|
|
||||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Division of Healthcare Quality Promotion, National Center for Infectious Diseases (Drs. Pati, Arduino, Jernigan, and Srinivasan, and Ms. Jensen) and the Epidemic Intelligence Service, Epidemiology Program Office, Office of Workforce and Career Development (Drs. Estivariz and Bhatti), Centers for Disease Control and Prevention, Department of Health and Human Services, Atlanta, GA; and the Department of Pediatrics and Communicable Diseases (Dr. LiPuma), University of Michigan Medical School, Ann Arbor, MI.
Correspondence to: Concepcion F. Estivariz, MD, Centers for Disease Control and Prevention, 1600 Clifton Rd NE, MS-E05, Atlanta, GA 30333; e-mail: CEstivariz{at}cdc.gov
Abstract
Background: Species within the Burkholderia cepacia complex (Bcc) can contaminate medications and disinfectants and cause severe pneumonia in critically ill patients or persons with cystic fibrosis. In March 2004, we investigated a hospital outbreak of Bcc possibly associated with a contaminated nasal spray.
Methods: We conducted a matched case-control study, environmental sampling, and observations of infection control practices. Case patients had infection or colonization with Bcc, and control patients had sputum culture not yielding Bcc. Isolates from patients and environmental samples were compared by pulsed-field gel electrophoresis (PFGE).
Results: Bcc was recovered from sputum in 18 patients. Compared with matched control patients (n = 18), case patients were more likely to be receiving mechanical ventilation (p = 0.01), to have been hospitalized > 6 days (p = 0.01), and to have received antimicrobial treatment within 7 days before sputum collection (p = 0.03). Bcc was cultured from opened, but not unopened, multidose albuterol bottles, a nebulizer attached to a ventilator, and opened and unopened nasal spray bottles from contaminated lots. PFGE showed that isolates from albuterol samples and from patients were indistinguishable but unrelated to the nasal spray strain. Observations revealed improper aseptic techniques during respiratory therapy procedures and inadequate nebulizer cleaning.
Conclusions: Despite a temporal association with use of a contaminated nasal spray, this outbreak was caused by extrinsic contamination of multidose albuterol used for nebulization treatments and lack of adherence to infection control precautions. Implementation and re-enforcement of infection control measures successfully terminated the outbreak.
Key Words: critical care health-careassociated pneumonia ventilation
The Burkholderia cepacia complex (Bcc) is a group of bacterial species or genomovars, of which one, genomovar I, is designated B cepacia.1 Members of the complex are widely distributed in the environment and are intrinsically resistant to many antimicrobial agents, disinfectants, and antiseptic solutions.234 These species have emerged as opportunistic respiratory pathogens in persons with cystic fibrosis (CF)5 and critically ill patients.167 Bcc was isolated in 0.2% of all health-careassociated pneumonias reported to the National Nosocomial Infections Surveillance System between 1998 and 2004 (unpublished Centers for Disease Control and Prevention [CDC] data). Though infrequent, Bcc infections can be severe, with reported mortality rates as high as 83% among patients with lower respiratory tract infections.89 Bcc also has been implicated in nosocomial outbreaks of respiratory infections due to both intrinsic610 and extrinsic contamination of medical devices and products.71112131415161718
In March 2004, several lots of an over-the-counter nasal decongestant were recalled by the manufacturer for intrinsic contamination with Bcc.1920 Following this notification, an adult acute care facility (Hospital A) in Missouri discovered the recalled product in their pharmacy supplies and suspected a possible association between the contaminated nasal spray and an outbreak of Bcc respiratory infections ongoing in the facility for several months. On March 22, 2004, the CDC was invited to investigate the outbreak, to explore the potential connection with the nasal spray, and to recommend prevention and control measures.
Materials and Methods
Case Finding
A case patient was any patient admitted to Hospital A between October 2003 and March 2004 who had at least one culture specimen from which Bcc was isolated (Bcc positive). Active surveillance was conducted from March 25 to June 2, 2004, to identify and isolate patients with colonization. Surveillance culture samples from patients in the ICU, receiving mechanical ventilation, or capable of producing sputum without aggressive induction were collected on hospital admission, weekly, and on discharge from the ICU.
We used pharmacy logs to identify and contact patients who had received the recalled nasal spray. Patients who had clinical symptoms after use of the nasal spray were encouraged to return to the hospital for examination and collection of sputum cultures and nasal swabs.
Case-Control Study
Control patients were randomly selected from a list of patients admitted for at least 72 h between October 2003 and March 2004, who had at least one Bcc-negative sputum sample collected during that hospital admission. Because patients often had several cultures, we selected a "reference" culture collection date. For case patients, the reference culture was the first sputum yielding Bcc; for control patients, it was the culture with closest exposure time (number of days between hospital admission and culture collection) to the matched case. We matched case patients and control patients for gender, age within 15 years, and reference sputum samples collected within 40 days (median, 4 days; range, 0 to 37 days). Clinical information abstracted included underlying diseases, hospital and ICU length of stay (LOS), mechanical ventilation, and medications and invasive procedures (line insertions, endoscopies, intubations or surgery) in the 7 days prior to collection of the reference sputum.
Environmental Investigation
We reviewed infection control techniques during administration of respiratory therapy and patient care in the ICU, as well as procedures for maintenance, cleaning, and disinfection of ventilator equipment.
Microbiologic Studies
Samples for culture of medications, health-care worker hands, and the environment were collected on several dates in February and March. One milliliter of solutions and swabs from environmental surfaces and hands was inoculated onto MacConkey or Pseudomonas cepacia agar (Remel; Lenexa, KS). Water samples (250 mL) were collected in sterile bottles containing sodium thiosulfate and cultured with a method described previously.21 All samples were incubated for 48 h at 35°C. Gram-negative organisms were identified using an automated system (Vitek bioMérieux; Durham, NC; or manual RapID NF; Remel). Putative Bcc isolates were further analyzed using selective media, biochemical analyses, 16S recombinant RNA-specific and recA species-specific polymerase chain reaction assays, and recA restriction fragment length polymorphism analyses as described previously.2223
Available Bcc isolates from patients and environmental samples were compared using pulsed-field gel electrophoresis (PFGE) following digestion of chromosomal DNA with the restriction endonuclease Spe I.24 Isolates from the outbreak and patients exposed to contaminated nasal spray in other facilities nationwide were also compared by random amplification of polymorphic DNA.25
Statistical Analysis
Data were analyzed using statistical software (SAS version 9.1; SAS Institute; Cary, NC). Matched univariate analysis was performed using the Cochran-Mantel-Hanszel test to determine odds ratios. Conditional logistic regression was performed for risk factors with p < 0.05.
Results
Case Patients
Between October 1, 2003, and March 31, 2004, 18 patients had culture specimens that grew Bcc, all from respiratory samples (Fig 1
). This was a significant increase from the baseline of zero to one Bcc culture per year detected in from 2000 to 2003. Two more cases were detected after the implementation of active surveillance in March 2004, but no other cases were identified for the next 3 months despite active surveillance.
|
|
A committee of infectious diseases and ICU specialists reviewed charts of cases using CDC definitions for nosocomial infections2627 and a priori established definitions for attributable mortality. This committee determined that two case patients had Bcc pneumonia, while positive culture results in the other 16 patients represented colonization. Both patients with pneumonia died, and in both cases the committee attributed the deaths, at least in part, to the Bcc pneumonia. Mortality could not be attributed to Bcc infection for the other six case patients who died.
Case-Control Study
There were no differences between case patients and control patients with respect to age, underlying disease, or prior hospitalization (Table 2
). For control patients, the median interval between hospital admission and the reference sputum collection was 4 days (range, 1 to 24 days). Therefore, case patients were more likely than matched control patients to have stayed in the hospital or ICU for > 6 days before sputum collection (Table 2). This difference was significant despite selecting the reference sputum collection date with the closest exposure time to that of matched cases. Other risk factors associated with Bcc respiratory acquisition included mechanical ventilation, receipt of nebulized albuterol for > 3 days, and antimicrobial treatment in the previous week (Table 2). Daily nebulized albuterol treatments consisted of six doses for patients receiving mechanical ventilation and four doses for patients not receiving mechanical ventilation, but missed doses were not always recorded. Exposure to other potential risk factors for respiratory infections in hospitalized patients was not different between case patients and control patients. None of the identified risk factors was independently associated with recovery of Bcc in the conditional logistic regression.
|
Nationwide, Bcc colonization after using this nasal spray was reported in 12 children in six states, 4 of them with CF. However, none had clear evidence of either infections or prolonged colonization following exposure to the contaminated nasal spray.20
Microbiologic Studies
A total of 370 environmental samples were collected. Swabs were obtained from respiratory and ventilator equipment, sink faucets and drains, otolaryngology instruments, and the hands of 80 health-care workers with direct care of ICU patients. In addition, samples of medications, IV fluids, tube feedings, and antiseptics used by case patients and tap water and environmental disinfectants were collected in the ICU.
The following 10 environmental samples were positive for Bcc: (1) three opened and unopened bottles of "Twice-A-Day" nasal spray, lot K 4496, recalled by the manufacturer (Propharma; Miami, FL); (2) two opened albuterol multidose bottles collected on different days from the pocket of a respiratory therapist (two unopened bottles from the same lot were negative); (3) an in-line nebulizer cup attached to the ventilator circuit of a case patient; and (4) internal surfaces and suction cups of ventilator tubing used by a case patient. Although Bcc was not cultured in any of the hand swabs or wipes, Gram-negative rods, Bacillus spp, or yeast were detected on the hands of 3 of 15 respiratory therapists (20%), compared to 3 of 53 nurses and health technicians (6%), and 0 of 12 physicians.
PFGE analysis of available clinical isolates showed that Bcc isolates recovered from the sputum of the last four case patients were indistinguishable from the strain isolated in the albuterol bottles. However, isolates from patients and albuterol were different from the strain isolated in the nasal spray bottles (Fig 2 ). Based on polyphasic testing, the Bcc species implicated in this outbreak was found to be B cepacia (genomovar I of the Bcc). The isolates recovered from contaminated nasal spray nasal belonged to an unnamed species within the Bcc.
|
Discussion
This investigation demonstrated that extrinsic contamination of albuterol medication used for nebulization therapy in multiple patients, combined with suboptimal infection control practices, was the most likely mechanism for transmission of Bcc. Despite a temporal association, the intrinsically contaminated nasal spray was not the source of this outbreak.
Several findings support this conclusion. First, Bcc contamination was detected in two open multidose albuterol bottles used at different times during the outbreak, and the PFGE patterns of isolates from these bottles were indistinguishable from those of available patient isolates. Second, contamination with Bcc was also demonstrated in the in-line nebulizer cup used by one case patient. In-line nebulizers were used for all patients receiving mechanical ventilation in the facility. Previous studies2829 have shown that contamination with oropharyngeal and tracheal secretions is likely to happen when nebulizers stay attached to the ventilator tubing between treatments. Use of contaminated nebulizers generate small-particle bacterial aerosols that may increase the risk of ventilator-associated pneumonia.29 Likewise, contamination of the in-line nebulizers increases the risks of contaminating multiple-dose vials of medications when nondisposable syringes or droppers are used to dispense the medications. For this reason, guidelines developed by CDC and the Healthcare Infection Control Practices Advisory Committee2830 recommend that nebulizer cups be removed, cleaned, rinsed with sterile water, and allowed to air dry between treatments. This policy was not followed in Hospital A at the time of the outbreak. Third, despite collection of multiple environmental samples, no other potential sources were found to be contaminated with Bcc. Fourth, review and observation of respiratory therapists revealed lapses in infection control practices that could have facilitated Bcc transmission.
We hypothesize that Bcc was transmitted when secretions from a patient receiving mechanical ventilation with Bcc infection or colonization contaminated the nebulizer reservoir that was left in-line between treatments. When the respiratory therapist added new medication to the nebulizer reservoir, the dispensing dropper became contaminated. Bcc was then introduced into the albuterol bottle, where it multiplied and was transmitted to other patients receiving nebulized albuterol. Though multidose albuterol contained benzalkonium chloride as a bacteriostatic agent, Bcc can survive in solutions containing benzalkonium chloride, especially when bottles stay open several days.413
The original source of the Bcc remains unclear. It is possible that Bcc was present at certain point in the hospital water supply,31 and water cultures failed to detect the bacteria. Likewise, it is possible that an unrecognized patient was the initial source.
Mechanical ventilation and treatment with broad-spectrum antimicrobial agents were risk factors for Bcc respiratory acquisition in this investigation, and a similar association has been described in other Bcc outbreaks.1315 Oropharyngeal bacterial colonization during intubation, poor cough reflex, and direct inhalation of contaminated aerosols into the lower respiratory tract have been involved in the higher risk for pneumonia of patients receiving mechanical ventilation.28 Treatment with cephalosporins and/or vancomycin may facilitate growth and culture isolation of Gram-negative bacteria resistant to both antimicrobials, such as Bcc.131528 Factors that result in Bcc respiratory infection following colonization are less clear. The new strain recovered from the nasal spray did not appear to be associated with any adverse outcomes, even in the presence of CF. However, the outbreak strain was associated with patient-to-patient transmission and severe pneumonia and death in two patients. These results support other studies13233 suggesting higher transmissibility and virulence for certain strains within the complex, although the molecular mechanisms are still unclear.
Our investigation had several limitations. First, by choosing exclusively patients with sputum culture during hospitalization, we selected a control population with a high likelihood to require nebulized medications; and the risk of receiving nebulized albuterol among case patients may have been underestimated. Second, it was not possible to find ward-, age-, and gender-matched control patients with sputum samples collected the same day as the case patient sputum. As a result, risk factors on specific days may not be present in all matched pairs. Third, the exposure time for case patients was longer than that of control patients. We tried to minimize this effect by selecting the sputum collection date for a control patient with the closest exposure time to that of the matched case patient. To assess the impact of this difference, exposure time was evaluated as a risk factor in the multivariate analysis and was not found to be an independent risk factor for Bcc acquisition. Fourth, it is possible that the true number of case patients was higher than reported because active surveillance for Bcc in the respiratory tract started late in the outbreak. Finally, Bcc isolates were available for genotypic characterization from only a few of the case patients.
In conclusion, implementation and re-enforcement of infection control measures successfully halted transmission of Bcc in this facility. Intrinsic contamination of medical products remains an important cause of health-careassociated outbreaks. However, as our investigation demonstrates, the fact that a contaminated product is in use at a facility is not sufficient evidence that it is the cause of an outbreak. Thorough investigations of outbreaks remain important in determining the true routes of transmission so that appropriate interventions can be implemented.
Acknowledgements
We thank personnel in Hospital A and the Missouri Department of Health and Senior Services for their cooperation and assistance in making this investigation possible.
Footnotes
Abbreviations: Bcc = Burkholderia cepacia complex; CDC = Centers for Disease Control and Prevention; CF = cystic fibrosis; LOS = length of stay; PFGE = pulsed-field gel electrophoresis
This study was performed at Hospital A, Missouri, with the collaboration of The Centers for Disease Control and Prevention, Missouri Department of Health and Senior Services, and University of Michigan Medical School.
This work has no funding research support sources.
The authors have no conflicts of interest to disclose.
Received for publication March 13, 2006. Accepted for publication May 3, 2006.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |