Journal Club 2014

Chickenpox and risk of stroke: a self-controlled case series analysis

Thomas SL, Minassian C, Ganesan V, Langan SM, Smeeth L. Clinical Infectious Diseases 2014; 58(1):61–8.

Langan SM, Minassian C, Smeeth L, Thomas SL. Clinical Infectious Diseases 2014; 58(11):1497–503

Background: Varicella is associated with central nervous system (CNS) complications, including acute ischaemic stroke (AIS) following chicken-pox. In one case series, chicken-pox occurred in 31% of 70 cases of AIS in the preceding 12 months. In a case-control study involving 11 children with AIS, 7 (64%) had chickenpox in the preceding 9 months compared to 4 (9%) of 44 controls (OR 17.5 (95% CI, 2.8–126.1). In October 2014, Thomas et al published a large study using self-controlled case series (SCCS) analysis to test whether children are at increased risk of stroke (first/index AIS used) following chicken-pox . They analysed four large general practice databases in the UK resulting in data from >100 million person years of observation. SCCS analysis is a technique to determine the incidence ratio (analogous to relative risk) of rare events in which the association between an event (AIS) and the exposure of interest (varicella) in a defined ‘risk period’ (in this case 12 months) is determined by considering an individual as both a case of the disease (risk of event occurring within the observed ‘risk period’) and a control (risk of event occurring within the whole period of study observation excluding the risk period or ‘baseline period’). This, so-called, self-matching eliminates fixed confounding as a source of bias.

Main Findings: The meta-analysis showed children were at a four-fold increased risk of stroke in the 6 months after chicken-pox (IR = 4.07; 95% [CI], 1.96–8.45). In adults the association was less pronounced, (IR = 2.13; 95% [CI], 1.05–4.36). In both adults and children, the association ceased to be statistically significant in the 7-12 months after chicken pox. The same group have also now reported on an increased risk of AIS in adults following zoster that was reduced using anti-viral medication.

Take Home Message: This study confirms the association of childhood stroke with varicella, which is also biologically plausible as varicella is known to cause an inflammatory arteriopathy. The accompanying editorial proposes an analysis of the effects of varicella vaccine on the incidence of childhood stroke. This should be achievable in those countries where primary childhood varicella immunisation is now in place. The question remains if any treatment should be should be given to prevent stroke in a child who has chicken-pox?

Submitted by: Philip Britton, Children’s Hospital at Westmead NSW, 15 December 2014.

Efficacy and safety of tenofovir disoproxil fumarate in pregnancy to prevent perinatal transmission of hepatitis B virus

Greenup AJ, Tan PK, Nguyen V, Glass A, Davison S, Chatterjee U, Holdaway S, Samarasinghe D, Jackson K, Locarnini S, Levy M. Journal of Hepatology 2014 ; 61: 502–507. PMID24801414.

Background: Globally, perinatal transmission accounts for around one third of all chronic hepatitis B infections. The vast majority of vertical hepatitis B transmission occurs during labour and delivery but in-utero transmission has also been reported. E antigen positivity and high maternal viral load are recognised risk factors for vertical transmission. In recent years antiviral therapy has been used during the third trimester of pregnancy to reduce vertical transmission in high-risk women. Tenofovir, with its high resistance barrier, potency, and established profile in the setting of HIV PMCT, is now being used for prevention of perinatal Hepatitis B, however there is very little published literature regarding its safety and effectiveness. This observational study describes outcomes of 130 pregnancies in 120 pregnant women with highly viraemic hepatitis B at a single Australian centre, from 2007 onwards.

Main findings: 58 women received tenofovir, 52 received lamivudine and 20 had no antiviral treatment in the third trimester. 4 of 58 did not tolerate tenofovir due to GI side-effects, whereas lamivudine was well-tolerated by all 52. Virologic response was better with tenofovir, where the mean VL fell from 7.98 log copies to 4.49 log copies at birth. In the lamivudine group mean viral load fell to 5.6 log copies (P=0.01). Follow-up data was available for 95 of 130 infants (73%). 1 of 44 in the tenofovir group acquired hepatitis B, despite good maternal adherence and virologic response. None of the 43 in the lamivudine group was infected. 2 babies of 8 untreated mothers were infected. One lamivudine-exposed infant died of SIDS at 3 months of age. There were no significant differences in obstetric or infant outcomes.

Summary: Both tenofovir and lamivudine in the third trimester reduced perinatal hepatitis B transmission compared to no maternal antiviral treatment. Tenofovir in the third trimester appears to be safe, produces a superior virologic response compared to lamivudine, and is an attractive option due to its potency and high resistance barrier, however GI side-effects may be problematic.

Submitted by: Rachael Webb, Starship Children’s Hospital NZ, 18 November 2014.

Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data

Van Boeckel T, Gandra S, Ashok A et al. Lancet Infect Dis. July 10, 2014; 14:742-50 PMID 25022435

Background: This is the first study since 1987 to ambitiously map and compare the evolution of antibiotic consumption on a global scale, involving 71 countries.

Main Findings: Between 2000 and 2010 antibiotic consumption increased by 36%. India was the largest consumer of antibiotics in 2010 (12.9 X 10^9units), China was the second largest (10 x 10^9 units) and USA the third largest (6.8 x 10^9 units). The study used Standard Units rather than the preferred Defined Daily Dose (DDD) as there were some gaps in global data available for DDD. Countries experiencing population and economic growth; Brazil, Russia, India, China and South Africa accounted for more than three-quarters of the rise in antibiotic use. Antibiotic consumption was stable or had moderately decreased in high-income countries with two exceptions: consumption substantially increased in Australia and New Zealand.

Take home message: These findings are consistent with surveillance data for antibiotic use in Australia and New Zealand and reveal high antibiotic consumption when compared with other developed European countries. The precise reason for this remains unclear and further investigation and funding for antimicrobial stewardship programs are warranted. Although antibiotic resistance is a major public health threat, far more deaths occur due to poor access to antibiotics in low-income and lower-middle income countries and addressing this issue is also of paramount importance.

Submitted by: Anita Campbell, Women’s and Children’s Hospital, SA, 3 November 2014.

A Prospective, International Cohort Study of Invasive Mold Infections in Children

Wattier RL, Dvorak CC, Hoffman JA et al. Journal of the Pediatric Infectious Diseases Society Advance Access, July 16, 2014. (No PMID available yet)

Background: Invasive mold infections (IMIs) are a major cause of morbidity and mortality in immunocompromised children. This prospective study from the International Pediatric Fungal Network is the first to describe the epidemiology, treatment and outcomes of an international cohort of children with proven and probable IMIs.

Main Findings: 131 children with IMI from 22 sites were enrolled (12 US and 10 international) over 4 years, median 3 (IQR 1 – 10) patients per site. Invasive aspergillosis (IA) accounted for 75% of all infections, with mucormycosis (13%) and other IMIs making up the remainder. Risk factors including neutropenia (69%), corticosteroid therapy (39%), other immunosuppression (27%) and GVHD (8%) were similar when comparing IA and non-IA IMIs. Patients with non-aspergillus IMIs were more likely to have received a mold active antifungal agent compared with patients with IA (67% vs 40% p=0.009). Voriconazole (34%) and polyenes (31%) were the most frequently used agents in initial monotherapy. Combination therapy was used initially in 25% of patients and by 12 weeks after diagnosis, 53% of patients had received a period of combination therapy. There was no significant association between combination therapy and treatment outcome (OR 0.93 [0.44-1.97]) and children receiving combination therapy were more likely to experience adverse effects. Overall mortality was 30% at 12 weeks (31% IA vs 35% mucormycosis p=0.623). In a multivariate logistic regression, preceding corticosteroid therapy was associated with treatment failure (OR 2.73 [1.17-6.37]) whilst surprisingly, neutropenia at diagnosis was not (OR 0.48 [0.25-0.93])

Take Home Message: This multi-centre, international study confirms findings in children with IMI to be similar to adult cohorts. These include prior exposure to mold-active agents as a risk factor for non-aspergillus IMI, and the association of previous steroid therapy with treatment failure. More than half the cohort received combination therapy, without evidence of treatment success but with more adverse events. Mortality is high.

Submitted by: Daniel Yeoh, Asha Bowen, Princess Margaret Hospital WA, 15 Oct 2014

Short Course oral co-trimoxazole vs. IM benzathine benzylpenicillin for impetigo in a highly endemic region: an open-label, randomized, controlled, non-inferiority trial

Bowen AC, Tong SYC, Andrews RM et al. The Lancet, Aug 26 2014 [Epub ahead of print] PMID25172376

Background: Impetigo is often associated with mild skin infections, but can also present with severe, recurrent skin sores that may be complicated by invasive bacterial infections or post-streptococcal inflammatory conditions such as rheumatic heart disease (RHD) and post-streptococcal glomerulonephritis (PSGN). There is a heavy burden of impetigo and its complications in Australian Indigenous children. Intramuscular benzathine penicillin has been used in tropical settings for treatment of impetigo since the 1970s, but administration is painful and benzathine penicillin provides inadequate cover against S. aureus. Co-trimoxazole has traditionally been considered ineffective in treatment of infections caused by S. pyogenes.

Main Findings: Five hundred and eight children from remote NT communities with purulent or non-bullous impetigo were randomly assigned (1:1:1) to receive benzathine penicillin, twice daily co-trimoxazole for 3 days, or once daily co-trimoxazole for 5 days. Treatment success was determined at day 7 based on digital image scoring by independent reviewers who were masked to treatment allocation. Treatment was successful in 85% of children with no significant difference between treatment groups, absolute difference 0.5% (95% CI -6.2 to 7.3%). Baseline swabs grew S. aureus in 81% and S. pyogenes in 90% of cases. Co-trimoxazole was more effective than benzathine penicillin at reducing S. aureus culture positivity on follow-up swabs (20% vs 52%), and both were equally effective in clearing S. pyogenes (7% at follow-up).

Take home message: This important study, published in the Lancet, demonstrates that short-course co-trimoxazole is comparable to benzathine penicillin for microbiological clearance of S. pyogenes and resolution of skin sores in children. Clinical improvement is associated with eradication of S. pyogenes but not S. aureus suggesting a predominant role for S. pyogenes in the pathogenesis of impetigo. Whether or not short-course co-trimoxazole effectively prevents RHD and PSGN is not answered by this study, but given its success in eradicating S. pyogenes, it seems likely that it is a satisfactory alternative to benzathine penicillin – an attractive alternative given that it is cheap, easy to administer and rarely associated with side effects.

Submitted by: Charlie McLeod, Josh Francis, Royal Darwin Hospital NT, 15 September 2014.

Immunising pregnant women against influenza to protect their infants

Madhi SA et al. N. Engl. J. Med. 2014; 371: 918-31. PMID 25184864

Because they are relatively immunocompromised, pregnant women are at risk of severe influenza from the second trimester onwards. It is recommended to immunise pregnant women, although vaccine efficacy in pregnancy is unknown. In a Bangladesh randomised controlled trial (RCT), immunising pregnant women against influenza with trivalent inactivated vaccine (TIV) reduced the incidence of influenza in their infants in the first 6 months of life by 63% (95% CI 5 to 85%) (Zaman NEJM 2008, 359:1555). Two RCTs from South Africa examined TIV efficacy for mothers and their babies in 194 pregnant women with HIV and in 2116 pregnant women without HIV infection for 24 weeks after birth. In HIV-infected mothers, the attack rate for influenza confirmed by polymerase chain reaction (PCR) was 17.0% with placebo and 7.0% in vaccine recipients, giving a vaccine efficacy of 57.7% (95% CI 0.2 to 82.1%). The vaccine reduced influenza by 26.7% in the infants of HIV-infected mothers, which was not statistically significant. In both HIV-uninfected placebo mothers and their infants the influenza attack rate was 3.6%. The incidence was reduced to 1.8% and 1.9% respectively in HIV-uninfected vaccine recipients, giving a vaccine efficacy of 50.4% (95% CI 14.5 to 71.2%) for mothers and 48.8% (95% CI 11.6 to 70.4%) for their infants. Thus TIV approximately halves the incidence of influenza in HIV-infected pregnant women and in HIV-uninfected pregnant women and their infants.

Submitted by: David Isaacs, Westmead Children's Hospital, NSW,10 September 2014.

Gut Microbiome in SIDS: a new understanding

Highet AR, Berry AM, Bettelheim KA, Goldwater PN. Int J of Med Micro 2014; 304: 735–741. PMID 24951305.

Background: Until now and despite many decades of SIDS research, there has been no congruent and plausible patho-mechanism for SIDS. Because of the importance of the gut microbiome in the development of the infant immune system and in maintaining the physiological integrity of the gut, Highet et al (2014) have conducted an investigation into the gut microbiome of SIDS and live babies. This paper extends the infection hypothesis proposed that a brief bacteraemia proceeds death. The bacteraemia could result from defective gut wall integrity or defences. An altered or “abnormal” gut microbiome may influence gut defences or cause possible inflammation and therefore could affect translocation of pathogens such as S. aureus and certain pathotypes of E. coli resident in the infant gut. Extraintestinal and uropathogenic E. coli pathotypes are over-represented in SIDS cases, as is S. aureus.

Main Findings: The paper showed that the gut microbiome of SIDS babies differs from that of normal babies. The overarching finding was of significantly more babies dying prone with S. aureus colonisation than babies similarly colonised but dying in other positions (supine/side) (odds ratio = 20·25; CI = 3·60-134·92). This supports the hypothesis that prone sleeping increases the risk of colonisation and therefore provides a congruent explanation for the higher risk of prone sleep position. Moreover, there was a highly significant association between proneness and S. aureus in both gut and a sterile site (OR = ∞; CI = 2·04 - ∞).

Take home message: The microbiome findings suggest a possibility exists for trials of a prevention program to provide the “correct” gut microbiome to babies early in life (particularly prems and those delivered by caesarean section). If sustained colonisation is achievable, this would be a simple and relatively cheap and novel approach to help eliminate this tragic and enigmatic condition.

Submitted by: Paul Goldwater, Women’s and Children’s Hospital, SA, 8 August 2014.

Spinal Rod Infections: a retrospective review

A retrospective case series from a single US institution identified 486 patients who had posterior spinal fusion with insertion of rods over 66 months. The rods became infected in 27 (5.6%), but only 23 had adequate data for analysis. Of these 23, aged 8-20 years (average 14.8 years), the commonest presentation was wound drainage. Most infections occurred about 2 weeks post-surgery (median 16 days) but the range was 8 to 1052 days. No single organism was predominant. Among Gram positive organisms, there were 8 cases of S. aureus, of which 3 were methicillin-resistant; the Gram negatives were Enterobacter (5), Pseudomonas (4), Proteus (2), E. coli (1) and Serratia (1). Six patients grew more than one organism. Two patients grew Mycobacterium abscessus. Infection was cured in 78% with antibiotics alone without removing the rods, but 4 patients failed medical therapy. Antibiotics were given for a median of 131 days (range 42->597 days). It appears most spinal rod infections can be cured with antibiotics alone with preservation of the implant, although the optimal duration of antibiotics is unknown.
Submitted by: David Isaacs, Westmead Children's Hospital, NSW, 5 August 2014.

Invasive Pneumococcal Disease in Children Can Reveal a Primary Immunodeficiency.

Gaschingnard J, Levy C, Chrabieh M et al. Clin Infect Dis. 2014;59(2):244-51. PMID 24759830

Background: Streptococcus pneumoniae results in a clinical spectrum ranging from asymptomatic carriage to invasive pneumococcal disease; the latter has a mortality rate of about 10%. Primary immunodeficiencies predispose to invasive pneumococcal disease, but the prevalence of significant immune dysfunction has previously never been evaluated systemically.

Main findings: 163 children with confirmed invasive pneumococcal disease were enrolled between 2005 and 2011 from 28 paediatric wards throughout France. PCV7 and PCV13 were included in the French national immunisation schedule from 2005-9 and 2010-11, respectively. Confirmed invasive pneumococcal disease was defined as positive pneumococcal culture, polymerase chain reaction result, and/or soluble antigen detection in a normally sterile site (blood, cerebrospinal fluid, synovial fluid, pleural fluid, but not sputum). Meningitis was the most common manifestation of invasive pneumococcal disease (87%), followed by pleuropneumonitis, isolated bloodstream infection, osteomyelitis, endocarditis and mastoiditis. An immunological assessment was performed on all of the children enrolled, including abdominal ultrasound, whole-blood counts and smears, determinations of plasma immunoglobulin and complement levels, and the evaluation of proinflammatory cytokines. 26 patients (16%) had abnormal immunological assessment and 17 (10%) were diagnosed with a primary immunodeficiency. Transient and primary antibody deficiencies (including two patients with XLA) were the most common finding, followed by complement deficiencies, asplenia and isolated disorders of the TLR signaling pathway. A primary immunodeficiency was more common in children >2 years (26% vs 3%, p < 0.001), in boys (even accounting for X-linked disorders), and in patients from consanguineous families (4/10 vs 13/152, p = 0.01). There were no significant differences in rates of primary immunodeficiency in children hospitalised for the first episode vs. recurrent episode of invasive pneumococcal disease (15/146 vs 2/16, p = 0.68).

Take home messages: Children with invasive pneumococcal disease should be considered for assessment of possible immunological abnormalities, particularly if > 2 years of age, male, or from a consanguineous family. Investigations may include abdominal ultrasound, WBC counts with blood smears, plasma Ig levels (and IgG subclasses for patients aged >2 years) and classic (and alternative) complement pathways analysis.

Submitted by: Freya Summons, David Burgner, Monash Children’s Hospital, VIC, 14 July 2014.

Beyond Malaria - Causes of Fever in Outpatient Tanzanian Children.

D'Acremont V, Kilowoko M, Kyungu E et al. N Engl J Med. 2014 Feb 27;370(9):809-17. PMID24897089

And letter: Osowicki J, Curtis N. Causes of Fever in Outpatient Tanzanian Children. The New England Journal of Medicine. 2014;370:2242-2244.PMID14897091

Background: With declining rates of malaria and improving access to diagnostics, empirical protocols for the management of febrile children in under-developed settings are being reconsidered. There are a paucity of high-quality data describing the microbial ecology and aetiology in acutely febrile African children.

Main findings: Febrile children 2 months to 10 years of age, without severe malnutrition or requirement for immediate lifesaving procedures, presenting to rural and urban Tanzanian clinics, were systematically and rigorously investigated with rapid diagnostics including molecular tests, serological studies, and culture of blood and nasopharyngeal specimens (25,743 tests). Final diagnoses were carefully attributed according to prespecified criteria and algorithms incorporating historical, clinical, radiological and laboratory findings. A cause of fever was ascribed in 96.8% of 1005 children, with multiple diagnoses in 22.6%. 70.5% of the children had viral disease, 22.0% bacterial disease, and 10.9% parasitic disease. Acute respiratory infection was present in 62.2%, malaria and gastroenteritis in 10% each, UTI in 5.9%, typhoid fever in 3.7%, skin/mucosal infection in 1.5% and meningitis in 0.2%. There were substantial differences between diagnoses made on the basis of predefined clinical and laboratory criteria, and pathogens identified but not necessarily implicated in the febrile episode.

Take home messages: This is an extraordinarily ambitious yet accessible study. The methodology is forward thinking, and includes prospective collection of data to address key concerns such as differentiating pathogens from passengers. The results reinforce the aphorism that common infections are common, even in sub-Saharan Africa, and question the role of routine empirical antibiotic and antimalarial treatment. The contribution of high fidelity new diagnostics in the study is appropriately placed in a clinical context, demonstrating the gulf between microbial ecology and aetiology, highlighting the importance of future research refining the application of these tests in clinical care across all settings.

Submitted by: Josh Osowicki, Royal Children's Hospital Melbourne, VIC, June 6th 2014.

A Randomized Trial of Hyperimmune Globulin to Prevent Congenital Cytomegalovirus.

Revello MG, Lazzarotto T, Guerra B et al on behalf of the CHIP Study Group. N Engl J Med. 2014 Apr 3;370(14):1316-26. PMID24693891

Background: Congenital CMV is a leading cause of neonatal morbidity and mortality, with approximately 0.6% of all newborns infected and about 20% of these symptomatic. Primary antenatal CMV infection results in around 30% CMV transmission. Previous non-randomised studies suggested CMV hyperimmune globulin showed promise for prevention of congenital CMV infection and disease.

Methods: This phase 2, randomized, placebo-controlled, double-blind study, involved 11 Italian centres. Pregnant women with primary CMV infection between 5-26 weeks gestation were eligible. Intravenous hyperimmune globulin or placebo (0.9% saline) was given every 4 weeks until 36 weeks of gestation, detection of CMV in the amniotic fluid (if amniocentesis done), or spontaneous termination of the pregnancy.

Results: Of 338 eligible women, 124 were randomised and 123 completed study. Baseline characteristics were similar between the 2 groups. The median number of infusions received by each woman did not differ significantly (hyperimmune globulin group 5 infusions and placebo group 4 infusions; P = 0.60). Intrauterine CMV transmission occurred in 45 of 123 cases (37%). Congenital infection was diagnosed before birth in 18 fetuses and at birth in 27 newborns. The rate of congenital infection in the hyperimmune globulin group was 30% (18 fetuses or infants of 61 women) and in the placebo group was 44% (27 fetuses or infants of 62 women) (P = 0.13). Symptoms at birth were documented in 3/10 (30%) hyperimmune globulin group and 4/17 (24%) placebo. Obstetric complications affected 13% in the hyperimmune group and 2% placebo, P=0.06.

Summary: This RCT did not show a significant reduction in the rate of transmission of CMV infection or decrease in symptomatic CMV at birth with antenatal hyperimmune globulin as compared with placebo, and there was a trend towards increased obstetric complications. Hyperimmune globulin cannot be recommended as a routine intervention at this time, although 2 other RCTs are currently underway and may provide further data.

Submitted by: Julia Clark, Royal Children's Hospital Brisbane, 14 May 2014.

Human Parechovirus Central Nervous System infections in Southern Californian children.

Felsenstein S, Yang S, Eubanks N, et al. Pediatr Infect J 2014;33:e87-e91 PMID: 24104958

Background: Human parechovirus (hPeV) infections have been associated with a wide variety of clinical presentations in children and it has been increasingly recognised as an important pathogen causing meningoencephalitis in young infants. CNS infections caused by HPeV in neonates have been associated with white matter lesions which suggest possibility of long-term sequelae. This study assessed prevalence, epidemiology and clinical presentation of CNS infections with hPeV and compared it to that of enterovirus (EV). This was done by examining paediatric CSF samples over a 5 year period (2008-2012) in a tertiary paediatric hospital in Los Angeles. Detection of hPeV was by RT-PCR which included types 1-6.

Main findings: This study included 440 CSF samples of which 12 (2.7%) were positive for hPeV. EV PCR was only performed in 402 specimen which had sufficient volume for testing, of which 43 (10.7%) were positive. All children tested positive for hPeV were ≤5yrs and 58% were <3 months. Median age was 3.7 months. 11/12 hPeV CNS infections occurred in even numbered years with no distinct seasonality. Neither maximum temperature nor duration of illness distinguished EV and hPeV infections. Seizures were much more common in children with hPeV compared with those positive for EV (41.7% vs 14%, p=0.05). Similarly so were URTI symptoms (58.3% vs 16.3%, p=0.01). Of note, prevalence of pre-existing seizure disorders and structural brain anomalies were higher in the hPeV positive group than in EV positive or the remaining cohort (25% vs 2.3% vs 5.8%, p=0.03). In terms of CSF parameters, only 25% of hPeV positive children had CSF pleocytosis whereas 82% EV positive patients had a CSF WBC above the age adjusted normal range (p=0.004). Neurologic imaging was performed in 75% of hPeV positive cases and none of them demonstrated acute abnormalities.

Take home messages: HPeV is an important cause of CNS infections in young infants, particularly those with neurological symptoms. It is not uncommon to have no CSF pleocytosis. It is difficult to comment on the significance of higher prevalence of pre-existing seizure disorders and structural brain anomalies in the hPeV group. It may be that hPeV has a propensity to cause neurologic symptoms such as seizures in vulnerable children with pre-existing neurologic morbidity, exceeding that of EV.

Submitted by: Sophie Wen, Children's Hospital Westmead, 11 April 2014.

Protective Association Between Rotavirus Vaccination and Childhood Seizures in the Year Following Vaccination in US children.

Payne D, Baggs J, Zerr D et al CID 2014:58(2);173-7. PMID 24265355

Background: This retrospective cohort study looked at the protective effect of rotavirus vaccination against seizures in infants. It used the Vaccine Safety Datalink, a collaborative project that collects data on approximately 3% of the US population annually for the purposes of vaccine safety research. Rotavirus vaccination status was obtained from managed care organisation clinical records, and the study compared fully vaccinated infants with those who had received at least one vaccine, but no rotavirus vaccine. The diagnosis of seizures came from emergency department discharge coding occurring 29 to 388 days following the last rotavirus vaccination.

Main findings: The study cohort included 250,601 children or whom 74.4% were fully vaccinated. Rotavirus vaccination provided significant protection against all seizures (RR = 0.790; 95% CI .714-.875) and against first ever seizures (RR = 0.816, CI .73-.91).

Take home messages: More than 99% of the infants in the study received the RV5 vaccine so the results cannot be generalised to RV1 vaccine. The authors outline that the results are biologically plausible given that rotavirus has been detected in CSF samples of children with diarrhoea and seizures and that one study reported seizures in 7% of children with laboratory confirmed rotavirus. Unexpected benefits of vaccination are great to hear about as they shift the focus away from vaccine adverse events.

Submitted by: Tony Walls, Christchurch, 20 March 2014.

Once Weekly Trimethoprim-Sulfamethoxazole for Prevention of PCP in Children with Cancer.

Caselli D et al, Single-day trimethoprim/sulfamethoxazole prophylaxis for pneumocystis pneumonia in children with cancer. J Pediatr. 2014 Feb;164(2):389-392.e1. doi: 10.1016/j.jpeds.2013.10.021. PMID: 24252793

Background: Before the use of trimethoprim-sulfamethoxazole prophylaxis, up to 43% of children with cancer developed Pneumocystis jiroveci pneumonia (PCP). Current recommendations for prophylaxis vary considerably, from daily dosing to thrice weekly dosing.

Main Findings: This multi-centre Italian prospective survey compared treatment regimens in 2466 children with newly diagnosed cancer from 2009 to 2011, including 1093 with solid tumor and 1373 with leukemia/lymphoma. Of these patients, 1371 received the 3-day/week prophylaxis regimen, 406 received the 2-day/week regimen, and 689 received the 1-day/week regimen (5-10mg/kg/day in 2 doses). Overall, only 2 cases of PCP (0.08%) were reported, both in the 2-day/week group. Both of these patients who failed had withdrawn from prophylaxis. The authors concluded that a one day per week regimen of trimethoprim/sulfamethoxazole may be sufficient to prevent PCP in immunosuppressed children with cancer.

Take home message: One day per week trimethoprim-sulfamethoxazole dosing for PCP prophylaxis appears sufficient to prevent PCP in immunosuppressed children with cancer, is simpler than twice or thrice weekly dosing, and may improve adherence. This regimen merits consideration in children with cancer and potentially in children at risk of PCP from primary immunodeficiency, transplantation or other medical immunosuppression.

Submitted by: Brendan McMullan, Sydney Children's Hospital, 14 February 2014.