Letter to the Editor The incomplete cleaning and disinfection and use of reusable oxygen humidifier for several patients probably created a reservoir for L. pneumophila and facilitated transmission.6 This matched caseecontrol analysis showed that the risk of LD increased with every additional minute of exposure to the contaminated humidifier although weakly statistically associated. Our regular water distribution system monitoring did not disclose potentially dangerous situations. A contamination level of L. pneumophila of 40% positive water samples out of a total of 15 samples was detected in September 2007, of which only one sample exceeded 103 cfu/L. Treatment of the water distribution system was repeated one month later and revealed no contaminated water samples. Humidifiers have been implicated in transmission of LD, but this is the first description of the involvement of sterile water refilling in a nosocomial outbreak of LD.1e5 Unfortunately, it was not possible to compare clinical and environmental isolates genetically, because the bacteria were not recovered from the patient and the small number of cases limited the power and precision of the study. We recommend using disposable oxygen humidifiers. Furthermore, staff adherence to infection control procedures should be reinforced. No new L. pneumophila infections were identified in our hospital following implementation of these measures.
Acknowledgements We thank maintenance personnel for their technical assistance, the physicians C. Merino and F. Tarazona for clinical management, the physician C. An ˜o ´ and the nurse supervisor M.A. Go ´mez for their help in the PARU, and the microbiologist V. Domı´nguez for the sample management. Conflict of interest statement None declared. Funding sources None.
References 1. Woo AH, Goetz A, Yu VL. Transmission of legionella by respiratory equipment aerosol generating devices. Chest 1992; 102:1586e1590. 2. Mastro TD, Fields BS, Breiman RF, Campbell J, Plitkaytis D, Spika JS. Nosocomial Legionnaires disease and use of medication nebulizers. J Infect Dis 1991;163:667e671.
383 3. Moiraghi A, Castellani Pastoris M, Barral C, et al. Nosocomial legionellosis associated with use of oxygen bubble humidifiers and underwater chest drain. J Hosp Infect 1987;10: 47e50. 4. Arnow P, Chou T, Weil D, Shapiro EN, Kretzschmar C. Nosocomial Legionnaires’ disease caused by aerosolized tap water from respiratory devices. J Infect Dis 1982;146:460e467. 5. Zuravleff JJ, Yu VL, Shonnard JW, Rihs JD, Best M. Legionella pneumophila contamination of a hospital humidifier. Demonstration of aerosol transmission and subsequent subclinical infection in exposed guinea pigs. Am Rev Respir Dis 1983; 128:657e661. 6. Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R, CDC, Healthcare Infection Control Practices Advisory Committee. Guidelines for preventing health care associated pneumonia, 2003: recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee. Morb Mortal Wkly Rep 2004;53(RR-3):1e36. 7. Fields BS, Benson RF, Besser RE. Legionella and Legionnaires’ disease: 25 years of investigation. Clin Microbiol Rev 2002; 15:506e526.
R. Bou* P. Ramos Infectious Diseases Group, Hospital Universitari de La Ribera, Alzira, Valencia, Spain E-mail address: [email protected]
Available online 5 February 2009 * Corresponding author. Address: Hospital Universitari de La Ribera, Ctra. de Corbera, km. 1, 46600 Alzira, Valencia, Spain. Tel.: þ34 96 245 83 25; fax: þ34 96 245 81 56. ª 2009 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2009.01.002
Neonatal sepsis in a neonatal intensive care unit in Indonesia*
Madam, In developing countries neonatal mortality is about 34 per thousand newborn infants. Most of these deaths occur in the first week of life.1 The World Health Organization estimates that infection, prematurity and birth asphyxia are the main causes of death.2 A study from Pakistan showed that the rate of neonatal infections is high in neonatal intensive care units (NICUs) in developing countries, Gram-negative bacteria being the most * Presented in part at the Society for Pediatric Research Meeting, San Francisco, USA, 2006.
384 common causative agents.3 We studied the incidence of neonatal sepsis on day 1 and day 3e5, the bacteria causing infection, and the relation between the incidence of sepsis and prescribed antibiotics in our NICU. We included all newborn infants admitted to the NICU of Harapan Kita Hospital, Jakarta, Indonesia, between May 2003 and June 2005. Data on mothers and infants were recorded. Antibiotics given prior to delivery were also noted. The most commonly used antibiotics were third-generation cephalosporins. Neonatal septicaemia was suspected when infants showed any signs of infection (respiratory insufficiency, bradycardia, lethargy, poor feeding, seizures or temperature instability). Blood cultures were analysed with the BACTEC 9240 Rapid Detection System (BectoneDickson, Sparks, MD, USA) using the PEDS Plus culture bottles and media. When neonatal septicaemia was suspected, broad-spectrum antibiotics were started. When the first culture was negative but there was clinical improvement the blood culture was repeated on day 3e5. Neonatal septicaemia was confirmed by positive blood culture. In the study period 6600 infants were born. Of these, 216 were admitted on the first day of life to the NICU. In 163 out of these 216 infants the mother received antibiotics before delivery. In 133 of the 216 infants an infection was suspected on the first day of life. Four out of nine with proven sepsis on day 1, and 45 out of 63 with proven sepsis on day 3e5, were due to Serratia species. Klebsiella pneumoniae accounted for a further nine of these 63 infants. Most cases of proven sepsis either on day 1 or 3e5 had been exposed to antenatal antibiotics. We did not find a relationship between infections either on day 1 or day 3e5 with any maternal risk factors and no relation to birthweight, gestational age and Apgar score. We observed an important difference in the rates of sepsis between day 1 and day 3e5. On day 1 we found that nine out of 216 infants had a positive blood culture compared with 63 positive cultures in these same infants on day 3e5. Infections in newborn infants are often divided into early-onset, diagnosed within the first days of life, and late-onset sepsis occurring after the first week. Early-onset sepsis is mainly caused by perinatally acquired bacteria, Group B streptococci, H. influenzae and Escherichia coli. Late-onset infections are considered nosocomial and caused by staphylococci and Candida spp.4 The rate of earlyonset sepsis in the NICU of the University Medical Centre Groningen is w0.5 per 1000 deliveries including all gestational ages (P. van den Broek and S. Lusyati, unpublished data). In our hospital
Letter to the Editor we observed an incidence of early neonatal sepsis of 1.4/1000 deliveries. This incidence in our hospital is higher than in NICUs in developed countries, when taking into account the higher weight and gestational age in infants in our unit. In our unit Serratia spp. was the commonest infecting agent, both on day 1 and 3e5. It is unclear whether infants in our unit were contaminated with Serratia spp. during delivery or soon after birth in our unit. In our hospital the majority of mothers received broad-spectrum antibiotics before delivery. Whether this caused an abnormal bacterial flora in the mothers, which then transmitted to the infants is unclear. Almost all proven sepsis on day 3e5 was caused by Gram-negative bacteria with Serratia spp. as the main organism, followed by K. pneumoniae. These infections must be considered as nosocomial. The postnatal age of our infants developing a nosocomial infection is lower than that found in studies in developed countries, whereas the incidence in our unit is much higher. Our results therefore indicate a serious problem with nosocomial infections in our unit starting soon after admission to the unit. Efforts therefore have to be directed to reduce the incidence of nosocomial infections. Conflict of interest statement None declared. Funding sources Supported by Nestle ´ Nutrition Institute Pediatric Scholarship Award ‘in Neonatology studies’, Switzerland, 2005.
References 1. Costello A, Francis V, Byrne A, et al. The states of the world’s newborns. Washington: Save the Children Fund; 2001. 2. Anonymous. Perinatal mortality. Report No.: WHO/FRH/ MSM/967. Geneva: WHO; 1996. 3. Zaidi AK, Huskins WC, Thaver D, Bhutta ZA, Abbas Z, Goldmann DA. Hospital-acquired neonatal infections in developing countries. Lancet 2005;365:1175e1188. 4. Stoll BJ, Hansen N. Infections in VLBW infants: studies from the NICHD Neonatal Research Network. Semin Perinatol 2003;27:293e301.
S. Lusyatia,b,* P. van den Broekb P.J.J. Sauerb a Paediatrics Department, Women and Children’s Harapan Kita Hospital, Jakarta, Indonesia
Letter to the Editor
Paediatrics Department, Beatrix Children’s Hospital, University Medical Center Groningen, The Netherlands E-mail address: [email protected]
Available online 12 February 2009 * Corresponding author. Address: Pediatric Department, Women and Children’s Harapan Kita Hospital, Jakarta, Indonesia. Tel./fax: þ62-21-53152570. ª 2009 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2009.01.004
Bacterial flora in residents of long-term care facilities: a point prevalence study
Madam, The bacterial flora of residents of long-term care facilities (LTCFs) differ from the flora of residents in the general community. Factors like poor hand hygiene and selective pressure of antimicrobials contribute to infection in the LTCF population with serious health and financial consequences.1e3 The present study focused on factors other than antibiotic consumption that may promote carriage of resistant micro-organisms. This point prevalence study was performed during the period June 2002 to May 2004 in 41 LTCFs of the Athens region. In total, 1523 residents participated after giving written informed consent. Residents with any signs of infection were excluded. Swabs were taken from decubitus ulcers and anterior nares and urine samples were collected. Sampling was done once for each resident irrespective of the length of stay in the LTCF and once for each LTCF during the study period. Sample cultures, identification of isolates and susceptibility testing were performed according to standard techniques.4 Production of extended-spectrum b-lactamases (ESBLs) by Klebsiella pneumoniae was assessed by the double disk approximation test.5 Full clinical data were collected for each participant. Residents were divided into two groups: group A, consisting of residents of 32 LTCFs with perceived poor hygiene, i.e. four residents per room sharing the same toilet, one nurse per 16 residents and a physician available only on demand; and group B, consisting of residents of nine LTCFs with good hygiene infrastructure, i.e. one
resident per room with single toilet, one nurse per four residents and physician available daily at a 2:1 rate of nurses to physicians.6 In total, 328 residents were male and 1195 female. Their mean SD age was 84.6 6.3 years. A total of 1139 belonged to group A and 384 to group B; 474 (41.6%) and 128 (33.3%) respectively had a positive urine culture (P ¼ 0.005). The commonest isolates were Escherichia coli (19.1%) and Proteus mirabilis (9.4%) in group A; and P. mirabilis (11.5%) and E. coli (10.9%) in group B. Carriage of a micro-organism was found in the nares of 309 (27.1%) and 32 (8.3%) residents of groups A and B respectively (P < 0.0001). The commonest isolates were Staphylococcus aureus (15.0%) and P. mirabilis (4.5%) in group A; and S. aureus (3.9%) and P. mirabilis (1.6%) in group B. Carriage of a micro-organism was found in the decubitus ulcers of 152 (13.4%) and 12 (3.1%) residents of groups A and B respectively (P < 0.0001). The commonest isolates were P. mirabilis (4.0%) and E. coli (2.4%) in group A; and P. aeruginosa (1.5%) and E. coli (0.5%) in group B. Their resistance patterns are shown in Table I. Analysis failed to disclose any effect of the length of LTCF stay on the colonisation in group A. Mean SD length of stay of group B residents with sterile urine was 34.6 20.8 months; for those with positive urine cultures it was 39.7 23.4 months (P ¼ 0.029). Factors favouring colonisation of the urinary tract were urinary catheterisation (OR: 3.83), decubitus ulcers (OR: 2.11), lack of mobility (OR: 1.61), and type 2 diabetes mellitus (DM2; OR: 1.75). Those favouring nasal colonisation were nasogastric feeding tubes (OR: 92.51), decubitus ulcers (OR: 4.06) and intake of sedatives (OR: 1.54); those favouring colonisation of decubitus ulcers were urinary catheterisation (OR: 2.14), nasogastric feeding tubes (OR: 600.00) and intake of sedatives (OR: 1.65). The single factor affecting the urinary tract colonisation by ESBL-producing K. pneumoniae was the intake of broad-spectrum b-lactams (OR: 6326.1; P < 0.0001) and of trimethoprim-sulfamethoxazole (OR: 27.5; P < 0.0001) during the previous trimester. It was evident that the patient flora was dependent on the level of hygiene provided by the LTCF. The most heavily colonised sites were the urinary tract followed by anterior nares and decubitus ulcers. Colonisers of urine and decubitus ulcers were predominantly Gram-negative bacteria; that of anterior nares was mostly S. aureus. These findings are in general agreement with those reported by other authors.1,7 The most