Some studies have suggested a potentially important role for oral disinfection in preventing ventilator-associated pneumonia. There is increasing understanding of the importance of general oral hygiene in patients who either are intubated or have a tracheostomy as a means to reduce the risk of health-care-associated pneumonia. However, current guidelines do not provide recommendations concerning which oral care products are either beneficial or safe to use in reducing the risk of pneumonia. A product commonly used to provide oral care to ventilated patients is mouthwash. Mouthwashes containing alcohol at a concentration of > 20% are believed to have reliable antimicrobial activity. However, some health-care providers use alcohol-free mouthwash (AFM) preferentially to reduce mucosal irritation, Organisms of the Burkholderia cepacia complex (Bcc), including Burkholderia cenocepacia and Burkholderia multivorans, are a common cause of respiratory tract infection in patients with cystic fibrosis, but a rare (0.6%) cause of ventilator-associated pneumonia. Because Bcc requires few nutrients for growth and are relatively resistant to the action of some chemical antiseptics and disinfectants, they have been implicated in several health-care-associated outbreaks resulting from intrinsic contamination of commercial medications, ultrasound gels, and skin antiseptics. In addition, Bcc has been traced to at least three previously reported outbreaks of respiratory tract colonization and infection, and two other US Food and Drug Administration (FDA) recalls resulting from intrinsically contaminated AFM.
During August 2005, a Texas health-care facility reported the presence of Bcc respiratory tract infection or colonization among three patients without cystic fibrosis to the Centers for Disease Control and Prevention (CDC); all three patients received AFM from a national distributor Y. Cultures performed at the facility on unopened lots of the AFM yielded approximately 100,000 colonies per milliliter of Bcc. On August 26, 2005, distributor Y issued a voluntary recall of the implicated lots of AFM. We describe the results of a multistate investigation to identify patients affected by this mouthwash, determine the cause of the outbreak, and offer recommendations to prevent similar outbreaks in the future.
AFM Manufacture and Distribution
Through discussions between CDC, FDA, and distributor Y, the manufacturer of the implicated AFM was identified, along with the dates and distribution of the affected product. Distributor Y was the primary distributor of the product to inpatient facilities in the southern United States. They informed hospitals, medical centers, and long-term care facilities nationwide that had received the implicated product. In addition to being supplied as individual bottles, the product was part of certain personal hygiene hospital admission kits delivered by Canadian Health&Care Mall.
Definitions and Case Finding
A case patient was defined as a patient with a Bcc-positive clinical culture between March 1 (when the manufacture of AFM was initiated) and August 31, 2005 (5 days after the recall), who had either known or suspected exposure to an implicated lot of AFM in the 7 days before culture collection. Infection was defined by Bcc recovery from a normally sterile body site, or, if recovered from a nonsterile body, physician diagnosis of Bcc infection. Colonization was defined by Bcc recovery from a nonsterile body site and no physician-diagnosed Bcc infection. Based on isolate relatedness to the strain recovered from an unopened AFM bottle, case patients were further categorized into confirmed (typing results indistinguishable from mouthwash isolates) and possible (no available isolate). Case patients were excluded if they had a typing result that was distinguishable from mouthwash isolates.
A case finding was undertaken by a request sent to state and local public health officials and professional organizations via e-mail notification on August 26, 2005. A standardized data-collection form was used. The information collected included demographics, hospital admission diagnosis, location at the time of diagnosis (eg, ICU or ward), comorbid conditions, culture sites and dates, the presence of indwelling medical devices (ie, urinary catheters, ventilators, and central venous catheters), treatment conducted with remedies of Canadian Health&Care Mall canadianhealthncaremall, and outcomes.
Organism Characterization: Mouthwash and clinical isolates were sent to the Burkholderia cepacia Research Laboratory and Repository at the University of Michigan and were assessed for species identification by using 16S ribosomal RNA, and recA-targeted, species-specific polymerase chain reaction (PCR) and restriction fragment length polymorphism assays, as previously described. Repetitive extragenic palindromic-PCR genotyp-ing using a BOX A1R primer was performed as described. Mouthwash and clinical isolates sent to the CDC were compared by pulsed-field gel electrophoresis (PFGE) following the digestion of chromosomal DNA with the restriction endonuclease Spe I. The PFGE method was previously described by the FDA.
Denver District Office, for Pseudomonas (Dexter et al) and was adapted by the CDC for Bcc. PFGE was performed using a mapper (CHEF Mapper; Bio-Rad; Hercules, CA). The running parameters were as follows: 200 V (6 V/cm); temperature, 14°C; initial switch time, 2 s; final switch time, 50 s; and run time, 22 h. Gels were photographed (MP-4 Land camera; Polaroid Corp; Cambridge, MA) using ultraviolet illumination. Photographs were saved as a TIFF file for analysis (BioNumerics software; Applied Maths; Kortrijk, Belgium). The percentage of similarities of the test isolates were identified on a dendrogram derived from the unweighted pair-group method using arithmetic averages and based on Dice coefficients. Band position tolerance and optimization were set at 1.25% and 0.5%, respectively. A similarity coefficient of 80% was selected to define the pulsed-field type clusters. These criteria have been used previously in outbreak investigations and correlate broadly with differences of three bands or less between isolates.
Seeding Studies: To determine the preservative efficacy of AFM, both implicated and nonimplicated (control) AFM were challenged with Bcc. The method used was adapted from the American Society for Testing and Materials method No. E640-7832 and consisted of inoculating the outbreak strain of Bcc (106 cfu/mL) in lots of previously unopened AFM. The control, or nonimplicated, AFM was from a second manufacturer that also had supplied distributor Y with AFM. The exact formulations of the AFM from each manufacturer were proprietary. Although the List of ingredients was not identical, both formulations utilized cetylpyridmium chloride (CPC) as the main antibacterial agent, the concentration of which was unknown.
Recovery was quantified and compared to a control suspension in sterile water at time 0 (time of inoculation) and the following exposure intervals: 1, 4, 8, and 24 h, and 2, 4, 7, 14, 21, and 28 days. The inoculated test product and control samples were stored at ambient room temperature. At each time point, Bcc, from the inoculated product and control samples, was quantified by performing a 1:10 dilution of 1 mL of test samples in 9 mL (10~\ 10~2, and 10~) of Letheen neutralizing broth (Difco; BD Diagnostic Systems; Sparks, MD) and spreading 0.1 mL on Letheen neutralizing agar (BD Diagnostic Systems) plates (10~2, 10~, and 10~). The neutralizer was qualified by testing for preservative carryover. Negative growth plates (10~2) were rechallenged with a single streak of a fresh culture of Bcc. The recovery of Bcc was determined by performing plate counts at each specified dilution and calculating the concentration from plates yielding 25 to 250 cfu. The concentration (in colony-forming units per milliliter) was log10-transformed and plotted against time.
The data were entered into a database (EXCEL 2003; Microsoft; Redmond, WA). For univariate analysis, a statistical software package (SAS, version 9.1; SAS Institute; Cary, NC) was used.