Computer Keyboards as Reservoirs for Acinetobacter baumannii in a Burn Hospital

SIR—In the January 1999 issue of Clinical Infectious Diseases, there were 2 articles plus an editorial about Acinetobacter [1—3]. One of the articles was from Hong Kong [1], and the other was from Germany [3], thereby indicating the widespread distribution of this increasingly emerging nosocomial pathogen. In one of these articles, Wisplinghoff et al. [3] found that prior nosocomial colonization with Acinetobacter baumannii at a distant site was strongly associated with the acquisition of A. baumannii bloodstream infections in burn patients; they aptly concluded that their data underscore the need for effective control measures for this emerging nosocomial problem. In the past few years at our pediatric burn hospital here in the United States, we too have become increasingly aware of the presence of A. baumannii and have developed improved control measures because of this microbe. In contrast to the studies from Germany and France, which showed A. baumannii was highly endemic in hospitals there, we found the microorganism is not highly endemic in our hospital. Of the 587 acute burn patients admitted to our hospital between January 1996 and December 1998, only 8.3% had cultures of specimens obtained at admission that yielded Acinetobacter. Further, our rate of transfer of A. baumannii to patients tends to be quite low; specifically, in 1996 and 1998, only 4.2% of our acute burn patients acquired Acinetobacter colonization (less than 1 transfer per month). In the spring of 1997, however, we noticed an increase in the number of acquired A. baumannii colonizations in our pediatric burn patients (figure 1). Cultures of environmental specimens showed that Acinetobacter was on various surfaces in the patients' rooms, especially on the plastic covers that fit over the bedside computer keyboards. Observation revealed that gloved patient care staff moved back and forth between the patient and the keyboard and that ungloved support staff, who never touched the patient, entered and retrieved data from the same computer. We hypothesized that the ungloved personnel were contaminating the keyboard with A. baumannii that was then picked up by the gloved patient care staff who transferred it to the patient. In essence, the keyboard was acting as an A. baumannii reservoir, a hypothesis that is consistent with studies showing that Acinetobacter species can survive for long periods on dry surfaces [4, 5]. Effect of computer keyboard—related infection control interventions on number of acquired Acinetobacter baumannii colonizations in a burn hospital. Two control measures were instituted. First, it was required that all personnel put on gloves before using the computer and that those gloves be removed before the person left the room, thereby preventing the support staff from contaminating the keyboards with Acinetobacter or from carrying the microorganism from one keyboard to another. Historically, these computers were introduced into the patients' rooms in the last half of 1996 and were used increasingly in 1997. At the time of our investigation in early 1997, it was discovered that there were no specific procedures for the disinfection of the plastic covers on the keyboards. Hence, as a second controlmeasure, a defined cleaning procedure for the plastic keyboard coverswas provided to the housekeeping staff, and this procedure was followed daily. These two control measures were initiated at the end of May 1997 with the result that the number of acquired A. baumannii colonizations during the last 7 months of 1997 was significantly less (P <.05, x2 test) than the number in the first 5 months (figure 1). These control measures have remained in place, and the transfer of this microorganism has remained low (figure 1). Although other factors may have been involved in the 1997 spring outbreak of acinetobacter transfers, the temporal relationship of the introduction of computers at the bedside before the increase in A. baumannii transfers and the significant decrease in these transfers once two control parameters relative to keyboard use were instituted strongly suggest that the keyboards were acting as reservoirs for transfer of this microorganism to the patients. It should be noted that, consistent with the findings of the German study [3], two of the acquired colonizations in our patients were associated with blood cultures positive for A. baumannii; both of these colonizations were successfully resolved. This investigation provides an example of how various departments (housekeeping, nursing, support staff, infection control, etc.) need to and can work together to provide the effective control measures that Wisplinghoff et al. [3] so appropriately recommended at the conclusion of their case-control study of A. baumannii bloodstream infections in burn patients.