by our study, as well as better communication between service providers, better procedures for management of complications, and better assessment and patient selection. We maintain that naltrexone implants should not be routinely used until research has demonstrated their safety and efficacy, and, importantly, until there is a licensed naltrexone implant product with appropriate regulatory safeguards for patients and providers.

Nicholas Lintzeris

To the Editor: Two studies1,2 and an editorial3 on the vexed issue of medication errors in hospitals have recently appeared in the Journal. Yet none of them have mentioned patients, except as the passive victims of error, and all three have focused exclusively on public hospitals. This is unfortunate as, arguably, the public hospital system is evolving predominantly into a way station along the chronic illness journey within the wider health care system. To quote an 86-year-old patient of mine: “I’ve been taking the same tablets for 30 years, but as soon as I come to hospital they take them away and give me other ones, as if I’m incapable.” Indeed. Could patients continue taking their own regular medications (patients’ own drugs [PODs]) themselves? This would eliminate many errors of transcription and administration and eradicate in-hospital dispensing errors altogether. Although it might seem risky for patients to be given free rein with their medications, this is precisely what they are doing at home, and generally with no supervision whatsoever. Furthermore, if patients do make errors, they are more likely to omit some medications than to take toxic doses, and the consequences of such errors of omission are likely to be minor. If a self-administration system for PODs were to be implemented — including newly prescribed drugs that patients were expected to continue using on discharge, as well as medication strategies used in the home (eg, dosette boxes) — then any errors could be noted and corrected, just as a patient’s mobility is assessed before discharge. On the other hand, it is pertinent to note that many hospital admissions arise from medication errors, with one study demonstrating that up to 30% of admissions in patients aged 75 years and over are medication-related and up to 75% are preventable.4 Clearly, certain groups of patients and certain drugs should be excluded from a POD system. A sample protocol is available from the author on request. Would a POD self-administration system be cost-effective? Such a system has been used successfully in at least one Australian private hospital (in which I have worked) and in several hospitals in the United Kingdom.5 A systematic review has also been undertaken to examine the risks and benefits of the use of PODs, but it did not discuss the issue of self-administration.6 A MEDLINE search from 1996 onwards for patients’ own drugs/medications and self-administration yielded no results. As always, further research is required. What my patient’s comment highlights is that patients do have autonomy and some competence, and that these factors should be taken into account in any strategies aimed at reducing medication errors.

Frank T Formby

To the Editor: In May 2007, I began working in an Australian public hospital as a permanent consultant endocrinologist and physician. There are many differences between the health systems of Australia and Sweden, where I previously worked. In particular, there are striking differences in on-call work for the rostered consultant. In Sweden, the compensation (usually taken as leave) for every phone call, whether it was 1 minute or 30 minutes long, was half an hour during weekday hours of 4.30–9 pm and 7–8 am, and 1 hour during nights and weekends. Compensation also applied to ward rounds on weekends. I also received an on-call allowance of 6 minutes per hour during weekdays and 12 minutes per hour during weekends. During public holidays, rates were tripled. Every year, I earned 5–7 extra weeks on top of my normal 5 weeks of annual leave, some of which was used for well paid locum work, as private practice is almost non-existent in Sweden. Compensation could also be taken as payment; each department had its own agreement with staff on the preferred format, depending on staff levels and budget. The on-call frequency was usually 1 in 10, and the physician on call was responsible for all general internal medicine patients. For most of my career I worked in a large tertiary referral centre,1 but the rostering and compensation were similar in local general hospitals, with the only difference being fewer subspecialties with their own on-call roster. The surgical specialties had the same system. As in Australia, I was rarely called to attend the hospital, although contact during an on-call session was more frequent in Sweden (2–10 v 0–2 phone calls). Junior doctors in Sweden were more inclined to discuss patients and, in some wards, the nurses were instructed to call the physician on call first if the matter could be solved by phone. I was happy for every call, as I received extra compensation, but I never received any further phone calls once the on-call session finished. In Sweden, I did ward rounds on weekends, either of all medical wards in the local general hospitals or only the general medical wards in the tertiary centre when I was physician on call. Before seeing patients, I did a paper round with the nurses, when we decided which patients I needed to see. Patients were not admitted directly under me, but under the specialists responsible for the wards during office hours. However, I was temporarily responsible for all medical patients in the hospital. In Australia, at least in my current hospital, I have “my patients” in the wards (usually spread out over the entire hospital). If something suddenly happens to one of them, the intern, resident medical officer or medical registrar will phone me, even in the middle of the night or if I am away on leave. As the compensation is the same whether phone calls are received or not, and most medical patients are admitted under the physician on call, there is little incentive for the physician to volunteer for extra on-call sessions or to encourage junior doctors to phone.

Henrik Falhammar

COPD medication risk Some medications used in the treatment of newly diagnosed chronic obstructive pulmonary disease (COPD) may increase the risk of mortality, according to US researchers. Over 145 000 patients with newly diagnosed COPD were followed for up to 5 years, during which 32 130 died. A large control group of over 320 000 people were matched to the case patients on the basis of sex, age, and year of diagnosis. Measurements included all-cause mortality, respiratory and cardiovascular deaths, and exposure to COPD medications. Theophylline was associated with an increased risk of respiratory death and ipratropium with a greater risk of cardiovascular death. Inhaled corticosteroids were associated with a reduced risk of cardiovascular death. The researchers point out that a limitation of their study is that potential confounders such as smoking status and severity of COPD were unknown. They call for more research to determine whether the observed associations reflect causal relationships. Ann Intern Med 2008; 149: 380-390 A Mediterranean feast The health benefits conferred by the Mediterranean diet have become well known, particularly in terms of prevention of cardiovascular disease and improved quality of life. The pattern of eating represents that usually consumed among people bordering the Mediterranean Sea — a diet rich in vegetables, fruits, legumes, cereals and fish, with a moderate intake of red wine during meals. Most studies have focused on adherence to the diet overall, rather than relating health benefits to individual nutrients, as this approach reflects actual eating habits more closely. A systematic review with meta-analysis of all available prospective cohort studies between 1966 and 2008 has confirmed that greater adherence to the Mediterranean diet is associated with significant improvement in health status. The cumulative analysis of 12 studies included over 500 000 individuals followed for a period of 3–18 years. Scores were attributed in each of these studies to estimate the degree of adherence to the diet. An increased score was significantly associated with a reduced risk of mortality overall, and a lower risk of death from cardiovascular disease and cancer. The incidence of Parkinson’s disease and Alzheimer’s disease was also reduced in those with greater adherence to the Mediterranean diet. The authors of the study comment that the findings are particularly clinically relevant in terms of public health, the production of guidelines for healthy lifestyles, and the prevention of major chronic diseases. BMJ 2008; 337: a1344 CT vs colonoscopy There is consensus that screening for colorectal cancer is effective, but adherence to guidelines may be suboptimal due to the perceived invasive nature of the procedures involved. The results of a US study assessing the accuracy of computed tomographic (CT) colonography suggest this less invasive test may be an adjunct to improve screening effectiveness. A group of over 2500 asymptomatic participants aged 50 years or older underwent CT colonography followed by optical colonoscopy and histological review of any lesions identified. CT colonography was found to have a sensitivity of 0.90 for large adenomas and cancers, meaning that the technique identified 90% of subjects with a lesion measuring 10 mm or more in diameter. The authors comment that CT colonography is a rapid, acceptable technique with few adverse effects, but that controversy surrounding its effectiveness is probably related to variability in imaging protocols and the qualifications of radiologists performing the procedure. N Engl J Med 2008; 359: 1207-1217 Nasal insulin and diabetes prevention Despite promising results in mouse models, prophylactic insulin administered to infants at risk of developing type 1 diabetes does not appear to prevent or delay the onset of the disease. Finnish researchers analysed the cord blood of over 100 000 babies to identify HLA susceptibility alleles for type 1 diabetes. In a double-blind trial, 224 infants and 40 of their siblings who tested positive for the antibodies were assigned to receive intranasal short-acting insulin or a placebo once a day. After a median of almost 2 years, there was no difference in the number of children diagnosed with diabetes in the treatment and placebo groups. The trial was terminated early due to the inability to show a beneficial effect of intranasal insulin in children with HLA-conferred susceptibility to type 1 diabetes. Lancet 2008; 23 Sep [Epub ahead of print]

Tanya Grassi