Infectious diseases

To the Editor: Community strains of methicillin-resistant Staphylococcus aureus (MRSA) are increasingly seen in Australia, particularly in certain population subgroups, such as Pacific Islander1 and Aboriginal2 people. We report the case of an African with community MRSA to highlight its existence in yet another subgroup. Given the increase in people arriving in Australia from Africa under the Humanitarian Program (with around 8500 arrivals from Africa in 2004–2005)3 and their wide dispersal around the country, it is possible that African community MRSA will be seen increasingly in Australia. A 53-year-old Burundi refugee presented with an infected wound overlying the left lateral malleolus after laceration 6 weeks previously in a Tanzanian refugee camp. An unknown antibiotic was given for 2 weeks before travel to Australia. On the patient’s arrival in this country, the wound appeared purulent, erythrocyte sedimentation rate was 82 mm/h (reference range [RR] < 10 mm/h), and C-reactive protein level was 9 IU/L (RR < 5 IU/L). Plain x-rays and bone scans suggested osteomyelitis. A wound swab grew S. aureus, Streptococcus pyogenes and Pseudomonas species. The patient was initially given intravenous cefazolin, and then definitive therapy (for MRSA and S. pyogenes, ignoring the colonising pseudomonad) with oral clindamycin (450 mg three times daily). Clinical resolution was complete, and levels of acute-phase reactants returned to normal. The antibiotic sensitivity pattern of the S. aureus isolate raised suspicion that it might be an unusual strain: it was resistant to methicillin, tetracycline and trimethoprim–sulfamethoxazole, but sensitive to erythromycin, clindamycin, ciprofloxacin, gentamicin, vancomycin, linezolid, mupirocin, rifampicin, fusidic acid and chloramphenicol. The mecA gene was detected by polymerase chain reaction testing, confirming methicillin resistance. The organism possessed staphylococcal cassette chromosome mec (SCCmec) element type IV. The Panton–Valentine leukocidin gene, staphylococcal enterotoxins A to E and toxic shock syndrome toxin-1 were not detected. As DNA fingerprinting with standard pulsed-field gel electrophoresis showed a novel banding pattern, the “gold standard” of multilocus sequence typing was used for identification. This confirmed an ST140 allelic profile, which has not been seen previously in Australia.4 On the balance of probabilities, the isolate represents an African community MRSA strain, not previously detected in Australia. Non-multiresistant community MRSA is not widely recognised in African countries. Hospital MRSA rates vary widely in Africa (eg, between 21% and 30% of all S. aureus isolates in Nigeria, Kenya and Cameroon, and fewer than 10% in Tunisia and Algeria5), but most are multiresistant. Medical practitioners in Australia who treat African refugees need to be aware that pyogenic soft tissue infections could be caused by community MRSA, and these MRSA strains may have a different antibiotic sensitivity profile to Australian community MRSA strains. It is essential to take appropriate specimens for microbiological analysis (wound swabs and possibly blood cultures and/or tissue samples), as antibiotic susceptibility profiles are increasingly unpredictable.

Annabelle Donaldson MB ChB · Iain B Gosbell MD, FRACP, FRCPA

To the Editor: We recently treated two patients with invasive meningococcal disease presenting with cellulitis. This presentation contributed to a delay in diagnosis and appropriate antibiotic treatment. The first patient was a 33-year-old woman, recently diagnosed with nephrotic syndrome, who had been unwell for a week with mild upper respiratory tract symptoms. During this time, her nephrologist began treating her with prednisolone (15 mg daily). The day before presentation, she developed abdominal pain, vomiting, chills, myalgia and headache. A rash developed on the day she presented to hospital. The temperature was 39.2°C, heart rate 148 beats per min, and blood pressure 146/57 mmHg. She had an area, measuring 20 cm × 20 cm, of tender cellulitic rash on the right thigh (Box) and mild neck stiffness. The diagnosis was initially unclear, leading to a delay of several hours before ceftriaxone was administered, and a lumbar puncture performed. Cerebrospinal fluid (CSF) examination revealed a leukocyte count of 4500 × 106/L (98% polymorphs) (reference range [RR], < 5 × 106/L), gram-negative diplococci, and protein concentration of 2073 mg/L (RR, 150–450 mg/L). The patient subsequently required intensive care admission for non-invasive ventilation and inotropic support. Neisseria meningitidis serotype C was detected in the CSF by polymerase chain reaction testing. The patient was discharged well except for mild headache and lethargy after 6 days. At 1-week review, she remained lethargic but was otherwise well. The rash was slowly resolving. The second patient was a 51-year-old woman with fever and a 2-day history of progressive pain, swelling and erythema of the anterolateral area of the neck. The temperature was 38.5°C, heart rate 115 beats per min, and blood pressure 134/86 mmHg. There was no evidence of upper airway involvement. The anterior area of the neck and upper chest wall were swollen, erythematous, tender and warm. No fluid collections or masses were detected on ultrasound examination. The patient was admitted to hospital with a diagnosis of cellulitis, and treatment was begun with intravenous flucloxacillin and metronidazole. After 17 hours, culture of blood taken on admission showed N. meningitidis serotype W135. Antibiotic treament was changed to ceftriaxone. After 5 days, the patient had mild residual inflammation and tenderness of the neck. She completed another week of treatment with oral amoxycillin. Only 14 cases of N. meningitidis cellulitis have been published.1-3 Seven cases involved children with periorbital cellulitis. In adults, three cases involved the face and neck, and four the limbs. N. meningitidis was isolated from blood (eight patients), conjunctival swabs (three), aspirates of the cellulitic areas (two) or CSF (one). There was one death: an elderly woman with bacteraemia and cellulitis of the face and neck.2 As illustrated by our cases, the many guises of meningococcal disease continue to challenge clinicians. Cellulitis in Patient 1 Cellulitic area on the right thigh of a 33-year-old woman with Neisseria meningitidis meningitis.

Karina J Kennedy MB BS · Jhumur Roy MB BS, FRCPA · Paul Lamberth MB BS, FACEM

To the Editor: Q fever is a serious disease caused by Coxiella burnetii, and usually occurs in people exposed to livestock. In Australia, acute Q fever generally manifests as nonspecific febrile illness. We report a case of spontaneous splenic rupture as a complication of acute Q fever acquired in Australia. A 29-year-old, previously well man presented in March 2005, with 5 days of fever, rigors and severe headache. On examination he had a temperature of 40°C, was tachycardic at 100 beats per minute, but normotensive. There were no focal examination findings. Investigations revealed thrombocytopenia at 123 × 109/L, a normal white cell count, and clear chest x-ray. The patient worked at a factory that processed animal placentas and fetal products. A provisional diagnosis of Q fever was made based on the illness and this exposure. The man was admitted and treated with empiric doxycycline, penicillin, and ceftriaxone. On Day 1 of the admission, he developed sudden, severe, left upper quadrant abdominal pain with shoulder tip radiation, diaphoresis and hypotension. An urgent computed tomography scan of the abdomen revealed splenomegaly with a diameter of 14 cm, and a crescentic, subcapsular splenic haematoma with rupture into the peritoneal space (Box). There was no history of trauma. He was admitted to the high dependency unit for monitoring. His haemoglobin level dropped from 151 g/L to 102 g/L, but he was managed conservatively and discharged from the high dependency unit 24 hours later. Acute Q fever was confirmed by polymerase chain reaction on Day 2, and antibiotic therapy was simplified to doxycycline 100 mg twice daily for 14 days. He was discharged from hospital on Day 5, and recovered fully. Seroconversion to Q fever was subsequently confirmed. His workplace now practices Q fever prevention policies, including pre-employment vaccination. Common presentations of Q fever include nonspecific febrile illness, pneumonia and hepatitis. There are five reported cases of Q fever associated with spontaneous splenic rupture, but this is the first Australian case.1-5 The other patients presented with flu-like symptoms and abdominal pain of 2–14 days’ duration, and required splenectomy between Day 1 and 4 of admission. Clinicians should be aware of splenic rupture as a potential complication of Q fever in Australia. Abdominal computed tomography scan A crescentic, subcapsular splenic haematoma is visible (arrow).

Amanda J Wade MB BS · Tim Walker MB BS · Eugene Athan MB BS, FRACP · Andrew J Hughes MB BS, FRACP

Kari A J Jarvinen,* Bradley J McCall,† Clare B Nourse,‡ Joe G McCormack,§ Martyn H Tilse¶ * Senior Public Health Registrar, † Public Health Medical Officer, Communicable Disease Control, Brisbane Southside Public Health Unit, 39 Kessels Road, Coopers Plains, QLD 4108; ‡ Paediatric Infectious Diseases Physician, § Director of Infectious Diseases, ¶ Director of Microbiology, Mater Health Services, South Brisbane, QLD. kari_jarvinenAThealth.qld.gov.au To the Editor: Pertussis continues to be a significant public health problem in Australia. Children aged under 1 year are most at risk from severe, life-threatening complications from the disease.1 Traditionally, erythromycin has been the drug of choice for treatment of cases and prophylaxis in selected contacts. However, its use in neonates is known to carry a risk of infantile hypertrophic pyloric stenosis.1,2 Its propensity to cause QT prolongation and ventricular arrhythmias is also well described.2,3 Both azithromycin and clarithromycin have been recently recommended as suitable alternatives for management of pertussis.2,4 The US Centres for Disease Control now regard azithromycin as the agent of choice for neonates less than 1 month of age.1 There is evidence suggesting azithromycin has less pro-arrhythmic potential than erythromycin or clarithromycin.5,6 Azithromycin does not interact significantly with the hepatic cytochrome P450 system and has less potential for significant drug interactions than other macrolide antibiotics.3,5,6 Azithromycin and clarithromycin also require less frequent administration (1–2 doses per day) and shorter treatment regimens (5–7 days) than erythromycin. In Australia, roxithromycin is the most widely prescribed macrolide antibiotic. However, there are no clinical studies on its effectiveness in pertussis, and in-vitro sensitivity studies suggest it may be inferior to erythromycin. Thus, roxithromycin cannot be recommended in pertussis.4 Updated versions of Australian antibiotic guidelines to be released later this year will recommend azithromycin for pertussis treatment and prophylaxis. However, access to azithromycin for this purpose in Australia is currently limited by the restrictions placed on prescribing through the Pharmaceutical Benefits Scheme (PBS). Azithromycin is currently approved for Chlamydia trachomatis urethritis, cervicitis and trachoma. Pertussis is an approved indication only for the use of 500 mg tablets under the Repatriation PBS. This restriction has important implications for the effective and safe management of pertussis in Australia. Widespread use of newer macrolides in the community is not advisable because of the propensity of macrolides to induce antibiotic resistance, and their greater cost. However, for pertussis infection, Australians need to be able to access agents such as azithromycin. PBS restrictions for this indication need to be revised, for both tablet and liquid formulations.

Kari A J Jarvinen · Bradley J McCall · Clare B Nourse · Joe G McCormack · Martyn H Tilse

Clinical records Patient 1 A 75-year-old woman, who was born in Greece and migrated to Australia 50 years ago, presented with acute shortness of breath secondary to cardiac failure due to valvular heart disease. Her only cardiovascular risk factor was hypertension. A transthoracic echocardiogram 1 month earlier had shown moderate aortic regurgitation, moderate left ventricular dysfunction, and a markedly dilated ascending aorta measuring 5.8 cm (normal, < 3.4 cm). Her symptoms were alleviated with diuresis and medication to treat cardiac failure. She underwent cardiac catheterisation, which revealed a dilated aortic root and ascending aorta, measuring 4.3 cm and 6 cm, respectively, with normal coronary arteries. Syphilis serology, requested because of the dilated aortic root, was positive with a rapid plasma reagin titre of 1 : 2 and a reactive Treponema pallidum particle agglutination test. Because of penicillin allergy, intravenous ceftriaxone 1 g daily was given for 21 days. No lumbar puncture was performed. She had an aortic valve replacement, and replacement of the ascending aorta with a synthetic graft. Macroscopically, the aortic wall was extremely thick. Histological sections showed features consistent with syphilitic aortitis (Figure 1). Her recovery was slow, but she is now living independently 15 months after surgery. 1 Aortic wall of Patient 1 A: Aortic wall (2 x objective) with thickened media (m) and intima (i). B: Aortic media with inflammatory infiltrate of plasma cells (arrowhead) and lymphocytes (arrow) (40 x objective). C: Aortic adventitia with narrowed vasa vasorum (endarteritis obliterans) (20 x objective). Patient 2 A 40-year-old Indonesian man who had lived in Australia for 14 years presented with acute onset severe central chest pain, on a background of similar but less severe pain over the previous 2 months. He smoked one packet of cigarettes per day, but had no other risk factors for coronary artery disease. He was in a monogamous heterosexual relationship of 18 months. His electrocardiogram revealed 3 mm ST elevation in the anterior leads (V1–3) and reciprocal ST depression in the inferior leads (II, III, aVF). His serum troponin level was 4.9 μg/L (reference range, 0–0.4 μg/L). A coronary angiogram revealed a completely occluded left main coronary artery at the ostium, and a 90% ostial lesion of his right coronary artery; the remainder of his right coronary system was clear of disease (Figure 2). He underwent emergency coronary artery bypass grafting. The operative notes do not comment on the appearance of the native coronary arteries, nor were these sent for histological examination. An aortic wall biopsy showed mild degenerative changes only — syphilis could not be excluded as causing these changes. In view of the bilateral coronary ostial lesions, syphilis serology was performed. His rapid plasma reagin titre was 1 : 64, together with a reactive Treponema pallidum particle agglutination test. Treatment with intravenous penicillin 1.8 g every 4 hours was given for 15 days, and prednisolone 25 mg twice daily for 2 days to prevent a Jarisch–Herxheimer reaction. The patient declined a lumbar puncture. 2 Coronary angiograms of Patient 2 \ A: Tapering of the aortic root (thin arrows) and left main coronary stump (arrowhead). B: 90% occlusion of right coronary ostium with no distal disease. Details of five patients The details of these two patients and another three patients seen between 1998 and 2004 are summarised in the Box. Although confirmatory histology of cardiovascular syphilis was not available for four of the five patients, this was considered the most likely diagnosis. All patients were born outside Australia. None of the three men reported having sex with men. No patients reported a history of syphilis. Syphilis has become a rare disease, although peaks of syphilis notifications occurred in Australia in the mid 1970s to mid 1980s and again more recently.1 Clinicians need to be aware of the manifestations of syphilis, and to consider the diagnosis outside the groups considered at risk of this infection in the modern era, such as men who have unprotected sex with other men. The five patients in our series ranged in age from 40 to 77 years. Notably, all were born overseas, and none belonged to a group considered at high risk of syphilis in the contemporary Australian context. Serology was consistent with acquisition of infection in the distant past in four of the five patients. Among patients with untreated syphilis, aortitis occurs in up to 70%–80%, and the clinically apparent manifestations of aortic regurgitation, coronary ostial stenosis and aortic aneurysms are seen in 10%–15%.2 Although aortic regurgitation is rarely caused by syphilis, it occurs in 20%–30% of patients with syphilitic aortitis.3 Coronary ostial lesions may be seen in 20%–25% of patients with syphilitic aortitis,3 but it is uncommon for such coronary ostial lesions to lead to acute myocardial infarction.4 In contrast, ostial lesions are only seen in 0.1% of patients with coronary artery disease, and bilateral lesions are even less common.5 The high plasma reagin titre seen in Patient 2 is also unusual in cardiovascular syphilis, but has been previously noted.6 Lessons from practice Cardiovascular syphilis may occur in patients outside of the usual risk groups for syphilis. Testing for syphilis serology should be considered in patients with aortic regurgitation, and particularly bilateral coronary ostial lesions or aortic aneurysms. Cerebrospinal fluid should be examined in patients with cardiovascular syphilis to exclude neurosyphilis. Patients with bilateral coronary ostial lesions but no distal coronary artery disease and those with ascending aortic aneurysms should be screened for syphilis. Screening should be considered for people with aortic regurgitation, especially those with risk factors, including birth in a country where syphilis has been or continues to be common. Although rates of neurosyphilis are low in patients with cardiovascular syphilis,3,7 a cerebrospinal fluid examination is recommended to exclude neurosyphilis. In addition, screening for other sexually transmitted infections, including HIV, should be considered, and appropriate contact tracing instituted. Sexual transmission of syphilis occurs only when mucocutaneous lesions are present, but long-term sexual partners of patients with late latent or tertiary syphilis should be screened serologically. The duration and route of penicillin therapy for cardiovascular syphilis are controversial. Each of our five patients received a different therapy. Penicillin has never been formally evaluated as treatment for cardiovascular syphilis, but a study from the 1950s showed that few patients had progressive disease after penicillin therapy, and up to 60% reported symptomatic relief.8,9 The Australian Therapeutic guidelines: antibiotic suggest intravenous benzylpenicillin for 15 days,10 the Australian National management guidelines for sexually transmissible infections suggest intramuscular procaine penicillin for 20 days,11 and the World Health Organization, European and United States guidelines suggest intramuscular benzathine penicillin 2.4 million units once weekly for three doses.12-14 These variations reflect the paucity of good clinical data comparing the different regimens. Although definitive evidence of the efficacy of benzathine penicillin is lacking, there is also no evidence to the contrary. There is little evidence that Jarisch–Herxheimer reactions complicate treatment of cardiovascular syphilis.8,9,15 Cardiovascular syphilis: one proven and four presumptive cases Age, sex and country of origin Indications for syphilis testing Syphilis serology and cardiovascular risk factors Management and follow-up 75 years, female, Greece Aortic regurgitation; dilated aortic root RPR: 1 : 2 TPPA: reactive Hypertension Aortic valve and aortic root replacement. Histological confirmation of syphilitic aortitis. Treated for cardiovascular syphilis with intravenous ceftriaxone 1 g daily for 3 weeks. No evidence of a Jarisch–Herxheimer reaction. Lumbar puncture not performed. Follow-up 15 months after therapy: mild exertional shortness of breath. Follow-up syphilis serology has not been requested. 40 years, male, Indonesia Acute myocardial infarction; bilateral ostial lesions and no visible disease in the right coronary artery RPR: 1 : 64 TPPA: reactive Smoker CABG. Histology from aortic wall biopsy was inconclusive for cardiovascular syphilis; no biopsy of the coronary arteries. Presumptively treated for cardiovascular syphilis with intravenous penicillin 1.8 g every 4 hours for 15 days. Oral prednisolone 25 mg twice daily for 2 days was given to prevent a Jarisch–Herxheimer reaction. No evidence of a Jarisch–Herxheimer reaction. CSF normal (lumbar puncture performed after treatment completed). Follow-up 2 years after therapy: symptom-free with RPR 1 : 8. 56 years, female, Vietnam Angina; bilateral ostial lesions; minor irregularities of the left circumflex artery, but no visible disease in the left anterior descending and right coronary arteries RPR: 1 : 1 TPPA: reactive Type 2 diabetes mellitus Hypercholesterolaemia Ostial lesions were not sufficiently severe to require revascularisation procedures, so no biopsies were taken of the aortic wall or coronary arteries. Presumptively treated for cardiovascular syphilis with intramuscular procaine penicillin 1 g daily for 15 days. No evidence of a Jarisch–Herxheimer reaction. CSF normal (lumbar puncture performed after treatment completed). Follow-up 1 year after therapy: symptom-free with RPR 1 : 1. Cardiovascular MRI revealed no further evidence of ostial coronary disease. 77 years, male, Hong Kong Aortic regurgitation; dilated aortic root RPR: 1 : 2 TPPA: reactive Hypertension Type 2 diabetes mellitus Aortic regurgitation was not sufficiently severe to warrant valve replacement, so no aortic wall biopsy was obtained. Presumptively treated for cardiovascular syphilis with intramuscular benzathine penicillin 2.4 g weekly for 3 doses. No evidence of a Jarisch–Herxheimer reaction. CSF normal. Follow-up: returned to Hong Kong and lost to follow-up. Follow-up syphilis serology has not been requested. 72 years, male, Timor Wife had been found to have positive syphilis serology. Following positive syphilis serology, aortic regurgitation was noted on clinical examination and then echocardiographically RPR: Non-reactive TPPA: reactive No cardiovascular risk factors present Aortic regurgitation was not sufficiently severe to warrant valve replacement, so no aortic wall biopsy was obtained. Presumptively treated for cardiovascular syphilis with intravenous penicillin 1.8 g 4 hourly for 14 days. Intravenous hydrocortisone 50 mg four times a day given for 48 hours because of concern about a Jarisch–Herxheimer reaction when atrial fibrillation occurred on antibiotic therapy. No other manifestations of a possible Jarisch–Herxheimer reaction. Atrial fibrillation reverted to sinus rhythm with amiodarone and sotalol therapy. CSF normal. Follow-up: ongoing medical management for cardiac failure. Repeat syphilis serology 3 years after therapy, showed RPR non-reactive. CABG = coronary artery bypass graft. CSF = cerebrospinal fluid. MRI = magnetic resonance imaging. RPR = rapid plasma reagin titre. TPPA = Treponema pallidum particle agglutination test.

Steven Y C Tong MB BS · Alan C Street MB BS, FRACP · Haris Haqqani MB BS

Graham Simpson,* Paul Clark,† Trevor Knight‡ * Director of Thoracic Medicine and Regional TB Control Unit, † Resident Medical Officer, ‡ Nurse Unit Manager, Department of Thoracic Medicine, Cairns Base Hospital, Cairns, QLD 4870. fgsimpsonATiig.com.au To the Editor: Australia has a low incidence of tuberculosis (TB), which has remained constant for over a decade.1 However, the incidence is not uniform across the population; immigrants and Indigenous Australians have higher rates. An audit of all cases of TB in Far North Queensland over 5 years showed an incidence of 35.9/100 000 per annum in Indigenous Australians, and poor outcomes in this group.2 This finding led to a number of policy changes, including an increase in directly observed therapy (DOT), made possible by increased use of Aboriginal health care workers in remote communities, and more aggressive and prolonged treatment of relapses. A follow-up audit was undertaken to assess the effect of these changes. The results are shown in the Box for both time periods. New cases of TB in Indigenous Australians were significantly reduced (P < 0.0001 by Fisher’s exact test), and DOT had increased significantly (P < 0.0001). The number of deaths from TB had declined, as had relapses, but these falls were not statistically significant. There were no deaths among Indigenous Australians during the second 5-year period. Of the people who died in this period, three were elderly men suspected of having cancer, and one was a patient from Papua New Guinea (PNG) who had HIV co-infection with TB. The most striking finding was the dramatic increase in cases in people from PNG (P < 0.0001). The outer Australian islands in the Torres Strait are only 3 kilometres from the PNG coast, and there is free movement of people across the border under a treaty arrangement. Although there are no precise figures,3 it is clear that there are epidemics of both TB and HIV in PNG, and that these have extended to rural areas. Specialist outreach clinics with x-ray facilities have been established on the outer islands, but numbers have continued to rise. In 2005, of 38 cases of TB in Far North Queensland, 26 were from the Torres Strait including seven cases of multidrug resistant TB. This represents a significant public health threat and highlights the importance of local audits of TB control, as state and national data may not be adequate to identify emerging local problems. Findings of two 5-year audits on tuberculosis in Far North Queensland Findings 1993–1997 1998–2002 Total cases 87 92 Indigenous Australians 50 22 Non-Indigenous 30 26 Papua New Guineans 7 44 Pulmonary tuberculosis 54 57 Sputum smear positive 67% 47%* Directly observed therapy 18 (21%) 67 (73%) Death from tuberculosis 10 4 Deaths in Indigenous Australians 7 0 Total early relapses 7 2 Indigenous Australians 7 0 Drug resistance 6 7 Multidrug resistant tuberculosis 1† 3‡ HIV co-infection 0 2‡ * Queensland average, 48%. † Patient from the Philippines. ‡ All in Papua New Guineans.

Graham Simpson · Paul Clark · Trevor Knight

To the Editor: Johnson et al detailed an intensive hand hygiene program planned to reduce the burden of nosocomial methicillin-resistant Staphylococcus aureus (MRSA) infections.1 The results were based on observations before and after the program. Hand hygiene compliance rates reached only 42% despite the program, and there was no effect on patient MRSA colonisation or environmental colonisation or contamination. Outside the intensive care unit, there was no effect on health care worker colonisation. Despite this evidence of ineffectiveness, the program was held responsible for a reduction in hospital-wide rates of clinically important MRSA infections. The literature on hand hygiene is inadequate. The recent edition of Clinical evidence contains no randomised controlled trials of hand hygiene.2 In fact, the only published randomised trial is the Mortimer study,3 which is now more than 40 years old. Huynh and Commens made the point that hand hygiene procedures involving application of chemical agents or scrubbing are hazardous for staff and suggested using mechanical barriers (ie, gloves) on clean unscrubbed hands.4 The hand hygiene bandwagon rolls on despite the absence of evidence of benefit for patients and its hazardous nature for staff. Mechanical barriers together with reduced contamination opportunities (hand-shaking, touching telephone handsets and computer key boards) may be better options. We need properly conducted studies to find an effective means of protecting patients from nosocomial infections by MRSA and other agents.

Keith V Woollard

To the Editor: We congratulate Johnson et al on their article, which illustrates a successful hand hygiene program associated with a fall in transmission of multidrug-resistant organisms.1 Their publication is significant for three reasons: It is only the second article2 to demonstrate the anticipated relationship between increased hand hygiene and a fall in multidrug-resistant organism transmission; It again indicates that 100% compliance in hand hygiene is not necessary to significantly improve outcomes;2,3 and It suggests that environmental contamination with MRSA has little relationship to patient colonisation. However, the specific aspects of their program that led to success are not obvious, and may not relate to a sustained response to either education or the provision of alcohol/chlorhexidine hand hygiene solution. Alternative explanations include: The education program and/or overt observation induced a Hawthorne effect on hand hygiene practice; The screening and treatment program induced the same effect on hand hygiene behaviour; or The treatment of MRSA carriers reduced the size of the MRSA reservoir and thus the probability of transmission and subsequent colonisation. Although we agree that the approach of Johnson et al is laudable, without teasing out those causal factors that induce the improvement in hand hygiene in health care workers, it remains expensive to implement and maintain. Moreover, there is no evidence that improved hand hygiene would continue if the alcoholic gel alone remained, without all other aspects of the program. This was recognised in the successful Geneva program on which the protocol used by Johnson et al was modelled. In that study, the authors remained so uncertain as to what elements of the program were causal that they stated: Whether improved hand-hygiene practice will outlast the intervention remains uncertain; we decided to refrain from testing this issue by maintaining a permanent component of the intervention.2 Evidence currently available4 and soon to be amplified5 suggests that hand hygiene practice in health care workers is simply an extrapolation of their community behaviour. Unfortunately, community hand washing behaviour is not microbiologically founded, being developed on the basis of emotion not science. Both the Austin and Geneva protocols supported the introduction of alcoholic hand gel with strong promotion of specific behavioural elements to induce change in hand hygiene practice. Our findings suggest that alcoholic gel is not pivotal to the improvement of hand hygiene, in that behavioural modelling suggests its effect is relatively small and very dependent on concomitant behavioural change.4,5 The World Health Organization World Alliance for Patient Safety has recently advocated the introduction of alcoholic gel into all hospitals.6 While not denying that this is a step toward improvement, we strongly caution against unrealistic expectations of this single intervention. The hand hygiene practices of health care workers are learned behaviours of childhood, continued as professionals, and reinforced in everyone’s daily lives.4,5 Entrenched, longstanding behaviour patterns will not be changed in a sustained fashion by the introduction of a new hand hygiene product.

R Michael Whitby · Mary-Louise McLaws

In reply: Woollard is critical of the lack of randomised controlled data to support the use of alcohol/chlorhexidine hand rub solution (ACHRS). Although a placebo-controlled study would be ideal, it is doubtful whether one could be performed. Apart from the complexity of design and cost, there would be the requirement to ask patients to consent to being treated in a hospital where there was a substantial risk of nosocomial sepsis, but where half the health care workers would not have clean hands when attending them. Woollard argues that our failure to reduce colonisation or contamination with MRSA shows that our project failed. However, he offers no alternative explanation for the reduction in MRSA bacteraemia, clinical MRSA isolates and resistant gram-negative bacteria that we reported. Our project was a multimodal quality intervention, and we cannot know which component of the project resulted in the benefit, or whether the improvement should be attributed to other confounders, as suggested by Whitby and McLaws. However, we have presented all our data so that readers can draw their own conclusions. It seems unlikely to us that the intervention on which we concentrated our major effort, the progressive introduction and promotion of ACHRS, would be the one component that failed to contribute to the improvement. Woollard also mentions the potential toxicity of asking health care workers to scrub with a chemical agent, and proposes the use of gloves instead. Our ACHRS is a quick to apply, self-drying solution. It is rubbed on the hands, but scrubbing is not required. We actively monitored rates of cutaneous reactions and found it to be extremely well tolerated.1 Gloves must be changed between patients or when moving from a dirty to a clean site.2 We know that busy health care workers often do not have time to do this, and that hands can become contaminated despite the use of gloves.3 We agree with Whitby and McLaws that simply providing ACHRS, without an active campaign to support its use, is pointless. The provision and promotion of ACHRS is a tool to assist health care workers improve hand hygiene, and is just one component in a web of interventions needed to control nosocomial sepsis. Whether it is cost-effective depends on the largely unknown costs to Australian hospitals of preventable infections. At our institution, we believe that it is worth the money, and continue to require all clinical staff and students to know where to find and when to use ACHRS before they start work.

Paul D R Johnson · M Lindsay Grayson

Annabelle D Donaldson,* Raymond C Chan,† Iain B Gosbell‡ * Infectious Diseases Registrar, Department of Microbiology and Infectious Diseases, Liverpool Health Service, Sydney, NSW; † Infectious Diseases Physician, ‡ Director, Department of Microbiology and Infectious Diseases, South Western Area Pathology Service, Locked Bag 7090, Liverpool BC 1871, and Conjoint Associate Professor, School of Medical Sciences, University of New South Wales, Sydney, NSW. i.gosbellATunsw.edu.au To the Editor: Community-acquired methicillin-resistant Staphylococcus aureus (CA-MRSA) is usually susceptible to a wider range of antibiotics than nosocomial or multiresistant MRSA, but evidence is lacking to guide antibiotic choices. Rifampicin plus fusidic acid is commonly used in multiresistant MRSA infection, and familiarity makes this combination attractive for CA-MRSA. We report two cases of CA-MRSA infection in which rifampicin resistance developed during treatment. Non-compliance may have been a significant factor, but high bacterial load and persistent foci of infection were also likely contributors. We caution against using regimens containing rifampicin in such cases. Patient 1: A previously well 16-year-old girl presented with a 4-day history of left knee pain. She had fever and a moderate joint effusion. Plain x-rays were unremarkable, and blood cultures showed no growth. An aspirate of the knee joint had a white blood cell count of 1100 × 106 cells/L, but showed no growth on culture. Non-multiresistant MRSA (sensitive to all non-β-lactams) was subsequently grown from a synovial biopsy specimen. She was treated with 7 weeks of intravenous vancomycin, followed by 6 months of oral rifampicin (450 mg daily) plus fusidic acid (500 mg twice daily). Her compliance was uncertain, and she did not return for follow-up appointments. The patient presented again 12 months later with increasing pain in the left knee and thigh. X-ray and magnetic resonance imaging revealed extensive osteomyelitis of the distal femur. Culture of purulent material removed at sequestrectomy grew MRSA with the same sensitivity profile as previously, but now resistant to rifampicin. She was treated with clindamycin, but required further sequestrectomy. She continues to take clindamycin long term. Patient 2: A 74-year-old man developed severe cellulitis and an abscess of his left hand after a golfing injury. He had bacteraemia with non-multiresistant MRSA (sensitive to all non-β-lactams). He also had significant pain in his left prosthetic hip. After 7 weeks of intravenous vancomycin and oral moxifloxacin, levels of inflammatory markers remained high. He was treated with oral rifampicin (600 mg daily) plus fusidic acid (500 mg twice daily) for several weeks until lost to follow-up. Four months later, the patient still had pain in the hip, and rifampicin plus fusidic acid treatment was begun again. After 3 months with no improvement, he underwent surgical exploration, and the loose femoral prosthesis was replaced. Culture of debrided material showed MRSA with the same sensitivity pattern as previously, but now resistant to rifampicin. He was treated with intravenous vancomycin for 6 weeks, followed by trimethoprim–sulfamethoxazole, which was later changed to clindamycin because of intolerance. He remains taking indefinite clindamycin suppression therapy. Traditional oral therapy for MRSA infection is rifampicin plus fusidic acid, but alternative oral agents for non-multiresistant CA-MRSA include clindamycin,1 tetracyclines2 and trimethoprim–sulfamethoxazole.3 There is wider experience with clindamycin, which penetrates well into skin and soft tissues, has good oral bioavailability, and has been used successfully.1 However, a concern is erythromycin-inducible clindamycin resistance, which may lead to clindamycin failure, particularly in severe infections.4 Its prevalence varies. Studies in vitro have shown that rifampicin resistance develops as readily in CA-MRSA as in nosocomial MRSA,5 through a single-step mutation. Resistance did not develop in vitro to clindamycin,5 suggesting it may be a more reliable alternative in situations where compliance is uncertain, the bacterial load is high, or the source (eg, an infected prosthesis) cannot be removed. Further in-vitro and in-vivo investigations are urgently required to determine the optimal drug therapy for CA-MRSA infections.

Annabelle D Donaldson · Raymond C Chan · Iain B Gosbell

Objective: To describe the epidemiology of invasive pneumococcal disease (IPD), and the impact of pneumococcal vaccines on IPD, in Indigenous people in north Queensland.Setting: North Queensland, 1999–2004; there are about 53 750 Indigenous people in the region, including nearly 6900 children < 5 years and nearly 5650 adults ≥ 50 years.Main outcome measures: Incidences of IPD in Indigenous children and in Indigenous adults compared between the 3 years before and after the introduction of a 7-valent pneumococcal conjugate vaccine (7vPCV) (1999–2001 versus 2002–2004).Results: Estimated annual incidence of IPD in Indigenous children < 5 years of age declined from 170 to 78 cases per 100 000 in the 3 years following the introduction of 7vPCV in 2001. The annual incidence of vaccine-preventable IPD in Indigenous adults had declined by 86% since a 23-valent pneumococcal polysaccharide vaccine (23vPPV) was introduced to the region in 1996, to 15 cases per 100 000 (95% CI, 8–25) in 2002–2004.Conclusion: Although there was a rapid decline in IPD in young Indigenous children, it is unlikely that the incidence will fall much further with the current 7-valent vaccine. There was a suggestion that vaccinating Indigenous children indirectly protected those aged 5–14 years and Indigenous adults ≥15 years of age. Incidence of IPD in Indigenous adults in 2002–2004 was the lowest on record in the region.

Jeffrey N Hanna MPH, FAFPHM · Jan L Humphreys · Denise M Murphy DipMedTech

A 40-year-old woman presented with sudden onset of a painful neck swelling, fever and dysphagia. Examination revealed a warm, tender lump in her thyroid gland. Five weeks previously, she had had an abdominal hysterectomy with bilateral salpingo-oophorectomy for pelvic inflammatory disease. There was no evidence of leukopenia or diabetes. Computed tomography and x-ray images were consistent with a diagnosis of thyroid abscess (Figure 1 and Figure 2). Anaerobic culture of a fine needle aspiration of pus confirmed the presence of Bacteroides fragilis. After failing twice to resolve the lump by aspiration, and incision and drainage, a left thyroid lobectomy was performed, with a satisfactory outcome. Antibiotics cefotaxime and metronidazole were administered after operation. There are about 400 cases of acute suppurative thyroiditis reported in the medical literature,1 but Bacteroides fragilis as the causative agent has been documented only once before.1 More than two-thirds of the women and half of the men with acute suppurative thyroiditis have pre-existing thyroid disease.2 Acute suppurative thyroiditis may be secondary to bronchial cleft fistula, pyriform sinus fistula, thyroglossal duct, HIV/AIDS, diabetes, or leukaemia.1,2 1: Computed tomography scan showing an abscess in the left lobe of the thyroid and the presence of air (arrow). 2: X-ray of the cervical spine (lateral view) showing soft tissue swelling in the anterior part of the neck and the presence of air (arrow).

Sashidhar Yeluri MBBS, MS · Jayshree P Mehta MB BS, MS · Siddharth Karanth MB BS, MS · Guneesh Dadayal MB BS

Clinical record A 65-year-old beef-cattle farmer with a 9-year history of sarcoidosis presented with a 2-year history of depressed mood with cognitive decline, weight loss and vomiting. Treatment with three antidepressant medications had proved ineffectual. After bronchoscopic diagnosis, his sarcoidosis had been treated for the ensuing 9 years with prednisolone (initially 30 mg/day and slowly reduced to 2 mg/day). At the time of initial diagnosis, three sputum samples had stained positive for acid-fast bacilli, and the patient was placed on conventional antituberculous therapy. This was ceased after 3 months, following negative mycobacterial cultures of sputa and bronchoscopic specimens. His family history was unremarkable. On examination, the patient appeared cachectic and older than his stated age. He was vague, with slow speech. Bedside cognitive testing revealed severe memory impairment, with moderate impairment in attention, visuo-constructional ability and executive function. The result of a mini-mental state examination (MMSE) was 17/30. Neurological examination revealed ataxia, brisk tendon reflexes, bilateral habituating palmomental reflexes and proximal weakness. There were no other physical abnormalities. Magnetic resonance imaging (MRI) scans of the brain showed ventricular enlargement, supratentorial white-matter hyperintensities, and extensive leptomeningeal enhancement in the posterior fossa, brain stem and basal cisterns, with extension along the cranial nerves (Figure 1). The MRI results were reported as consistent with neurosarcoidosis. A SPECT (single-photon emission computed tomography) scan, undertaken to differentiate possible causes of dementia, demonstrated global hypoperfusion (Figure 2). Cerebrospinal fluid (CSF) analysis revealed a raised protein level (3.02 g/L; reference range, 0.15–0.45 g/L) and an opening pressure within the normal range. Cryptococcal antigen was detected in the CSF at a titre of 1 : 4. Although India-ink staining for Cryptococcus was negative, Cryptococcus neoformans var. neoformans was isolated on culture. Serum cultures were negative for fungae. Notably, no malignant cells or bacteria (including acid-fast bacilli) were identified from stained smears. Mycobacterial cultures showed no growth after 8 weeks. A number of measurements showed that the patient’s sarcoidosis was stable: serum angiotensin-converting enzyme and 24-hour urine calcium levels were normal; respiratory function testing showed moderate air flow obstruction and gas transfer, with no change from previous measurements; and a chest x-ray (Figure 3) showed no new changes. The patient was diagnosed with cryptococcal meningitis and secondary hydrocephalus. He was initially given antifungal therapy for 4 weeks (intravenous amphotericin B 40 mg a day and oral 5-flucytosine 1.5 g four times a day, followed by intravenous ambisome 200 mg a day when he developed renal impairment with toxic accumulation of 5-flucytosine). Twelve months’ oral fluconazole therapy (400 mg a day) was planned, and a ventriculoperitoneal shunt was inserted. The patient’s mental state had improved when he was re-examined 2 weeks after commencement of therapy (MMSE result, 23/30), and cryptococcal antigen titres fell to levels below detection. The patient’s longer-term course was complicated by worsening confusion, with chronic subdural haematoma and bilateral posterior cranial fossa infarcts seen on MRI scans done at intervals of several weeks. To ensure that inadequately treated cryptococcosis was not contributing to his cognitive decline, treatment with amphotericin B and 5-flucytosine was reinstituted for 3 weeks. Three months after discharge, the patient was well, with moderate cognitive impairment. Sarcoidosis is a multisystem granulomatous inflammatory disorder of unknown cause. It affects the lungs in 90% of affected people and the nervous system in 5%–15%. About 5% of people with sarcoidosis have clinically evident neurological involvement (neurosarcoidosis).1 It most commonly presents as a cranial neuropathy, although other presentations — such as aseptic meningitis; mass lesions in the pituitary, hypothalamus, and other brain regions; spinal and peripheral neuropathy; and epilepsy — may occur. Hydrocephalus is uncommon, but can occur secondary to basal obstruction and may present with a progressive dementing syndrome.2 Neuroimaging may reveal focal lesions or periventricular and leptomeningeal enhancement, but results are often non-specific and do not correlate with disease activity.3 Tests on cerebrospinal fluid (CSF) may show an increased white cell count and/or raised protein and angiotensin-converting enzyme levels, but the CSF is normal in up to a third of patients.4 Steroids are the mainstay of treatment for neurosarcoidosis, but recalcitrant illness may require treatment with cytotoxic drugs and irradiation.5 Given the protean nature of presentation of neurosarcoidosis, it is very difficult to diagnose in patients without a history of systemic sarcoidosis. Furthermore, it may be restricted to the central nervous system in some patients.6 Additionally, in patients with known systemic sarcoidosis, the risk is that any new neurological condition will be attributed to the systemic disease without a new diagnosis being sought.2 Lessons from practice Chronic meningitis may present with atypical depression and/or dementia. Sarcoidosis is associated with an increased risk of infection with Cryptococcus neoformans, independent of immunosuppression. Cryptococcal meningitis is a treatable and reversible cause of dementia. An association has been postulated between systemic sarcoidosis and infection via inhalation of Cryptococcus neoformans, a dimorphic fungus found in soil. While it is well recognised that cryptococcosis is seen in patients with reduced cell-mediated immunity (such as those with AIDS), patients with sarcoidosis may also have an increased risk of infection independent of immunosuppressive therapy. Up to half of patients with sarcoidosis manifest alterations in cell-mediated immunity.7 The annual incidence of cryptococcal meningitis in Victoria is 4.9 cases per million population, which is lower than the national incidence of 6.6 per million.8 It typically presents with meningism and fever, but may masquerade as Alzheimer’s disease or vascular dementia,9,10 and is often complicated by elevated intracranial pressure or hydrocephalus.11 Typical CSF findings are lymphocytic pleocytosis, elevated protein levels, reduced glucose levels, and, commonly, an increased opening pressure. Cryptococcal meningitis is initially treated with amphotericin B, with or without flucytosine, followed by prolonged oral fluconazole.12 While there are data to support the exclusive use of fluconazole, complicated cases with severe presentations, such as that described in the case presented here, should be treated with amphotericin B and flucytosine.13 Cognitive impairment secondary to communicating hydrocephalus in cases of cryptococcal meningitis often responds to treatment with ventriculoperitoneal shunting.11 The differentiation of cryptococcosis from neurosarcoidosis in a patient with known sarcoidosis is difficult, as there is significant clinical overlap between the two disorders. CSF examination in both may show a raised cell count and protein level, and magnetic resonance imaging (as in our patient) may demonstrate non-specific leptomeningeal enhancement and hydrocephalus. In the setting of suspected neurosarcoidosis in patients with cognitive impairment and clinical or neuroradiological suggestion of meningeal involvement, the exclusion of cryptococcal disease through CSF analysis and culture is mandatory. Cryptococcal meningitis is among a range of inflammatory, infectious, metabolic and neoplastic/paraneoplastic conditions that can closely mimic the presentation and course of more established dementing disorders, and which can be diagnosed or excluded by a careful combination of clinical assessment and choice of investigations. Although it often presents with symptoms of meningism, it may progress indolently and lead to cognitive impairment (resulting from hydrocephalus) in more than half of individuals.14 It should be considered as a cause of reversible dementia10 in at-risk individuals — including patients with a history of sarcoidosis — and when its presence is suggested by clinical presentation and/or findings on neuroimaging. 1 Magnetic resonance imaging scans Sagittal view (left), view through posterior fossa (middle), and axial view through hemispheres (right). Sagittal and axial views show ventricular enlargement and global cortical loss; posterior fossa view shows enhancement of the leptomeninges in the basal cisterns as bright white structures. 2 Axial slices of SPECT (single-photon emission computed tomography) scan Scan shows global hypoperfusion. Areas in the cortical rim appear yellow/green (low perfusion) rather than red/white (high perfusion). Ventriculomegaly is visible as enlarged, non-perfusing blue spaces containing cerebrospinal fluid. 3 Chest x-ray Hilar adenopathy characteristic of sarcoidosis was present, but there was no change from previous films.

Siddhartha Deb MB BS · Mark Walterfang FRANZCP · Daniel Varghese MB BS, BMedSci · Bruce Tomlinson FRACP · Dennis Velakoulis FRANZCP · Damon P Eisen MD, FRACP

Frank H Beard,* Jeremy M McAnulty,† John W Tapsall,‡ Angelo M Zaia§ * Senior Medical Officer, Communicable Diseases Unit, Queensland Health, 147-163 Charlotte Street, Brisbane, QLD 4001; † Director, Communicable Diseases Branch, NSW Health, Sydney, NSW; ‡ Senior Specialist in Microbiology, SEALS, Prince of Wales Hospital, Sydney, NSW; § Molecular Biologist, University of Melbourne, VIC. Frank_BeardAThealth.qld.gov.au To the Editor: We report two cases of serogroup B meningococcal disease, caused by genotypically indistinguishable organisms, where transmission is likely to have occurred on a school bus. To our knowledge, transmission of meningococcal disease on a bus has been reported only once before.1 In June 2005, two cases of serogroup B meningococcal disease in teenagers from the same school were reported to the Northern Sydney Public Health Unit. Patient 1 had symptoms of fever, headache, vomiting, and an erythematous rash. Two days after onset of this patient’s symptoms, Patient 2 also developed fever, headache and vomiting. In both cases, diagnosis of meningococcal disease was confirmed by polymerase chain reaction (PCR) testing of cerebrospinal fluid (CSF). CSF cell counts were consistent with bacterial meningitis, but blood and CSF culture were negative, despite lack of prior antibiotic administration. The patients were in different school years. No obvious links, such as common classes, sporting teams, or mutual friends, could be found. However, they reported travelling on the same buses to and from school each day. These buses carry up to 78 students (53 seated and 25 standing) from the patients’ school and other nearby schools. The patients reported that the buses were usually crowded. Chemoprophylaxis was provided 8–9 days after symptom onset in Patient 2 to 132 students who claimed to have travelled on these buses during the exposure period. The two patients recovered fully and returned to school. No subsequent cases of meningococcal disease have occurred at the school. In the absence of meningococcal isolates, porA/porB genotyping2,3 was conducted on the meningococcal DNA contained in the CSF samples, and yielded identical sequences. While genotyping is not routinely conducted, porA/porB sequence nomenclature can be equated with serotype/subserotype (phenotypic expression of porA/porB). All 38 serogroup B isolates to date in NSW in 2005 have had serotyping/subserotyping performed, with no other cases having a serotype/subserotype equivalent to that of the two cases reported here. This supports the school bus as the most likely setting of transmission, given the lack of mutual friends and activities identified between the two cases. Asymptomatic nasopharyngeal carriage of meningococci is common, with about 10% of individuals being carriers at any one time.4 While crowding and close contact increase transmission of meningococci, factors leading to invasive disease are poorly understood.5 Antibiotic chemoprophylaxis is given to close contacts to eradicate nasopharyngeal carriage and limit disease spread.5 However, the absence of further cases among the 132 fellow travellers in this setting does not provide evidence of effectiveness of the chemoprophylaxis, given that the secondary attack rate, even in close household contacts, has been estimated at 2–4 per 1000 people.6 Current Australian guidelines recommend that, in school-based outbreaks, chemoprophylaxis be considered for a wider group than solely close contacts of a household nature.7 This report provides evidence to support chemoprophylaxis in similar circumstances, where linked cases are identified in school bus co-travellers, and no other, more specific natural grouping makes epidemiological sense.

Frank H Beard · Jeremy M McAnulty · John W Tapsall · Angelo M Zaia

Ami Schattner,* Noa Ben-Baruch† * Associate Professor of Medicine, † Head, Breast Cancer Unit, Hebrew University Hadassah Medical School, Kaplan Medical Centre, Bilu Junction, Rehovot, 76100, Israel. amiMDATclalit.org.il To the Editor: We describe a patient with malignant ascites and bacterial peritonitis, with no apparent intra-abdominal source of infection. A 56-year-old woman with advanced breast cancer was admitted to hospital with a 10-day history of diffuse abdominal pain, umbilical tenderness, and increasing abdominal girth. Breast cancer had been diagnosed 13 years before admission (T2 N0 M0) and treated with radical mastectomy and adjuvant chemotherapy. Five years before admission, disease had recurred at the chest wall, and 2 years before, liver metastases were found. Chemotherapy had failed, and her only current treatment comprised capecitabine and analgesics. On admission, the patient had tachycardia (110 beats per min), tachypnoea (26 breaths per min), blood pressure of 115/80 mmHg, and no fever or signs of sepsis. Examination showed evidence of local recurrence around the left mastectomy scar, and marked ascites with discoloration, warmth, tenderness and a diffuse infiltration over the umbilicus. There were no signs of peritoneal irritation, nor leg oedema. Blood test results were in the normal range, except for mild leukocytosis (10.6 × 109 cells/L; reference range [RR], 3.8–9.8 × 109 cells/L), with left shift (86% neutrophils, and 6% lymphocytes; RR, up to 67% neutrophils, and at least 32% lymphocytes). X-ray did not show abdominal free air. Abdominal computed tomography revealed ascites, retroperitoneal lymph nodes, omental and umbilical infiltration, and the known liver metastases without signs of portal hypertension. Peritoneal tap removed 2800 mL of fluid. Examination of the fluid revealed a white blood cell count of 2 × 109 cells/L (80% neutrophils), glucose concentration of 790 g/L, and albumin concentration of 22 g/L (serum–ascites albumin gradient, 0.7; a gradient < 1.1 indicates that the patient does not have portal hypertension). Gram stain of the ascitic fluid revealed gram-positive cocci, and culture identified Staphylococcus aureus. Cytological examination showed sheets of atypical enlarged epithelial cells highly suggestive of malignancy. Therapy was begun with parenteral cloxacillin and ciprofloxacin, but the patient died shortly after. This patient’s condition is unlikely to have been spontaneous (“primary”) bacterial peritonitis, which usually occurs in patients with liver cirrhosis, portal hypertension and ascites with high serum–ascites albumin gradient,1 and occasionally in patients with malignant ascites.2 However, in our patient, infection may have originated from the umbilical metastasis. The umbilicus has a direct connection with the intra-abdominal cavity and is also connected to a number of intra-abdominal organs via embryological remnants, such as the umbilical vein and the urachus. Both routes may be involved in the spread of malignant tumour to the umbilicus,3 but spread may also be in the opposite direction, causing peritonitis. This diagnosis may be missed by physicians who may not notice anything new in a patient with longstanding malignant ascites, and may fail to consider bacterial peritonitis, a treatable condition.

Ami Schattner · Noa Ben-Baruch

In view of the possibility of a human pandemic of avian influenza, a first-line strategy for many countries is stockpiling of antiviral neuraminidase inhibitors (oseltamivir [Tamiflu] and zanamivir [Relenza]), which can reduce mortality, morbidity and influenza transmission. However, global supply of the antivirals is controlled by the European-based patent owners, Roche and GlaxoSmithKline. This prevents competition in the manufacturing and distribution of antivirals and has reduced global supply capacity and affordability. The Australian Government has acknowledged that, in the event of a pandemic, its own stockpile of antivirals will be limited and reserved for those on a confidential rationing list. Pharmacies are running out of stocks, limiting opportunities for individuals to secure supplies privately. Compulsory licensing provisions, permitted under domestic patent law, would allow Australian generic manufacturers to start producing antivirals locally or import them from generic producers at affordable prices. Australia also has an opportunity and a responsibility to promote compulsory licensing and generic antiviral production in the Asian region, to ensure our neighbours can establish pandemic stockpiles in a timely and affordable manner.

Buddhima Lokuge MB BS, MPH · Peter Drahos GDLP, LLM, PhD · Warwick Neville BA, LLB, STD

Objective: To describe empiric community-acquired pneumonia (CAP) management in Australian hospital emergency departments (EDs) and evaluate this against national guidelines, including use of the pneumonia severity index and antibiotic selection.Design: A multicentre, cross-sectional, retrospective audit, April 2003 to February 2005.Setting: 37 Australian hospitals: 22 principal referral hospitals, six large major city hospitals, four large regional hospitals, four medium hospitals and one private hospital.Participants: Adult patients with a diagnosis of CAP made in the ED. Data on 20 consecutive CAP ED presentations were collected in participating hospitals.Outcome measures: Documented use of the pneumonia severity index, initial antibiotic therapy prescribed in the ED, average length of stay, inpatient mortality, and concordance with national guidelines.Results: 691 CAP presentations were included. Pneumonia severity index use was documented in 5% of cases. Antibiotic therapy covering common bacterial causes of CAP was prescribed in 67% of presentations, although overall concordance with national guidelines was 18%. Antibiotic prescribing was discordant due to inadequate empiric antimicrobial cover, allergy status (including contraindication to penicillin), inappropriate route of administration and/or inappropriate antibiotic choice according to recommendations. There was no significant difference between concordant and discordant antibiotic prescribing episodes in average length of stay (5.0 v 5.7 days; P = 0.22) or inpatient mortality (1.6% v 4.1%; χ2 = 1.82; P = 0.18).Conclusions: Antibiotic therapy for CAP prescribed in Australian EDs varied. Concordance with national CAP guidelines was generally low. Targeted interventions are required to improve concordance.

for the CAPTION Study Group*

Clinical record A 29-year-old Vietnamese woman who had migrated to Australia 5 years previously was referred with a 2-month history of erythematous painful nodules on her thighs and left buttock, the largest of which measured 3 cm × 4 cm. There was no history of recent infections or previous exposure to tuberculosis. She denied any symptoms of inflammatory bowel disease, vasculitis or connective tissue disease. Examination disclosed no abnormalities such as lymphadenopathy, hepatosplenomegaly or synovitis. A chest x-ray was normal. Blood tests showed a total white cell count of 7.3 × 109/L, haemoglobin level of 133 g/L, platelet count of 295 × 109/L and an erythrocyte sedimentation rate of 11 mm/h. No antinuclear antibodies or antineutrophil cytoplasmic antibodies were detected. Liver and thyroid function tests and coagulation studies were all normal. A working diagnosis of erythema nodosum was made, and the patient was commenced on naproxen. When treatment with naproxen was unsuccessful after 4 weeks, she was commenced on 45 mg (1 mg/kg) of prednisolone daily. As she showed only a partial response to prednisolone, a deep incisional biopsy of a lesion was performed. It revealed a granulomatous lobular panniculitis with necrosis (Box), consistent with erythema induratum (EI). There were no visible mycobacteria or fungi on staining. A subsequent Mantoux test was strongly positive, while a polymerase chain reaction test for Mycobacterium tuberculosis DNA performed on the paraffin-embedded tissue was negative. The patient was commenced on a 6-month course of quadruple antituberculous directly observed therapy, consisting of rifampicin, isoniazid, pyrazinamide and ethambutol, together with pyridoxine, while the prednisolone dose was quickly tapered and ceased. Within 2 months, the cutaneous lesions had resolved, and had not recurred a month after the completion of antituberculous therapy. Bazin1 first described erythema induratum (EI) in 1861 to illustrate chronic, painful, violaceous, indurated and occasionally ulcerative nodules that occurred predominantly on the lower limbs of patients (mainly women) with tuberculin hypersensitivity. Its appearance can easily be confused with erythema nodosum, polyarteritis nodosa or perniosis. The diagnosis of EI could easily have been missed, as the cutaneous lesions were similar to those characteristic of erythema nodosum. In addition, there was a partial response to corticosteroids. A diagnosis of EI is based on cutaneous characteristics, a positive Mantoux test, evidence of tuberculosis and histological findings.2 The diagnosis of EI in this patient was established with three of the four criteria. Furthermore, the patient responded clinically to antituberculous therapy. The negative polymerase chain reaction (PCR) result is not surprising. In other studies, only 56%–88% of patients previously diagnosed with cutaneous tuberculosis had a positive PCR result using the IS6110 primer.2,3 Schneider et al4 found a positive PCR result in only 5 of 20 patients with EI. Shimizu et al5 failed to isolate Mycobacterium tuberculosis by either culturing cutaneous EI tissue or inoculating the tissue into guinea pigs. The association of EI with tuberculosis is still a matter of debate. Some have considered EI to be a type III or type IV hypersensitivity reaction to M. tuberculosis antigens.6 This may explain the patient’s partial response to corticosteroids. Lessons from practice Erythema induratum (EI) can easily be mistaken for erythema nodosum, perniosis or cutaneous polyarteritis nodosa. The failure to detect Mycobacterium tuberculosis on microscopy or by a polymerase chain reaction test on the biopsy specimen does not exclude the diagnosis. EI lesions can resolve with corticosteroid therapy, which may lead to a misdiagnosis of erythema nodosum. EI should be treated with antituberculous therapy. Histological section of subcutaneous tissue from one of the nodules The lower-power image shows florid inflammatory changes involving fat lobules (L) with thickening of lobular septa (S), consistent with a florid, predominantly lobular, panniculitis. The higher-power image (inset) shows a classical combination of granulomatous inflammation (G) and eosinophilic granular fat necrosis (N).

Gary Y J Chew MB BS, BSc(Med) · John W Quin PhD, FRACP, FRCPA · Christopher Henderson MB BS, FRCPA

Benjamin P Howden,* Elizabeth A Grabsch,† Helen King,‡ Maryza Graham,§ Paul D R Johnson,¶ Tony M Korman** M Lindsay Grayson,†† Barrie C Mayall‡‡ * Infectious Diseases Physician, § Infectious Diseases Registrar, ¶ Deputy Director, †† Director, Infectious Diseases Department, Austin Health, Studley Road, Heidelberg, VIC 3084; † Infection Control Scientist, ‡ Scientist, ‡‡ Medical Microbiologist, Microbiology Department, Austin Health, Melbourne, VIC; ** Director, Infectious Diseases Department, Southern Health, Melbourne, VIC. benjamin.howdenATaustin.org.au To the Editor: Although community-acquired methicillin-resistant Staphylococcus aureus (CA-MRSA) has been reported, it is not recognised as important in Victoria.1 Also, health-care acquisition of “typical CA-MRSA” strains has been reported, but is uncommon.2 Because of an apparent increase in CA-MRSA at Austin Health (a tertiary referral centre in Melbourne encompassing The Austin Hospital, The Heidelberg Repatriation Hospital and Royal Talbot Rehabilitation Centre), we undertook a retrospective survey of cases to determine the frequency of isolation, and clinical and laboratory features of non-multiresistant MRSA (nmMRSA),1 as a marker for CA-MRSA. Patients with nmMRSA isolated from blood culture between January 2000 and December 2003 or from any specimen between March 2002 and August 2003 were included. Medical records were reviewed, and the mode of acquisition was defined as described previously.3 All isolates underwent polymerase chain reaction testing for genes encoding Panton–Valentine leukocidin, and pulsed field gel electrophoresis (PFGE). The frequency of isolation of nmMRSA was also reviewed for the period 1999–2004 at two major hospitals in Victoria — Austin Health and Southern Health. We identified 53 patients with nmMRSA infection or colonisation (Box). Thirteen cases (25%) were community acquired, including one fatal case of endocarditis, nine cases of soft tissue infection, two of bacteraemia, and one of bone/joint infection. Forty cases were health-care-acquired, including 17 cases of skin or soft tissue infection and seven of bacteraemia. Ten hospital patients had nmMRSA colonisation without infection. Patients with health-care-acquired nmMRSA were commonly in the renal or spinal unit or from nursing homes. PFGE revealed that the 53 nmMRSA isolates belonged to 18 groups, including six that have been previously described (WA-1 and 2, Queensland, SWP, UK EMRSA-15 and 16) and 12 novel groups (A-L). Health-care-acquired infection or colonisation was documented for all PFGE groups, including “typical CA-MRSA” strains (WA-1, WA-2, Queensland, and SWP). Twelve isolates (23%), including four health-care-acquired isolates, were positive for Panton–Valentine leukocidin (PVL). This is of particular concern given the association of PVL-positive S. aureus strains with severe skin disease, necrotising pneumonia, and high mortality in some studies, and the increased virulence of CA-MRSA compared with typical health-care-acquired multiresistant MRSA.4,5 Ten patients with CA-MRSA (77%) received ineffective antimicrobial therapy in the first 48 hours of treatment, including patients with bacteraemia. From 1999 to 2004, there was a significant decrease in the total number of MRSA cases per hospital discharge at Austin Health and Southern Health, but the number of nmMRSA cases increased, from 0.6 to 1.1 per 1000 discharges at Austin Health (1999 to 2004; P = 0.004) and from 0.08 to 0.35 per 1000 discharges at Southern Health (2001 to 2004; P < 0.001). Multiple genotypes of CA-MRSA are increasingly causing community-acquired and, importantly, health-care-acquired infections in Victoria. Clinicians need to be alert to this problem, and further research to understand the epidemiology and risk factors is urgently required to help guide changes in therapy and infection control policy. Characteristics of patients and isolates of non-multiresistant methicillin-resistant Staphylococcus aureus at a hospital in Victoria Characteristic Total Community-acquired Health-care- acquired Number of cases* 53 13 40 Age (years): median (range) 31 (17–58) 69 (18–94) Sex (M/F) 8/5 16/24 Infection type* Skin/soft tissue 26 9 17 Bacteraemia 9 2† 7 Endocarditis 1 1‡ 0 Bone/joint 4 1 3 Pneumonia 1 0 1 Urine 1 0 1 Pelvic abscess 1 0 1 Colonisation 10 0 10 PFGE group WA-1 4 2 2 WA-2 3 1 2 Queensland 6§ 5§ 1§ SWP 5§ 3§ 2§ UK EMRSA-15 10 0 10 UK EMRSA-16 3 2 1 A-L 22¶ 0 22¶ PFGE = pulsed field gel electrophoresis. * Survey included blood culture isolates for the period 2000–2003, and isolates from any specimen for the period March 2002–August 2003. † One case was associated with pneumonia, the other with epidural abscess. ‡ Fatal case. § All isolates were Panton–Valentine leukocidin (PVL)-positive. ¶ One isolate was PVL-positive.

Benjamin P Howden · Elizabeth A Grabsch · Helen King · Maryza Graham · Paul D R Johnson · Tony M Korman · M Lindsay Grayson · Barrie C Mayall

Christopher K Fairley,* Jane Hocking,† Jane Gunn,‡ Marcus Y Chen§ * Professor, † Postdoctoral Research Fellow, School of Population Health, ‡ Associate Professor and Deputy Head, Department of General Practice, § Registrar, Melbourne Sexual Health Centre, University of Melbourne, 580 Swanston Street, Carlton, VIC 3053. cfairleyATunimelb.edu.au To the Editor: Each year, Australian general practitioners test only about 7% of women aged 16–24 years for chlamydia (according to Medicare records, 89 132 tests were performed on women in this age group in 2004). This compares with 30% of women aged 20–24 years in Sweden1 or Denmark.2 Why are Australian GPs not testing young women for chlamydia? Chlamydia screening in sexually active women is cost-effective and significantly reduces complications such as infertility.3 The Australian Government’s commitment to chlamydia control is evidenced by the $12.5 million recently provided for increased awareness, improved surveillance and a pilot testing program for chlamydia. In 2004, 81% of 15–19-year-old and 89% of 20–24-year-old women attended a GP at least once in Australia (unpublished data from Health Insurance Commission [HIC]), yet chlamydia testing was ordered for only a small fraction of these women. Possible reasons for the low testing rate include a lack of knowledge about the benefits of testing young women, inadequate support, difficulties in raising chlamydia testing during consultations not related to sexual health, or concern about violating HIC rules on screening.4 However, the Medicare Benefits Schedule does permit testing for chlamydia among young sexually active women. Benefits are payable for: health screening services . . . by the patient’s own medical practitioner . . . to ensure the patient receives any medical advice or treatment necessary to maintain his/her state of health . . . [B]enefits would be payable for the attendance and such tests which would be considered reasonably necessary according to the circumstances of the patient, such as age, physical condition, past personal and family history.5 If the HIC is concerned that “inappropriate practice” may have occurred, there is a clearly defined process for determining this.6 Part of the definition of “inappropriate practice” is “conduct in connection with rendering or initiating services that would be unacceptable to the general body of members of that profession”. A GP’s professional peers would be members of the Royal Australian College of General Practitioners. The College will soon release the latest edition of its Guidelines for preventive activities in general practice (“The Red Book”),7 which includes activities only if they are relevant to general practice and have a demonstrated benefit. The book now recommends that all sexually active women less than 25 years of age should be tested annually for chlamydia. Increasing chlamydia testing in young Australian women from the current low rate of 7% would reduce complications and save money in the long run.3

Christopher K Fairley · Jane Hocking · Jane Gunn · Marcus Y Chen

Benjamin C Cowie,* Alan Breschkin,† Heath Kelly‡ * Infectious Diseases Physician, † Senior Scientist, Infectious Disease Serology, ‡ Head of Epidemiology, Victorian Infectious Diseases Reference Laboratory, 10 Wreckyn Street, North Melbourne, VIC 3051. Benjamin. CowieATmh.org.au To the Editor: Hepatitis E virus (HEV) infection is uncommon in Australia. The HEV cases detected are almost always in patients who have recently arrived from HEV-endemic regions of the world.1 We previously reported a significant increase in highly reactive serology results for anti-HEV IgG antibodies measured by enzyme im-munoassay (EIA) at the Victorian Infectious Diseases Reference Laboratory (VIDRL) in the first 6 months of 2004.2 Nine of the 10 Victorian patients with highly reactive samples in the previous report had had a history of recent clinically compatible illness and travel in a disease-endemic region within the incubation period (2–9 weeks). This indicated a strong association between highly reactive anti-HEV IgG measured by EIA and acute HEV infection, as has been shown previously.3 At the time, we hypothesised an association with overseas HEV epidemics, particularly in India, as seven of the nine patients with acute HEV infection had travelled there. We have now reviewed HEV serology results at VIDRL for the subsequent 9 months and compared them with our ex-perience of the past 5 years (Box). In the first quarter of 2005, we recorded the highest quarterly number of highly reactive anti-HEV serology results since testing commenced at VIDRL. Also marked on the figure are the dates, over the same time period, when an outbreak of hepatitis in India (either suspected or confirmed to be caused by HEV) was reported on ProMED-mail, the global electronic reporting program for emerging diseases hosted by the International Society for Infectious Diseases (http://www.promedmail.org). It would appear that epidemic HEV activity in India is reflected in significant increases in the number of highly reactive anti-HEV serology results in our laboratory. In fact, as shown in the Box, increases in highly reactive anti-HEV serology at VIDRL have sometimes preceded an outbreak noti-fication on ProMED-mail, and may provide early warning of such an event. A similar association is not observed for epidemics in other countries. Travellers to developing countries must be advised of preventive measures against HEV and other enterically transmitted diseases, and a diagnosis of HEV infection should be considered in any febrile traveller recently arrived from an HEV-endemic area, particularly if jaundice or abnormal liver function tests are present. This is especially important in pregnant women because of the risk of fulminant hepatitis, with maternal mortality in excess of 20% in the third trimester.4 All cases should be notified to state health authorities. Highly reactive anti-HEV IgG EIA results at VIDRL per quarter, 1 Apr 2000 to 31 Mar 2005. Also marked are confirmed (in bold) or suspected epidemics of HEV in India listed on ProMED-mail* during the same period EIA = enzyme immunoassay. HEV = hepatitis E virus. VIDRL = Victorian Infectious Diseases Reference Laboratory. * Available at <http://www.promedmail.org>.

Benjamin C Cowie · Alan Breschkin · Heath Kelly

Noelene S O'Keefe,* Kristina A Heinrich-Morrison,† Bruce McLaren‡ * Project Officer, Legionella Program, Environmental Health, † Public Health Nurse, ‡ Medical Officer, Communicable Diseases Section, Department of Human Services, 17/120 Spencer Street, Melbourne, VIC 3000. bruce.mclarenATdhs.vic.gov.au To the Editor: A 23-year-old man presented to a Victorian hospital with a 4-day history of fever, rigors, confusion and malaise. A chest x-ray showed left lower-lobe pneumonia, and Legionella pneumophila serogroup 1 antigen was detected in his urine. No respiratory specimens were obtained. He recovered completely after treatment for community-acquired pneumonia, including intravenous ampicillin and oral roxithromycin, and returned to work 16 days after onset. Investigations for the source of the infection included environmental review and sampling of cooling towers near his workplace, home, and other sites visited during the incubation period. Active workplace surveillance prompted testing for and detection of L. pneumophila serogroup 1 urinary antigen in a second employee, a 53-year-old man who had presented 2 days earlier than the patient above to another Victorian hospital with fever, abdominal pain and diarrhoea. Legionellosis was not suspected on presentation. He had no symptoms, signs or radiological evidence of pneumonia. Treatment, including intravenous ampicillin and oral roxithromycin, began when the antigen result was obtained, and he was discharged after 10 days in hospital, although he felt unwell for 2 or 3 weeks after discharge. The two men worked near each other in a welding area. A water tank was placed at the entrance to the area, with the cover left open. This acted as a heat exchange for the welding cooling system. A high count of L. pneumophila serogroup 1 (1300 colony-forming units/mL) was grown from a sample of this water. It was common on hot days to cool the work place with an industrial fan. The open water tank was between the fan and the two employees during the incubation period. No L. pneumophila isolates were found in any linked cooling towers. Remedial action included commencing a disinfection program for the water reservoir, and a request to fit the cover correctly and move the fan. No further cases were detected. Because no clinical isolates were obtained, a direct subtype match between clinical and environmental specimens was not possible. Urine antigens are considered definitive laboratory tests given a compat-ible illness (fever or cough or pneumonia).1 Outbreaks of Legionnaire’s disease and Pontiac fever (legionellosis without pneumonia) with industrial sources other than cooling towers have been reported.2,3 This outbreak demonstrates that a simple change in the environment (adding a fan) and an apparently low-risk source (a warm water bath) have the potential to give rise to significant disease. It also shows the value of active workplace surveillance after a single case.

Noelene S O'Keefe · Kristina A Heinrich-Morrison · Bruce McLaren

Objective: To estimate the risk of death from healthcare-associated (nosocomial) bacteraemia caused by methicillin-resistant Staphylococcus aureus (MRSA), and compare it with that of nosocomial bacteraemia caused by methicillin-sensitive S. aureus (MSSA), by meta-analysis of selected studies. Data sources: Medline, EMBASE, Current Contents and Cochrane Library were searched for the period January 1978 (or earliest date of the database, if later than 1978) to December 2000. Study selection: Studies which compared mortality of nosocomial MRSA and MSSA bateraemia. Data synthesis: Nine studies were analysed. All but one found an increased relative risk (RR) of death from MRSA bacteraemia, with RR ranging from 0.89 to 4.94. Meta-analysis showed that patients with MRSA bacteraemia have an RR of death, compared with patients with MSSA bacteraemia, of 2.21 (95% CI, 1.76–2.57) using the fixed-effect method, and 2.03 (95% CI, 1.55–2.65) using the random-effect method. Conclusion: MRSA bacteraemia is associated with a real increase in risk of death, further justifying ongoing MRSA surveillance and control in healthcare facilities.

Michael Whitby · Mary-Louise McLaws · Geoffrey Berry

Objective: To determine immunity to tetanus and diphtheria in the Australian population.Design and setting: Analysis, using double antigen enzyme immunoassays, of a representative sample of sera (1950 samples tested for diphtheria and 2884 for tetanus) collected opportunistically from Australian laboratories between July 1996 and May 1999.Main outcome measure: Immunity to diphtheria and tetanus, defined as negative (susceptible) when the antitoxin level was < 0.01 IU/mL, positive (immune) when it was ≥ 0.1 IU/mL, and low positive (partially immune) when it was in the range 0.01–< 0.1 IU/mL.Results: About 99% of children aged 5–9 years had diphtheria and tetanus antitoxin levels ≥ 0.01 IU/mL (immune or partially immune). Antitoxin levels declined with age and generally more markedly for diphtheria than tetanus. For subjects aged 50 years and over, less than 60% were immune or partially immune to diphtheria and less than 75% to tetanus. Men and women had similar diphtheria antitoxin levels, while women had lower levels of tetanus antitoxin compared with men of the same age, with the difference being most marked in the age group ≥ 70 years (37% v 60%; P < 0.001).Conclusions: Immunity in children appears to be good, but adults, especially older people, may not be adequately protected. Recent changes to the Australian Standard Vaccination Schedule should improve immunity in cohorts now aged < 50 years. However, additional efforts are required to protect those over 50 years (especially travellers), who are most susceptible.

Heather F Gidding BAppSci, GradDipEpidBiostats, MAppEpid · Josephine L Backhouse BRurSc · Gwendolyn L Gilbert MD, FRACP, FRCPA · Margaret A Burgess MD, FRACP

Grant A Mackenzie Paediatrician and Postgraduate Student, Ear Health and Education Unit, Menzies School of Health Research, PO Box 41096, Darwin, NT 0811. grantmacATmenzies.edu.au To the Editor: Funding of universal varicella zoster vaccine (VZV) at ages 18 months and 10–13 years was recently announced in Australia. Health professionals should be aware of a number of related issues. Varicella vaccination was recommended in the United States from 1996 for all children aged 12–18 months, with catch-up vaccination to age 13 years. US surveillance shows: Decreased varicella mortality in all age groups except those aged ≥ 50 years (average varicella deaths per year: 145 in 1990–1994 versus 66 in 1999–2001).1 Decreased varicella cases in all age groups, with a non-significant reduction in hospitalisations (average annual hospitalisations in three surveillance regions: 40 before vaccination versus 14 after vaccination).2 US data on herpes zoster have not yet been published. There are concerns that, in the longer term, universal varicella vaccination may increase the incidence of adult varicella and herpes zoster, similar to the effect of pertussis vaccination on adult pertussis. Modelling in the United Kingdom predicted that universal infant vaccination would initially reduce varicella, but would result in increases in herpes zoster 5–10 years later and adult varicella 20–40 years later.3 In contrast, modelling of adolescent vaccination predicted a small decrease in varicella, but no increase in later adult varicella.3 Varicella is generally perceived as a mild illness, while vaccination is largely valued for preventing serious, life-threatening conditions. Anecdotal reports of low levels of private purchase of VZV in Australia suggest it may not be a priority for some families. With 36% of general practitioners concerned about unknown side effects of VZV,4 and public concern about vaccine adverse events in the face of low disease rates, the level of acceptance of universal varicella vaccination by providers and consumers is uncertain. Alternatives to universal varicella vaccination were a high-risk strategy (vaccination of children with chronic illness and family members of high-risk individuals) or waiting until US disease patterns were established. These were real options as: A high-risk strategy may prevent up to 45% of paediatric hospitalisations.5 Hospitalisation and herpes zoster contribute more to health costs than treatment in the community or acute varicella.5 Natural infection, at the cost of disease, immunises most of the population. Any increase in adult varicella and herpes zoster caused by varicella vaccination may be alleviated by booster doses, but the added cost, difficulty in reaching the target population, and potential impact on community confidence in vaccination may be significant problems. The universal varicella vaccination program will test providers’ and consumers’ acceptance of vaccination against what is perceived as a mild illness.

Grant A Mackenzie