Pharmacology

Targeted post-exposure prophylaxis represents a more efficient use of the stockpile than treatment alone With the emergence of H1N1 influenza 09 (novel human swine influenza A[H1N1] 2009), efforts to control the spread and mitigate the impact of this virus have been implemented. The Australian health management plan for pandemic influenza (2008)1 (AHMPPI) outlines a range of strategies aimed at eliminating an outbreak where possible (the “Contain” response), or reducing transmission sufficiently to allow distribution of a targeted vaccine (the “Sustain” response). Following evidence of sustained transmission within Victoria in May 2009, state and territory health departments began implementing the Contain response, with a switch to a modified Sustain response in Victoria within weeks. The AHMPPI recommends liberal distribution of the stockpile of neuraminidase inhibitors (oseltamivir, zanamivir) to constrain influenza transmission.1 This policy was based on modelling studies synthesising the best available evidence, including clear demonstration of the efficacy of antiviral drugs to prevent secondary infection in randomised controlled trials.2,3 Notwithstanding revision of the AHMPPI on 17 June 2009 to incorporate the present “Protect” phase, understanding the rationale for and defining the operational implications of the initial recommendations for antiviral drug use are priorities. Translation of evidence from epidemiological trials into pandemic policy is challenging, given the complexity of real-world factors that influence intervention effectiveness. For example, as has been observed with many vaccines, drug effects on transmission have far greater impact when implemented across a whole population. The chosen antiviral deployment strategy needs to take into account not only the direct effects benefiting the treated individual, but also indirect effects due to changes in subsequent transmission. In a rapidly growing epidemic, these secondary effects are critical to determining the optimal deployment strategy. Mathematical modelling provides a way to systematically investigate these issues. Using a diverse set of assumptions and frameworks, models consistently demonstrate that for a country with a sufficiently large stockpile of antiviral drugs, augmenting a patient treatment policy with targeted post-exposure prophylaxis represents a more efficient use of the stockpile than treatment alone.4-6 Modelling performed in the Australian context, where the number of stockpiled antiviral drug courses is 40% of the population size, has demonstrated that extensive drug distribution for preventive purposes does not compromise the ability to treat infected patients.6 It was therefore recommended in the AHMPPI that prophylaxis should be provided to as many readily identifiable contacts as possible during the Contain response, with continued provision to household contacts during the Sustain response. Provision of continuous pre-exposure prophylaxis to health care workers was also recommended, as this additive burden on the stockpile would not substantially impede efforts to delay the pandemic. Lessons learned so far in the effort to implement these recommendations have highlighted the importance of clear case definitions to guide treatment, particularly when disease is mild. Delays in confirming infection, associated with finite laboratory resources, posed significant challenges for front-line health care workers. This practical issue must be addressed should deployment of existing stockpile reserves be recommended in coming years, as potentially more virulent variants of the present pandemic strain arise. Meanwhile, further research and policy development are required on a largely unaddressed issue. A transmissible drug-resistant variant of the pandemic strain may arise either by de-novo mutation or by reassortment with drug-resistant seasonal strains (eg, the oseltamivir-resistant 2008 seasonal H1N1 strain carrying the H274Y mutation [histidine-to-tyrosine mutation at codon 274]), and its spread would be favoured by widespread antiviral use. The effort to delay the appearance of such a variant might stimulate a change to the recommended strategies for antiviral deployment. To date, novel H1N1 viruses demonstrating oseltamivir resistance have been isolated from individuals in Denmark and Japan.7 Several published models have been used to investigate the potential consequences of antiviral drug resistance and all demonstrate that emergence of a transmissible drug-resistant variant will reduce the effectiveness of antiviral distribution strategies, with obvious consequences for their utility in “buying time” before a targeted vaccine becomes available. Two recent studies have investigated strategies for reducing the negative impact of drug resistance.8,9 Both considered the case where the main stockpile (ie, oseltamivir) is supplemented with a smaller secondary stockpile of another drug (eg, zanamivir). One study considered four strategies for antiviral drug distribution when both drugs are available for treatment and prophylaxis.8 It was concluded that a strategy whereby the smaller stockpile (drug B) is reserved for treatment, while the main stockpile (drug A) is used for prophylaxis, will most effectively delay the peak of the epidemic and result in the lowest overall level of drug resistance. Alternative strategies of random allocation of drug A or drug B to each individual who is prescribed an antiviral drug, or use of drug B followed by drug A, were also shown to have significant benefits over a single-drug policy. Cycling between one drug and the other over a period of weeks or months was shown to be a high-risk strategy and cannot be recommended. Treatment strategies were also considered in the other study, which demonstrated that using a small amount of drug B followed by drug A will reduce the overall attack and greatly reduce the resistant attack rate.9 In addition, the global implications of a two-drug strategy were considered, taking into account regular entry of infectious individuals into countries and regions over the course of a pandemic. It was shown that if the primary source country implements a strategic two-drug distribution policy, any country into which strains are subsequently introduced will gain a significant benefit from implementing a similar policy. Both these studies provide strong evidence for jurisdictions to consider acquisition of a secondary drug to supplement their primary drug stockpile. Whether the stockpiles are deployed in order (drug B, then drug A) or separated for use as treatment only and prophylaxis only would largely depend on logistical constraints and overall feasibility of the alternative strategies. Either strategy is likely to provide significant benefits compared with deployment of a single drug. In a climate of great uncertainty surrounding characteristics of the current influenza outbreak,10 the challenge worldwide is for jurisdictions to implement flexible evidence-based policies for antiviral stockpile distribution that maximise their effectiveness.

James M McCaw BSc, PhD · James G Wood BSc, PhD · Emma S McBryde MB BS, PhD, FRACP · Terry M Nolan MB BS, PhD, FRACP · Joseph T Wu PhD · Marc Lipsitch DPhil · Jodie McVernon BMedSc, MB BS, PhD

The complete, up-to-date guidelines can be found at: http://www.mja.com.au/ public/rop/cheng/che10661_fm.html Since the initial reports of H1N1 influenza (human swine influenza; caused by influenza A/2009/H1N1/swl) in Mexico and the United States in mid April 2009, many thousands of cases have been reported worldwide. At the time of writing, community transmission is becoming established in many areas in Australia, but the number of reported cases is likely to be an underestimate of the true incidence due to policies for testing. These guidelines provide advice to clinicians on the use of antiviral agents for this newly emerged influenza virus. A full version of these guidelines, including all references, has been previously published online (http://www.mja.com.au/public/rop/cheng/che10661_fm.html). Recommendations for diagnosis of influenzaIn areas with established community transmission, patients presenting with an acute febrile respiratory illness (fever with cough and/or sore throat) are considered to have H1N1 influenza 09 and testing is generally not recommended. Early in the course of the pandemic, timely influenza diagnostic tests can be used to enable targeted antiviral treatment, but will be less useful once community transmission is widespread. Health care workers performing nose swabs or other high-risk aerosol-generating procedures (eg, suctioning, bronchoscopy or intubation) should use a particulate respirator (N95, P2 mask or equivalent), eye protection, impervious gowns, gloves, and, where possible, carry out the procedure in a negative pressure room. The use of nasopharyngeal aspirates are not recommended because of the risk to staff. The use of a combined nose–throat swab is recommended for diagnosis. Specimens for viral nucleic acid detection and culture at reference laboratories should be taken for epidemiological surveillance and to monitor for drug resistance. Once the pandemic is established, this is best facilitated through existing sentinel surveillance systems. Recommendations for treatment using antiviral agentsAntiviral treatment has been shown to reduce the duration of symptoms and may also reduce the incidence of lower respiratory tract infection. Neuraminidase inhibitors (oseltamivir and zanamivir) are the antiviral agents of choice for H1N1 influenza 09. Dose recommendations for treatment and prophylaxis are provided in Box 1 and Box 2, respectively. The decision to treat an individual patient, particularly before the results of confirmatory testing are available, depends on three factors: An assessment of the likelihood of influenza, based on the known prevalence of infection in the region, a history of contact and the characteristics of the illness. An assessment of the likely benefits of treatment (Box 3), based on the presence of established complications, comorbidities and risk factors (Box 4), and the time since onset of the illness. The phase of the pandemic and the public health policies regarding distribution of the national stockpile (Box 5). Recommendations for treatment in adultsAntiviral treatment should only be given to patients with confirmed or suspected influenza within 48 hours of symptom onset, except in cases of severe influenza. Treatment should be prioritised for patients with risk factors for poor outcomes, such as the morbidly obese, pregnant women, those with chronic disease (including asthma, cardiorespiratory disease, diabetes and renal failure) or immunosuppression, and those presenting with severe disease. Recommendations for treatment in childrenAntiviral treatment can be given to children as young as 1 year. Parents should be warned of the possibility of rare neuropsychiatric symptoms related to oseltamivir use in children and adolescents. There is a concern regarding central nervous system accumulation of neuraminidase inhibitors in infants aged under 1 year, based on animal data; a treatment decision must balance the potential benefits of treatment with potential toxicity. Recommendations for treatment in pregnant womenAntiviral treatment should be offered to pregnant women with suspected or confirmed influenza because of the risk of severe disease in this group. Oseltamivir and zanamivir are in the Australian Drug Evaluation Committee category B1, with limited evidence suggesting safety. Recommendations for treatment of severe influenzaAntiviral treatment should be given to hospitalised patients with severe influenza infection (especially pneumonia), even if commenced more than 48 hours after the onset of symptoms. Antibiotic treatment should not be given routinely for influenza-like illness, but antibiotic treatment should follow established national guidelines for treatment of community-acquired pneumonia. Recommendations for prophylaxis with antiviral agentsLong-term prophylaxis can be given to first-responder health care workers for up to 6 weeks for oseltamivir and up to 4 weeks for zanamivir. Use of antiviral prophylaxis for these groups should be in the context of agreement to use the national stockpile. Antiviral prophylaxis can be given to health care workers and close contacts of patients with influenza following exposure, and to residents of institutions to terminate outbreaks. Contacts not provided with prophylaxis should have access to early treatment with antiviral agents, where indicated. Updated informationWe acknowledge that the evidence on which these recommendations are based is rapidly changing. In particular, estimates of disease severity and case fatality, and risk factors for severity are poorly defined at present and may influence clinical decision making. We therefore include some resources for further information. Updates to these clinical guidelines will be posted on the websites of the Australasian Society for Infectious Diseases (ASID) (http://www.asid.net.au), the Thoracic Society of Australia and New Zealand (TSANZ) (http://www.thoracic.org.au) and the MJA (http://www.mja.com.au). Australian resources for pandemic influenza, including links to clinical and infection control guidelines (http://www.flupandemic.gov.au) and current information on the H1N1 outbreak (http://www.healthemergency.gov.au and http://www.influenza specialistgroup.org.au). For information on accessing personal protective equipment and antiviral medication, see links below: http://www.emergency.health.nsw.gov.au/swineflu/professionals/index.asp (New South Wales) http://humanswineflu.health.vic.gov.au/practitioners/index.htm (Victoria) http://www.health.qld.gov.au/swineflu/html/hc_resources.asp (Queensland) http://flu.sa.gov.au/Swineflu/InformationforGPs.aspx (South Australia) http://www.public.health.wa.gov.au/3/952/3/human_swine_flu_health_providers.pm (Western Australia) http://www.pandemic.tas.gov.au/what_does_it_mean_to_you/health_sector (Tasmania) http://www.health.nt.gov.au/H1N1_Influenza/General_Information_Resources/index.aspx (Northern Territory) http://health.act.gov.au/c/health?a=da&did=11044035&pid=1242181681 (Australian Capital Territory). 1 Dose recommendations for treatment of influenza Treatment Dose, interval, duration Oseltamivir Adults; children > 13 years 75 mg, twice daily orally, 5 days Renal impairment* 75 mg, daily orally, 5 days Children aged 1–13 years < 15 kg 30 mg, twice daily, 5 days 15–23 kg 45 mg, twice daily, 5 days 23–40 kg 60 mg, twice daily, 5 days > 40 kg 75 mg, twice daily, 5 days Zanamivir Adults 10 mg (2 inhalations), twice daily, 5 days Children > 5 years 10 mg (2 inhalations), twice daily, 5 days * Creatinine clearance, 10–30 mL/min. 2 Dose recommendations for prophylaxis against influenza Prophylaxis Dose, interval, duration Oseltamivir Adults; children > 13 years 75 mg, daily, 10 days Renal impairment* 75 mg, alternate days, 10 days Children aged 1–13 years < 15 kg 30 mg, daily, 10 days 15–23 kg 45 mg, daily, 10 days 23–40 kg 60 mg, daily, 10 days > 40 kg 75 mg, daily, 10 days Zanamivir Adults 10 mg (2 inhalations), daily, 10 days Children > 5 years 10 mg (2 inhalations), daily, 10 days * Creatinine clearance, 10–30 mL/min. 3 Factors to consider in deciding on likely benefits of treatment for H1N1 influenza 09 (human swine influenza) infection Established complications Hospitalised patients Patients with respiratory compromise Patients with pneumonitis or secondary bacterial pneumonia High risk of complications Pregnant women Patients with morbid obesity Indigenous Australians Patients with chronic respiratory disease; other comorbidities (see Box 4) Potential for transmission to others Health care workers and first responders (eg, paramedics) Household contact or carer of high-risk patient Low risk of complications Healthy adults Low likelihood of benefit Presentation > 48 hours after onset of illness High prevalence of circulating influenza strains with resistance to neuraminidase inhibitors Potential risks of treatment Infants < 1 year 4 Patients at risk of complications from influenza infection* Pregnant women Indigenous Australians Patients with: chronic respiratory disease (including asthma and chronic obstructive pulmonary disease); cardiac disease; morbid obesity; chronic diseases (eg, diabetes, chronic metabolic diseases, chronic renal failure, haemoglobinopathies); chronic neurological disorders; or impaired immunity, including HIV infection Homeless people Residents of nursing homes and long-term care facilities Children aged 6 months – 10 years on long-term aspirin therapy Older people (> 65 years) Children < 5 years * Adapted from the Australian immunisation handbook.1 5 Indications for antiviral treatment and prophylaxis for H1N1 influenza 09 (human swine influenza) infection, depending on likelihood of benefit and stage of pandemic Pandemic phase Delay Contain Sustain Protect Epidemiological setting Little or no community transmission; cases identifiable via exposure history Limited community transmission; cases not identifiable via exposure history Community transmission in some regions Widespread community transmission Treatment Patients with established complications Clinically presumed or laboratory-confirmed Clinically presumed or laboratory-confirmed Clinically presumed or laboratory-confirmed Clinically presumed or laboratory-confirmed Groups at risk of complications* Clinically presumed or laboratory-confirmed. Consider treatment > 48 h after onset if severe or not improving Clinically presumed or laboratory-confirmed. Consider treatment > 48 h after onset if severe or not improving Clinically presumed or laboratory-confirmed. Consider treatment > 48 h after onset if severe or not improving Clinically presumed or laboratory-confirmed. Consider treatment > 48 h after onset if severe or not improving Health care workers, carers for patients at risk of complications within 48 h of onset of illness Clinically presumed or laboratory-confirmed Clinically presumed (if appropriate exposure history) or laboratory-confirmed Clinically presumed or laboratory-confirmed Clinically presumed or laboratory-confirmed Otherwise healthy adults and children > 5 y within 48 h of onset of illness Clinically presumed or laboratory-confirmed Laboratory-confirmed Clinically presumed (depending on rationing policy and virulence) Not generally indicated Infants < 1 y Depends on clinical scenario Depends on clinical scenario Depends on clinical scenario Depends on clinical scenario Low likelihood of benefit (> 48 h after presentation, known high prevalence of resistance) Not indicated, unless severe infection present. Consider zanamivir if oseltamivir-resistant Not indicated, unless severe infection present. Consider zanamivir if oseltamivir-resistant Not indicated, unless severe infection present. Consider zanamivir if oseltamivir-resistant Not indicated, unless severe infection present Prophylaxis following exposure Groups at risk of complications* Indicated Indicated Indicated (depending on rationing policies) Not generally indicated, except immunosuppressed patients and closed communities Health care workers, carers for patients with comorbidities Indicated Indicated Indicated (depending on policy for national stockpile) Not generally indicated (depending on hospital policy) Healthy adults and children > 5 y within 48 h of exposure Indicated Indicated Not indicated (depending on rationing policy and virulence) Not indicated Children < 1 y Not generally indicated Not generally indicated Not generally indicated Not generally indicated Low likelihood of benefit (> 48 h after exposure) Consider up to 7 days after exposure to prevent transmission Depends on observed incubation period and public health policy Consider early treatment if symptoms develop Not indicated * Such as pregnant women, patients with comorbidities or immunosuppression, and Indigenous Australians (Box 4).

Allen C Cheng FRACP, MPH, PhD · Dominic E Dwyer MD, FRACP, FRCPA · A Thomas C Kotsimbos MB BS, FRACP, MD · Mike Starr MB BS, FRACP · Tony M Korman MB BS, FRACP · Jim P Buttery MB BS, MSc, FRACP · Christine R Jenkins MD, FRACP · Vicki L Krause MD, FAFPHM, DTM · Paul D R Johnson MB BS, PhD, FRACP

To the Editor: We agree with Stockigt that clinical accountability is needed in the production of consumer medicines information (CMI).1 However, providing CMI is a great deal more complex. The ways in which consumers access and absorb medical information are still largely unknown, and the most important issue is potentially that consumers are not receiving CMI. The benefits of involving consumers in shaping health research and medical policy are widely acknowledged.2,3 We recently completed a study funded by the National Health and Medical Research Council (NHMRC) on improving medication safety and chronic disease outcomes in people aged over 65 years, which was accepted for presentation at the national General Practice and Primary Health Care Research Conference in Melbourne, July 2009. Three community forums co-facilitated by the Health Consumers’ Council of Western Australia raised key issues around the safe use of medicines, and an expert panel of consumers aged over 65 years was convened to guide the study. The safe use of medicines was then qualitatively explored in a series of focus groups and individual interviews. Our findings add to those of Vitry and colleagues,4 and suggest that campaigns urging consumers to request CMI may fall short of what is required to make a difference. We found that most study participants did not know they could ask for a CMI printout at the pharmacy. They noted that if the pharmacist was busy, they would not ask for advice or CMI, even if they were uncertain of dosages or potential side effects or interactions. Given that medication side effects and interactions were the greatest concerns for this age group, it is troubling that CMI is not getting to those who need it. In addition, it was noted that participants were not selective in accessing alternative web-based information about medicines, believing it to be reliable because it is “written by medical professionals”. A further issue raised by participants was a perception of undue influence of pharmaceutical companies on the prescribing choices of general practitioners, leading to mistrust of medicines information provided by doctors. Clearly, there is still more to be done to ensure consumers receive the right information at the right time about their medicines.

Caroline E Bulsara · Anne McKenzie

To the Editor: We share Millar’s concerns about the conflicts of interest that influence the genesis and adoption of clinical guidelines1 specifically, and the lack of independent assessment regarding information provided by the pharmaceutical industry generally. Iain Chalmers puts it succinctly: I do not blame industry for trying to get away with anything that is normally considered to be its primary purpose, which is to make profits and look after its shareholders’ interests. It is our profession that has colluded in all of this and been prepared to go along with it — we are the people to blame because we need not have stood for it.2 We believe the reasons behind this acquiescence are complex, but worthy of discussion. A strong and viable pharmaceutical industry is essential for clinical improvement. Similarly, clinical involvement in industry research is necessary. We would not debate either of these statements, but we are concerned about the failure of our profession to stand back and exercise careful scrutiny of data. Classic examples are thalidomide in the 1960s and, more recently, the cyclooxygenase-2 (COX-2) inhibitors, but many less dramatic examples can be found, such as gatifloxacin or rosiglitazone. This failure on our part harms both patients and the standing of our profession. A recent article in this Journal suggested this failure of physician leadership may in part be due to the comfortable position we cultivate with industry,3 relationships that go beyond the business transaction of providing independent medical advice for a consulting fee. Further, the role of “key opinion leaders”, cultivated by industry, is reinforced by criteria for hospital accreditation and university promotion, leading to disproportionate value being placed on service to company boards (which is often paid and of modest time commitment) compared with service on hospital, state and national regulatory and quality committees (which is usually time-consuming and unpaid). The presupposition in this discrepancy is that physicians on the company circuit are better physicians than those who are not. We should all support the recommendations of Millar,1 Olver and Haines,3 and Van Der Weyden,4 including those for true independence and transparency of guideline development and dissemination, strengthening ethical administrative structures and placing appropriate value on public service. Upskilling of clinicians in epidemiology and critical analysis is thus urgently needed so the incremental benefit and costs of new therapies can be objectively examined.

Jennifer H Martin · Christopher Beer · Raymond G Morris · Matthew P Doogue

To the Editor: Articles by Olver and Haines1,2 have catalysed robust discussion about the relationship between the pharmaceutical and device industries and the medical profession.3 These authors advocate changes in the direction of clinical cancer research and in health policy.2 In an era in which research into medicines is dominated by industry, they argue for greater scrutiny of data in a resource-constrained environment, and for fundamental changes in the collection, interpretation and ownership of data. We find their arguments sound and equally applicable to other areas of medicines research and health policy. Their primary recommendation is for “. . . a system to follow up and evaluate the outcomes of all treatments . . .” That is, that we exercise our duty to patients by monitoring and analysing existing clinical data to inform health care policy. There is a great deal of valuable clinical data collected that are not readily accessible because of ownership or privacy issues. For example, much business involving public health dollars is labelled “commercial in confidence”, and laboratory data held in many pathology databases are not accessible at all. The likely benefits to patients and society of transparency and data linkage in health care are greater than possible benefits to individuals of secrecy and privacy. Quality use of medicines (QUM) is one of the central objectives of Australia’s national medicines policy. QUM means selecting management options wisely; choosing suitable medicines if a medicine is considered necessary; and using medicines safely and effectively.4 Olver and Haines also identify issues relating to quality use of research. Quality use of research might include: supporting research into monitoring clinical outcomes related to drug use; supporting research into better use of existing drugs; and supporting truly independent guideline development. There continue to be advances. For example, registration of trials in public databases, such as the Australian New Zealand Clinical Trials Registry, should reduce publication bias.5 However, the decline of independent public sector clinical drug research and the marketing-based design of phase III and, increasingly, phase II industry-funded studies contribute additional bias to the available information. Olver and Haines’ arguments apply to all therapeutics, and particularly to all drug therapies. We strongly support their proposals for health data linkage and for quality use of research. These fit within existing health policy, and our continued failure to make full use of clinical data is an ethically compelling reason for improved political and clinical governance.

Matthew P Doogue · Kathleen M Knights

To the Editor: Although many medicines transfer into breast milk, the amount received by the breastfed infant is usually low, with minimal risk to the infant.1 Because medicines are not tested on breastfeeding women, product information often states that the safety of use during lactation is unknown. This may lead to over-caution in prescribers, who may incorrectly advise women to stop breastfeeding during courses of medication.2 Even brief interruptions to breastfeeding can lead to complications, such as mastitis or breast refusal.1,2 Evidence-based assessments should be made for each mother–baby pair, depending on the baby’s age and the drug’s pharmacokinetics.2 Information about the safety of medicines during breastfeeding is available from books and websites,3 but doctors’ knowledge and use of these resources are not known. We carried out an observational study to describe general practitioners’ current and preferred sources of information about the safety of medicines during breastfeeding. We surveyed the 640 GPs who provided shared maternity care at Victoria’s largest maternity hospital, the Royal Women’s Hospital (RWH), Melbourne. A postal survey to be completed anonymously was sent in November 2007 with a reminder postcard 2 weeks later; a second copy of the survey was sent in February 2008. The study was approved by the human research ethics committees at La Trobe University, University of Melbourne and the RWH. Responses were received from 52% of GPs (335/640); most respondents were women (70%, 233/333), and most had personal experience of breastfeeding for longer than 6 months (68% of participants or their partners, 227/333). Over two-thirds (70%, 233/335) used the Internet during consultations. Eighty-two per cent (270/331) found the Internet helpful. Most participants (73%) obtained information about medicines and breastfeeding from their software program, or from dedicated books (61%; predominantly the RWH’s Drugs and breastfeeding4), and 51% used telephone advice (predominantly from the RWH pharmacy). When asked where they would prefer to access this information, most nominated their software prescribing program (68%) or a reliable Internet database (57%) in their top three preferences (Box). Although most participants (89%, 293/331) felt confident about prescribing for breastfeeding women, the majority were not aware that ibuprofen is considered safe for breastfeeding women; only 31% (102/330) agreed that “there is no problem taking this medicine while breastfeeding”. It appeared that some GPs erroneously believed that pregnancy drug ratings also apply to breastfeeding women. Ibuprofen has a category C pregnancy rating (drugs that have caused or may be suspected of causing harmful effects in the human fetus or neonate without causing malformations), yet the product information from Reckitt Benckiser (Slough, United Kingdom), the manufacturer of Nurofen, states that “no harmful effects are known in breastfed infants”.5 An additional problem is the contradictory advice given by different sources;6 another manufacturer, Abbott, does not recommend ibuprofen for breastfeeding mothers.5 As recommended by researchers in the United States, “We should replace the assumption ‘when in doubt, don’t breast-feed’ with the mandate ‘when in doubt, look it up in a reliable source’”.6 A central accessible source of up-to-date information about individual medications and lactation is urgently needed.7 Most GPs in our study would like this information available on the Internet. Sources of information used by general practitioners when prescribing for breastfeeding women No. of GPs (n = 332) Current sources* Preferred sources† Software prescribing program 242 (73%) 226 (68%) Reliable Internet database 33 (10%) 191 (57%) Dedicated books 203 (61%) 146 (44%) Australian medicines handbook 109 (33%) 125 (38%) Printed guidelines 0 112 (34%) Telephone advice 168 (51%) 106 (32%) Conference/seminars 2 (0.6%) 23 (7%) Journal articles 68 (20%) 19 (6%) One-on-one educational visiting (academic detailing) 0 10 (3%) Printed product information (eg, MIMS) 181 (55%) 3 (0.9%) Therapeutic guidelines 33 (10%) 3 (0.9%) Previous experience 202 (61%) 0 Pharmacist 71 (21%) 0 Colleagues 61 (18%) 0 Other books 3 (0.9%) 0 MIMS = monthly index of medical specialties. * More than one option permitted. † GPs were asked to number their top three preferences.

Lisa H Amir · Marie V Pirotta

A woman with metastatic breast cancer developed diarrhoea, vomiting, and hand and foot discomfort within about 10 days of ceasing therapy with capecitabine and starting docetaxel therapy. The palms of both hands and feet were inflamed and tender, with confluent blanching erythematous areas (Figure). Extensive desquamation occurred from Day 12 to Day 19 after admission, with return to normal skin by Day 30. Hand–foot syndrome has been reported after therapy with various antineoplastic agents, most commonly cytarabine, liposomal doxorubicin, capecitabine, 5-fluorouracil, sorafenib and sunitinib. Increased metabolism of capecitabine in the palms may contribute to the local reaction;1 as may the concentration of docetaxel in eccrine glands in the palms and soles.2 Management involves stopping therapy with the implicated drug, analgesia, and preventing superinfection. Palmar surfaces of the hands showing confluent and erythematous lesions, and the plantar surface of one foot showing a blanching erythematous rash.

Chitra Sivaramamoorthy · Eddy S Thientosapol · Martin H Tattersall

Individuals vary in their response to a medication with regard to efficacy and adverse effects. The human leukocyte antigen (HLA) region of DNA offers the key to predicting drug hypersensitivity reactions. Single nucleotide polymorphisms for hypersensitivity reactions with carbamazepine, abacavir and allopurinol have been identified. A randomised controlled trial demonstrated the effectiveness of prospective screening for the predisposing genetic marker in preventing all cases of the hypersensitivity reaction with abacavir. Further pharmacogenetic investigation of hypersensitivity reactions could be conducted in Australia by establishing a network of sentinel hospitals.

Kathlyn J Ronaldson BSc, MSc, DPhil · John J McNeil PhD, FRACP, FAFPHM

To the Editor: We read with interest Talley’s excellent and informative editorial about the risks associated with proton-pump inhibitors (PPIs).1 Other possible serious side effects of PPIs that need to be taken into account are potential drug interactions with aspirin and clopidogrel. Aspirin is a weak acid that crosses the mucosa in its lipid state. The suppression of acid production reduces the lipophilic nature of this drug and, theoretically, might reduce its absorption and bioavailability.2 On the other hand, clopidogrel is a prodrug that is converted in the liver to an active metabolite. This bioactivation is mediated by hepatic cytochrome P450 isoenzymes,3 with cytochrome P450 2C19 (CYP2C19) playing a particularly important role. There is evidence suggesting that some PPIs (omeprazole, lansoprazole and rabeprazole) can inhibit CYP2C19, which would alter the effectiveness of clopidogrel and potentially lead to an increased risk of adverse cardiovascular outcomes. In a recent Canadian case–control study among patients prescribed clopidogrel after acute myocardial infarction, current use of PPIs was associated with an increased risk of reinfarction (adjusted odds ratio, 1.27; 95% CI, 1.03–1.57).4 The risk was limited to patients currently taking a PPI (the authors did not find any association with more distant exposure to PPIs), and did not extend to pantoprazole, a drug that does not interfere with the conversion of clopidogrel to its active form.

Francisco J Fernández-Fernández · Gonzalo Pía · Pascual Sesma

To the Editor: In his recent editorial, Talley discusses a range of risks of proton-pump inhibitors (PPIs).1 Another rare but serious side effect of PPIs of which every doctor should be aware is hyponatraemia. Eleven cases of hyponatraemia caused by PPIs have been published.2,3 Consistent features were the rapid onset of hyponatraemia within days of commencement of the PPI therapy, the severity of hyponatraemia often being associated with confusion or delirium, and rapid recovery after cessation of the PPI medication. Test results in each case were consistent with inappropriate release of antidiuretic hormone. One case occurred 5 days after a patient changed from lansoprazole to esomeprazole.4 Hyoponatraemia needs to be considered whenever there is clinical deterioration, even after brief exposure to a PPI.

Adam P Morton

In reply: Proton-pump inhibitors (PPIs) are often coprescribed for patients taking aspirin and clopidogrel to reduce gastrointestinal bleeding. There are emerging data that omeprazole diminishes the therapeutic effect of clopidogrel because the active enzyme in the liver, cytochrome P450 2C19 (CYP2C19), metabolises omeprazole and activates clopidogrel.1,2 In a large cohort study of 8205 patients with acute coronary syndrome and taking clopidogrel, 64% were also taking a PPI (60% omeprazole); 21% of those who were taking clopidogrel but no PPI died or were rehospitalised for acute coronary syndrome, versus 30% of those taking clopidogrel as well as a PPI.3 Notably, not all the PPIs have the same metabolic pathway. For example, omeprazole and esomeprazole are principally metabolised by CYP2C19 in contrast to lansoprazole, which is metabolised by cytochrome P450 3A4 (CYP3A4), and pantoprazole, which is metabolised by CYP2C19 O-demethylation then rapid sulfate conjugation. Thus, the negative interaction with clopidogrel may not apply to all PPIs (and pantoprazole may be the drug of choice if a PPI is required, as cytochrome P450 interactions are least likely).1 However, until more data are accumulated, all PPIs should probably be avoided where possible in patients who have been prescribed clopidogrel, unless there is no alternative. It is correct that hyponatraemia has, rarely, been reported in patients taking PPIs. However, this knowledge is based solely on case report data, and therefore the level of evidence for cause and effect is relatively weak.

Nicholas J Talley · Aneta Dimoska · Kevin Gan

To the Editor: We report a case of proximal myopathy attributed to varenicline therapy to assist with smoking cessation. A previously fit and well 27-year-old man presented to the emergency department with a 1-week history of lethargy, myalgia and limb weakness. The pain and progressive weakness incapacitated the patient to the extent of him requiring assistance to move from bed to chair. The patient’s only medication on admission was varenicline (Champix [Pfizer]), which he had started taking about 6 weeks previously to assist with smoking cessation. He denied drinking excessive alcohol or using illicit drugs. There were no symptoms of recent infection. On examination, a proximal myopathy with weakness of grade 2/5 for both upper and lower limbs was noted. The patient’s muscles were mildly tender, reflexes were preserved, and both cranial nerves and sensory examinations were normal. Respiratory function was not impaired and there was no evidence of fatigability. The clinical impression was of proximal myopathy of uncertain cause. An adverse drug reaction to varenicline was considered and the medication ceased. Appropriate blood investigations and clinical neurophysiological studies were requested. A full blood count, thyroid function tests and autoimmune screen were all normal, as were levels of electrolytes, calcium, phosphate and magnesium. The erythrocyte sedimentation rate was 15 mm/h (reference range [RR], 1–15 mm/h), the C-reactive protein level was 20 mg/L (RR, < 5 mg/L), and the creatine kinase level was 1100 U/L (RR, 30–190 U/L). Cushing disease was excluded. Two days after cessation of varenicline, the muscle weakness had improved and the creatine kinase level had normalised. The patient declined neurophysiological studies and was discharged home. On review 1 week later, he had made a full recovery. Varenicline is a recently marketed smoking cessation drug treatment that has been shown to be more effective than placebo and bupropion treatment.1 Varenicline acts as a partial agonist at nicotinic acetylcholine receptors in the brain.2 The agonist activity at these receptor sites reduces the symptoms of nicotine withdrawal and craving, while the antagonist activity blocks the reinforcing and rewarding properties of nicotine binding. Recognised adverse effects include nausea, headache, insomnia and abnormal dreams. Musculoskeletal effects are uncommon and limited to joint stiffness and muscle spasms.2 Potential mechanisms for myopathy include muscle cell degeneration induced by excess acetylcholine activity at the neuromuscular junction3 or possibly by varenicline’s affinity for the serotonin receptor.4 Enquiries to the Australian Adverse Drug Reactions Advisory Committee (ADRAC) and searches of the Canadian adverse drug reaction database and PubMed database failed to identify any reports of myopathy associated with varenicline. Thus we believe this to be the first reported case of proximal myopathy due to varenicline, and have reported the case to ADRAC accordingly.

Shelley E Wood · P Gerry Fegan

To the Editor: We are pleased that Simpson and colleagues brought the important matter of drug-dose calculation skills among Australian hospital doctors to the attention of the wider medical community.1 Data similar to theirs, suggesting inadequate calculation skills, have been reported from the United Kingdom and Germany.2-4 As teachers in the MB BS course of the University of Adelaide, we have been concerned with deficiencies in the clinical numeracy skills of our students for some time. Such deficiencies among medical students have been reported from North Carolina,5 but to our knowledge no information has been available about Australian medical students. In 2008, we included three questions on drug-dose calculations in the 90-item multiple choice question (MCQ) section of the final examination for Years 1, 2 and 3 of our course. For each question (Box), students were required to select one correct answer from five options. Although Question 1 required the knowledge that one standard drink contains 10 g of ethanol, as well as calculation skills, Questions 2 and 3 solely examined numeracy skills. The exam was completed by 177 students in Year 1, 155 in Year 2, and 119 in Year 3. The distribution of their responses is shown in the Box (with correct responses in bold). The percentage of correct responses to these questions was significantly lower than the overall score for the MCQ paper, with the exception of Question 3, which may have been too easy (as it did not discriminate between students). We believe these data support our hitherto anecdotal concerns that many students in the MB BS program have inadequate calculation skills. Although tertiary students’ numeracy problems have been attributed, at least in part, to the level of mathematics teaching in secondary schools,6 we consider they must be addressed at university level. We plan to make available an online calculation learning tool that begins with real-life non-medical examples. We are currently developing this tool with the University of Adelaide’s Mathematics Learning Centre and plan to publish our experience, including evaluation, with a view to making the tool widely available. Distribution of medical student responses to examination questions requiring numeracy skills* Question 1 (Year 1 and Year 3) A 21-year-old woman recalls drinking 5 glasses of champagne at her birthday party. Her glass holds 200 mL and the champagne has an alcohol content of 12.5%. How many standard drinks did she consume during her party? % of respondents Options Year 1 (n = 177) Year 3 (n = 119) A. 12.5 standard drinks* 50% 49% B. 5 standard drinks 6% 7% C. 18 standard drinks 1% 2% D. 6.25 standard drinks 20% 15% E. 10 standard drinks 23% 27% Question 2 (Year 2 and Year 3) You are treating a 60 kg patient for a laceration, which you will need to suture under local anaesthetic. Given that the maximum safe dose of lignocaine is 3 mg/kg, what is the maximum volume of a lignocaine 1% weight per volume (w/v) solution that can be administered safely? Options % of respondents Year 2 (n = 155) A. 1.8 mL 46% B. 6 mL 2% C. 18 mL* 35% D. 20 mL 14% E. 60 mL 1% Year 3 (n = 119) A. 60 mL 1% B. 6 mL 0 C. 180 mL 41% D. 18 mL* 38% E. 180 μL 19% Question 3 (Year 3) A 10 kg infant has viral meningitis and a high temperature. You want to treat her fever symptomatically with oral paracetamol. The preparation is paracetamol 50 mg/mL and the recommended dose is 15 mg/kg. How many mL of paracetamol syrup is the equivalent of one dose? Options % of respondents (n = 119) A. 1.5 mL 0 B. 3 mL* 96% C. 6 mL 1% D. 12 mL 2% E. 15 mL 1% * The correct options are shown in bold.

Kingsley J Whittenbury · Hubertus P Jersmann · Anne L Tonkin

Objective: To investigate sociodemographic variation in antidepressant utilisation.Design and setting: Cross-sectional analysis of antidepressant prescription under the Pharmaceutical Benefits Scheme in Australia, 2003–2005.Main outcome measures: Antidepressant utilisation (defined daily dose/1000/day) by sex, age, socioeconomic status (SES) and geographichal area.Results: Total antidepressant utilisation increased with age. Among those aged ≥ 15 years, female utilisation was about double that of males. About half of antidepressant utilisation was accounted for by sertraline, venlafaxine, citalopram, and paroxetine. SES differentials in antidepressant utilisation changed across age groups for males and females: among those aged ≤ 19 years, total antidepressant utilisation was significantly less in lower SES groups (P < 0.001); there was no relationship to SES among 20–29-year-olds; and among those aged ≥ 30 years, antidepressant utilisation was significantly higher in lower SES groups (P < 0.001). SES differences were attenuated after adjusting for urban or rural residence, but remained statistically significant. Antidepressant utilisation rates were highest in regional centres.Conclusion: Antidepressant utilisation in Australia partially reflects sociodemographic differences in the prevalence of affective disorder. Discrepancies between treatment provision and treatment need suggest that not all social strata in Australia have equal access to these treatments.

Andrew N Page BA(Psych)(Hons), PhD · Sarah Swannell BPsych(Hons), GradCertBiostat · Graham Martin MD, FRANZCP, DPM · Samantha Hollingworth BSc, MPH, PhD · Ian B Hickie MB BS, FRANZCP, MD · Wayne D Hall BSc, PhD

A key issue is to recognise when a duality of interest has become a conflict of interest In Australia, provision of specialised product information and promotion by the pharmaceutical industry of drugs approved by the Therapeutic Goods Administration is an integral part of the health care environment. The pharmaceutical industry provides information and training to health professionals about new products; funding for conferences; support for professional and social activities secondary to medical education; support for the conduct of research and information about its outcomes; and opportunities to meet with peers. However, the primary goals of the pharmaceutical industry and health professionals differ: the pharmaceutical industry has a financial responsibility to shareholders, while health professionals have a moral responsibility to their patients. The challenge for both is to manage their responsibilities when interacting with one another. The pharmaceutical industry’s code of conduct1 upholds the principles of Australia’s Quality Use of Medicines program and National Medicines Policy. However, an interaction between pharmaceutical representatives and hospital employees will ultimately have a promotional intent. In itself, an indirect promotional activity is not a problem. However, the interaction will often influence prescribing.2 Many health professionals deny that such activity influences their behaviour, although, paradoxically, they believe their peers may be more easily swayed.3 Appropriate provision of patient care requires health professionals to understand these influences and keep them in mind in order to maintain independence of judgement. Ethics relating to promotional activities of pharmaceutical companies and managing conflicts of interest have been recently reviewed.4-10 Some researchers have argued that contact with the pharmaceutical industry should be more restricted and certain activities prohibited. In the United States, steps have been taken to prohibit all gifts (including meals) and to institute central management of product samples.8 A US report commented that “bias, either by appearance or reality, has been woven into the very fabric of continuing education” and called for cessation of commercial support from pharmaceutical and medical device companies.11 In Australia, while the move to state and federal funding and other non-commercial sources for educational and drug information activities is currently being debated, it is unrealistic to prohibit contact between health professionals and the pharmaceutical industry. It may be argued that industry plays an important role in health education — indeed, constructive engagement between industry and health professionals may be in the interests of patients. Severing all contact between industry and health care providers could limit open dialogue, hamper innovation and create a huge gap in educational support for health professionals. Initiatives to bridge the gap have been suggested.4-6 In the meantime, hospital staff must analyse the nature of their current interactions with the pharmaceutical industry and aim to improve it to optimise benefit to the patient. Codes of practice have been developed by professional bodies, societies, hospitals, government and the pharmaceutical industry in an attempt to ensure that interactions between hospital-based health professionals and the pharmaceutical industry are ethical and in the interests of the patient. However, a more practical framework is required to evaluate these interactions and to work towards achieving the highest standards of patient care and quality use of medicines. At the request of its members, the New South Wales Therapeutic Advisory Group (NSW TAG) recently updated its existing position statement on liaison between hospital staff in NSW and the pharmaceutical industry. The position statement provides evidence-based guidelines to help hospital staff recognise the activities that enhance clinical practice and those that potentially damage the relationship between health professionals and patients.12 It suggests steps to minimise potential conflicts of interest and ways to support ethical interaction, including making full use of independent sources of evidence-based medicine. It proposes that all health professionals adopt the approach of the Royal Australasian College of Physicians with regard to identifying and managing dualities and conflicts of interest.13 A duality of interest (where two or more interests coexist) is not unethical, but the key issue is to recognise when one interest is compromising the other (ie, when a conflict of interest is present). It is not enough to voluntarily disclose a duality of interest and then feel justified in proceeding regardless. Members of NSW TAG have discussed establishing a system of review and authorisation, deciding whether steps are necessary to separate or prohibit the conflicting activities and how open communication contributes to the transparency of the process. Our intention has been to ensure that the primary objective of professional interactions with pharmaceutical companies is to advance the health and wellbeing of patients. A recent article called for a set of guidelines for academic medical centres and opinion leaders.4 Extension of practical guidelines to all health professionals is a necessary next step. The pharmaceutical industry has established a system of self-regulation.1 In authorising the Medicines Australia code of conduct, currently under review, the Australian Competition and Consumer Commission requires details to be published of educational events provided or sponsored by member companies. All events have been reviewed by an independent auditor, and the first of these 6-monthly reports is now available.14 The audit had limitations with regard to investigation of high-cost activities and verification of data supplied. Nevertheless, such measures from industry to increase transparency support the intentions of NSW TAG’s position statement.12 The issues discussed in the position statement extend well beyond the pharmaceutical industry. They also include providers of medical devices, chemicals in pathology laboratories, and machines and consumables in radiology departments. Understanding the differences between the role of the health professional and that of the pharmaceutical industry is fundamental to understanding how to handle the interaction between the two groups. This process is evolving and the NSW TAG position statement is considered a “work in progress” to provide guidance within existing codes. The pharmaceutical industry and health professionals need to continue to foster a process of introspective challenge and regulation. Ongoing discussion by all stakeholders to find solutions that benefit patients is paramount.

Diana H Shipp BPharm, MRPharmS · Gordon Mallarkey BSc(Hons), PhD

Clinical guidelines are recommendations based on systematic identification and synthesis of the best available scientific evidence. The National Health and Medical Research Council (NHMRC) has published standards for guideline development. According to the NHMRC standards, guideline development must be a transparent and independent process, with full disclosure of any potential competing interests. Australian guidelines for prevention of venous thromboembolism have been published by an autonomous group. Several features of the processes used to produce and distribute these guidelines, such as pharmaceutical sponsorship, do not meet NHMRC endorsement standards. The guidelines may overstate the need for thrombo-prophylaxis in medical patients, and thus expose some patients to an unnecessary risk of bleeding complications. Despite this, these guidelines have been taken up avidly by national and state bodies responsible for safety and quality in health care, and mandated national application has been proposed.

J Alasdair Millar PhD, FRACP, FRCP

To the Editor: Isoniazid is used extensively for the treatment of active and latent tuberculosis (TB). It is generally well tolerated by children, and hypersensitivity reactions resulting in skin rash and requiring cessation of treatment are rarely reported in this age group.1,2 We report a case of isoniazid hypersensitivity in a 21-month-old boy potentially exposed to TB in a childcare setting. He was one of over 80 children screened after contact with a childcare worker who showed a positive smear result. His initial tuberculin skin test (TST) was negative and, in line with New South Wales guidelines,3 he was commenced on isoniazid 150 mg daily (10mg/kg/day) while awaiting a repeat TST. After 3 days of treatment, he developed a small number of round vesicular lesions on his tongue. They were associated with mild discomfort but his appetite was not affected. There were three small maculopapular lesions on his legs and back that reportedly looked like mosquito bites before blistering. The child remained afebrile and was systemically well. The family general practitioner considered that this presentation was possibly an allergic reaction and isoniazid was discontinued. Population health staff were consulted, and the risks and benefits of further isoniazid treatment were discussed with paediatric TB specialists. It was recommended that, after the rash had resolved, isoniazid be reintroduced at half the dosage and with close supervision. Two days after isoniazid 75 mg daily was recommenced, the rash recurred. The child’s mother described lesions appearing as “burns all over his tongue” and reported further sores around his lips and six welt-like lesions on his legs. Isoniazid was immediately discontinued, the skin lesions resolved within 5 days and no further antituberculous therapy was administered. His repeat TST 12 weeks after the initial test was negative and he remains well. No other potential triggers for a hypersensitivity reaction were identified. In particular, no other medications were administered during this period or for the week before commencing isoniazid. According to his mother, the child had experienced a similar reaction within 1 hour of a single dose of ibuprofen when he was 8 months old. Several tongue blisters were accompanied by a generalised fine maculopapular rash lasting several days. We concluded that the child most likely had a hypersensitivity reaction to isoniazid that required discontinuation of treatment. We reported this to the Therapeutic Goods Administration, which advised that it had received seven other reports since 1991 of suspected hypersensitivity, but none were for children under 10 years of age.

Tony D Merritt · Peter D Massey

Therapeutic guidelines. Toxicology and wilderness. Emergency Medicine Expert Group. Melbourne: Therapeutic Guidelines, 2008 (xxii + 311 pp). ISBN 978 0 9804764 0 8. Poisoning is a common presentation to Australian emergency departments and a common cause of death in those under 40, yet there is a real paucity of Australasian toxicology texts. Toxicology and wilderness, a new release in the Therapeutic Guidelines series, is a subset of topics prepared by the Emergency Medicine expert writing group for the electronic eTG complete. Toxicology and wilderness sensibly starts with in-depth information on the many aspects of resuscitation. There is a good overview on the approach to the poisoned patient, with a great nomogram to help assess the risk of torsades from a prolonged QT. The majority of the book is based on the toxicology of individual agents. It approaches each agent in a structured manner, detailing the indicators for toxicity, clinical presentation, key investigations and treatment. It has much helpful and sensible advice. The authors seem to advocate routine activated charcoal for most poisonings that present within 1 hour, which I would disagree with. The discussion on antidotes is understandably brief, although I was curious to read in detail about dicobalt edetate for the treatment of cyanide poisoning, and not the currently recommended and far safer hydroxocobalamin. I felt that there were sections where better emphasis on the potential for severity of the poisoning or management issues could have occurred. The book ends with a well written section on envenoming, then the unusual bedfellow of wilderness medicine. I think the strength of this book lies in the ready access of the electronic format for hospital practitioners. I found the information a good starting point but, due to the restrictions of the structure of this series, a little light in some areas. As to whether it sits on your bookshelf — you need to browse through to see if it fits a need.

Mark Little

We propose that various measures of brain structure or function, gene expression and proteomic technologies can be used to guide better treatment of schizophrenia and other psychotic disorders. These measures are not used to establish a specific diagnosis. Their purpose is to predict variations in underlying illness activity that predict severity, course of clinical illness, or other morbidity. We propose a new instrument that uses a composite scoring system of systemic biomarkers of illness-related changes in health status: the Brain and Mind Research Institute Biomarker Index. This may permit comparison of biological dysfunction among patients who are at similar points in their illness or have similar clinical features. A specific example of the use of a novel positron emission tomography marker of progressive brain disease in patients with schizophrenia is described.

Richard Banati MD, PhD · Ian B Hickie MD, FRANZCP, AM

A 55-year-old woman who was treated with long-term, high-dose clozapine for schizophrenia presented with bilateral decreased visual acuity. She had pigmentary changes affecting the cornea and the retina, as well as stellate cataract. Chlorpromazine use is known to produce similar changes, but this is the first report to our knowledge of pigmentation associated with clozapine use. (MJA 2009; 190: 210-211) Clinical recordA 55-year-old white woman presented to a tertiary hospital eye clinic with a progressive decline in vision that affected the left eye more than the right. She had a history of schizophrenia, depression, hypothyroidism, gastro-oesophageal reflux and back pain. Her medications were: clozapine 800 mg daily, lithium carbonate 500 mg daily, thyroxine 100 μg daily, and omeprazole 20 mg daily. She had been taking clozapine for 16 years, and her cumulative dose was 4.67 kg. The patient’s best corrected visual acuity was 6/9 in the right eye and 6/60 in the left. Bilateral pigmented deposits were present in the corneal endothelium, and these were most prevalent in the interpalpebral fissure. On dilation, bilateral pigment dusting of the anterior portion of the lens capsule with central stellate opacity was evident (Box, A), and posterior subcapsular and nuclear sclerotic cataract was noted in both eyes. Retinal changes included a right epiretinal membrane and bilateral pigmentary retinopathy. Macular atrophy was present in both eyes and affected fixation on the left. The patient’s skin was brown, particularly in sun-exposed areas including her face, neck and hands. Confocal microscopy of the corneas showed diffuse, highly reflective, irregular honeycomb-shaped deposits on the endothelium (Box, B) and small granular deposits on the posterior stromal layer. Morphology of the endothelium visible between the deposits was normal. Optical coherence tomography confirmed atrophy of the neuroretina, greater in the left eye than the right. Electroretinography showed reduced cone function, indicated by reduced amplitude and latency in the cone response. The patient was diagnosed with presumed clozapine-related ocular and skin pigmentation. In consultation with her psychiatrist, her clozapine dose was reduced to 600 mg daily. On follow-up at 6 months, her vision had not improved, and the deposits had neither reduced nor progressed. DiscussionPigment deposits in the cornea, lens and skin are well documented complications of long-term phenothiazine antipsychotic therapy.1-3 Our patient had ocular changes that were possibly side effects of chronic high-dose clozapine use. The changes were similar to the side effects of phenothiazines, and they were demonstrated by confocal microscopy, optical coherence tomography and electroretinography.1-4 Medication history taken from the patient, as well as a collaborative medication history supplied by the patient’s psychiatric team, revealed no evidence of prior phenothiazine use. Clozapine is a tricyclic dibenzodiazepine derivative with weak D2 and D1 dopamine-receptor blocking activity. It is a relatively new atypical antipsychotic that is used in place of phenothiazines, particularly for refractory schizophrenia. It has noradrenolytic, anticholinergic, antihistaminic and antiserotonergic properties, and its most common side effects are sedation and weight gain. Anticholinergic side effects such as constipation and dry mouth may also occur. Rarely, clozapine can produce potentially lethal agranulocytosis, for which blood count monitoring is required. Myocarditis is another possible lethal side effect. Clozapine therapy is usually commenced at a dose of 25 mg daily, and titrated up to 300–600 mg daily for therapeutic effect. Doses of up to 900 mg can be used for treatment-resistant cases.5 Clozapine is recommended as a substitute for patients who have experienced pigmentation secondary to chlorpromazine use — clinical signs of pigmentation are expected to resolve after a period of chlorpromazine abstinence.6 The aetiology of phenothiazine-related ocular side effects has not been determined. It has been postulated that photosensitisation of tissue proteins occurs in areas with increased sun exposure after accumulation of the drug in these tissues.7 Alternatively, phenothiazines may interact with melanin in the choriocapillaris and retinal pigment in the epithelium, which may induce damage to the photoreceptors. Altered dopaminergic regulation of melatonin is suspected to increase susceptibility of photoreceptors to damage by light.7 In our patient, clozapine may have produced similar side effects to the phenothiazines, as it also acts on dopamine receptors. The dopaminergic system of the retina may respond to accumulation of clozapine and phenothiazines in a similar manner to the nigrostriatal dopaminergic system.8 None of the other medications the patient was taking — namely lithium, omeprazole or thyroxine — are known to cause skin or ocular pigmentation.9,10 Our patient had significant irreversible loss of vision, which may have resulted from chronic, high-dose clozapine use. Anterior and posterior segments of the eye were affected. These changes should be considered as possible side effects of clozapine, particularly if it is given in high doses. If further similar cases become evident, patients on long-term clozapine therapy should be considered for regular ophthalmological review. Phenothiazine-like ocular changes in a patient who was treated with long-term, high-dose clozapine A: Pigmented deposits on the corneal endothelium (arrowhead) and anterior central stellate cataract (arrow). B: Confocal microscopy image of retina showing highly reflective honeycomb-shaped deposits on the endothelial layer (arrow) with an affinity to the cell margin rather than the centre, and endothelial cells with a regular morphology (arrowhead) (original magnification, × 40).

Armand M Borovik MB BS, BSurg · Martina M Bosch MD, FMHOphth · Stephanie L Watson MB BS, FRANZCO, PhD

Scattered: the inside story of ice in Australia. Malcolm Knox. Sydney: Allen & Unwin, 2008 (xi + 290 pp). ISBN 978 1 74175 358 5. Malcolm Knox is a Walkley Award-winning journalist and he has a keen grasp of both the pharmacology and psychiatric effects of crystal methamphetamine. Scattered provides a lucid, elegant description of the turbulent recent history of ice in Australia. Knox goes beyond the statistics and includes a series of case vignettes, exploring the human cost of this drug in the Australian context. Knox explains that the heroin drought in Australia since 2000 has led to an alternative, yet far more damaging, drug supplanting the somnolent effects of the opioids. Ice is relatively cheap, hitting the streets in all our capital cities (and everywhere a truck goes), and provides an instant, profound teeth-grinding hit, far more potent than any other methamphetamine in history. This book is a roller coaster read in three sections: going up, the high, and coming down. In many ways, the book’s structure follows the effects of this most potent psychostimulant on many unfortunate users. Many patients that I have seen in my addiction practice have suffered severe drug-induced ice psychoses requiring urgent hospitalisation and larger than usual doses of tranquilliser medication. Such patients are alarming to both doctors and nursing staff alike. Practitioners at the coalface can only hope that ice becomes less popular and that we are spared the prospect of such violent psychosis or a terrified patient hiding under a desk, paranoid about imaginary helicopters spying on his every move. This book will not be to everybody’s taste because of the graphic descriptions of ice-fuelled violence and sexual depravity, but it should be required reading for all doctors who encounter such patients on the edge of oblivion.

Raymond C Seidler

The current review of the Therapeutic Goods Administration is an opportunity to improve the system for updating product and consumer information on drugs Pharmaceutical product information (PI) and consumer medicines information (CMI) are mandatory for prescription products in Australia, and government regulations specify that CMI must be consistent with PI.1 Health professionals and consumers should be able to assume that these sources are up-to-date and consistent with evidence-based best practice. However, this is not necessarily so, particularly for older medications.2,3 There is a wide discrepancy between the high-quality information available for new medications (eg, through series such as NPS RADAR [National Prescribing Service Rational Assessment of Drugs and Research]) and some existing texts2-4 that originate from pharmaceutical sponsors, who pay fees to the Therapeutic Goods Administration (TGA) for review and approval of their submitted material. Officially sanctioned information may appear different from different perspectives: all may seem to be in order when assessed from the top down, and shortcomings may become apparent only when specific end products or outcomes are evaluated. Two examples demonstrate this problem. Current CMI for glucocorticoids fails to distinguish between the dosages for replacement and for anti-inflammatory and immunosuppressive effects, a potential health hazard for several thousand Australians with adrenal insufficiency.3 The CMI in question, presented without professional accountability, remains uncorrected 18 months after attention was drawn to it,4 and is clearly inconsistent with the corresponding PI and advice in the Australian medicines handbook.5 In a second example, review of the PI from four different sponsors for thyroid medications identified erroneous therapeutic recommendations and the omission of well established indications or important side effects, as well as inappropriate advice on dose adjustment.2 Two years after publication of a detailed critique of the PI for these medications,2 11 of 16 salient errors remain uncorrected.6 When medical professionals point out necessary improvements to current PI or CMI, official responses tend to be self-affirming, legalistic and defensive, rather than receptive to evidence and the consensus of clinical expertise. For example, when it was pointed out that the instruction in CMI, “Do not take Cortate if you have an uncontrolled infection”,5 was dangerous for those with adrenal insufficiency, the TGA responded with the unexpected sophistry that this advice meant only “before you commence taking Cortate”, rather than “before you take your continuing medication”.7 A response in the general press from NPS leadership denied any need to differentiate glucocorticoid replacement from other indications.8 What are the systemic weaknesses of Australia’s system of preparing, reviewing and updating PI and CMI? Australian pharmaceutical sponsors may lack the clinical resources and perspective to offer PI that reflects evidence-based best practice. The major publisher of PI and CMI, MIMS Australia, is restricted to publishing the TGA-approved texts.9 PI and CMI are currently presented without professional accountability, a prerequisite for effective review. How can these difficulties be addressed? Some recommend a defined “use by” date for PI.10 However, regular review would not necessarily address clinical concerns, and the cost might be prohibitive. What else can be done? Regulatory authorities must abandon the now familiar response to any critique, “PI is the responsibility of the drug sponsor”, which can be used by these authorities to deny responsibility for deficiencies in that information. Sponsors need stronger clinical support, whether through the TGA or other means, in presenting therapeutic advice. Those who publish and disseminate PI and CMI, such as MIMS Australia, should be able to review, and should be accountable for, those texts. MIMS names a distinguished senior honorary editorial panel,6 who could have a valuable role in endorsing published PI or suggesting necessary revisions. Abundant clinical expertise is available in Australia, often concentrated and coordinated in the clinical and scientific specialty societies, that could be brought together under the auspices of the Royal Australasian College of Physicians. Consensus advice from a specialty society, rather than individuals, would diminish the potential influence of commercial interests or pressure groups. The key to effective updating and improvement of Australian pharmaceutical information is more fluent incorporation of clinical input, as occurs for adverse drug events. A notification process, initiated by vigilant professionals and consumers, should make it possible to eliminate incorrect, misleading, ambiguous or obsolete PI and CMI. The alternative is a progressively widening gap between industry and the consumers and professionals who use or prescribe medications. Both groups have the right to expect reliable, officially sanctioned pharmaceutical information. A revision of structures within the TGA is currently in progress.11 It would be a further setback if this opportunity to incorporate expert professional advice in the preparation and improvement of PI and CMI were overlooked.

Jim R Stockigt MD, FRACP, FRCPA

Objective: To determine whether proton-pump inhibitor (PPI) use is associated with hospitalisations for pneumonia and with antibiotic use.Design and setting: Historical cohort study in the Australian veteran population, conducted from 1 January 2002 to 30 December 2006, comparing veterans exposed to PPIs with those not exposed.Participants: All 185 533 veterans who were Gold Card holders (ie, eligible for all health services subsidised by the Department of Veterans’ Affairs) and aged 65 years and over at 1 January 2002 and had been prescribed at least one medicine in the previous 6 months.Main outcome measures: The primary endpoint was hospitalisation for pneumonia. Secondary endpoints included hospitalisation for bacterial pneumonia and dispensings of antibiotics commonly used to treat respiratory tract infections.Results: After adjustment for potential confounders, we found an increased risk of hospitalisation for pneumonia among those exposed to PPIs compared with the unexposed group (rate ratio [RR], 1.16; 95% CI, 1.11–1.22). The risk was not increased for bacterial pneumonia (RR, 1.13; 95% CI, 0.98–1.31), which made up 8% of pneumonia cases. An increased risk of antibiotic dispensings was observed among those exposed to PPIs (RR, 1.23; 95% CI, 1.21–1.24).Conclusions: PPI dispensings were found to be associated with a small but significant increased risk of hospitalisation for pneumonia. While the increased risk is small, the prevalent use of PPIs means that many people could be affected.

Elizabeth E Roughead BPharm, MAppSci, PhD · Emmae N Ramsay BSc, GradDipApplStats · Nicole L Pratt BSc(Hons) · Philip Ryan MB BS · Andrew L Gilbert PhD

Objective: To assess the ability of doctors to calculate drug doses and their workplace prescribing and calculation habits.Design and setting: Prospective, questionnaire-based observational study conducted at a 570-bed teaching hospital in February 2007.Participants: Convenience sample of 190 doctors, representing all acute medical and surgical disciplines and diverse levels of experience.Main outcome measures: Demographic data, self-reported prescribing habits, predicted score on a 12-item test of ability to calculate drug doses, score considered adequate for peers, and actual score.Results: 141 doctors (74%) completed the questionnaire. The mean actual score on the test was 72.5% (95% CI, 67.8%–77.3%), which was similar to the group’s mean predicted score (74.7%; 95% CI, 71.0%–78.5%) but significantly lower than the mean of the score they considered adequate (91.6%; 95% CI, 89.5%–93.8%) (P < 0.001). Subgroup analyses showed that senior doctors and those in critical care specialties (intensive care, emergency medicine and anaesthesia) achieved significantly higher actual scores than junior doctors and those in non-critical care specialties, respectively.Conclusions: Doctors expect their colleagues to perform significantly better in a drug-dose calculation test than they expect to, or can achieve, themselves. Junior staff and those in non-critical care specialties should be targeted for education in the skill of drug-dose calculation to reduce the risk of medication error and its consequences.

Chanelle M Simpson MB BS(Hons), FACEM · Gerben B Keijzers MB BS, MSc(Epidemiology), FACEM · James F Lind MB BS, FACEM, MRCP

To the Editor: We question the conclusions drawn by Canfell and colleagues1 from their analysis of trends in hormone replacement therapy (HRT) prevalence and breast cancer incidence for Australian women aged 50 years or older. Their ecological analysis lacks individual-level information on HRT use and information on tumour oestrogen receptor (ER) status, and captures only 2 years following the decline in HRT prevalence. This is an inadequate design within which to judge issues of causality; it is at best an hypothesis-generating exercise.2 The examination of only 2 years of breast cancer incidence after the decline in HRT use is unsound because of substantial unexplained annual variability in national breast cancer incidence. This is evidenced, for example, by the graph in Canfell et al’s Box 2, which identifies a fall in 1998–1999 that was not ascribed to changes in HRT prevalence. The lack of data on tumour ER status is another weakness. HRT use increases the risk of ER+ tumours, so any decline in HRT prevalence would be expected specifically to reduce ER+ tumour incidence. We analysed Victorian Cancer Registry data for women aged 50 years or older for the period 2001–2005 — 2 years past the cut-off in Canfell et al’s analysis. Tumour ER status was available for 87% of breast cancer cases in 2001, rising to 90% in 2005. The demographic characteristics of the women for whom these data were available did not change between 2001 and 2005. Over this period, the proportion of all breast cancers that were ER+ increased from 65% to 71%, and the proportion of tumours with known ER status that were ER+ increased from 74% to 79%. Poisson regression analysis of the age-specific incidence rates for both total and ER+ breast cancer for women aged 50 years or older estimated an average annual decline of 1.7% in the total incidence rate (P = 0.0009) and an annual increase of 0.2% in the ER+ incidence rate (P = 0.7). Our findings are illustrated in the Box. Although there was an apparent small decline in total breast cancer incidence in 2001–2003, this trend was reversed in 2004–2005. More importantly, the trend in ER+ tumour incidence was stable across the entire period. Age-standardised incidence* of invasive breast cancer in women in Victoria aged ≥ 50 years, 2000–2005 Vertical bars represent 95% confidence intervals. ER+ = oestrogen receptor-positive. * Standardised to World Standard Population. These Victorian trends, covering a longer time period and including information on the tumour type most likely to be affected by changes in HRT use, provide no support for the hypothesis of Canfell and colleagues.

Graham G Giles · Richard Bell · Helen Farrugia · Vicky Thursfield