Ethics statement

This study was approved by the institutional ethics committees of the Academy of Military Medical Sciences and 309 Hospital of the Chinese People’s Liberation Army, Beijing, China. Written informed consent was obtained from all participants before the study.

Setting

The outbreak occurred in an ICU ward comprising a large open bedroom with ten beds, a buffer room, treatment room, and equipment room. Every bed was equipped with an alcohol-based hand rub. Fifteen doctors and thirty-one nurses worked in this ICU and approximately 12 nurses were on duty every day. There was only one bronchofiberscope in the ICU and bronchofiberscopy was performed once or twice each day for diverse examination and treatment indications such as corpus alienum removal, secretion clearance, tracheal intubations, and bronchoalveolar lavage. After each procedure, the bronchofiberscope was reprocessed by the professional staff in the Center for Cleaning and Disinfection of the hospital according to the Chinese guidelines for endoscopy cleaning and disinfection [23]. The standard procedure for reprocessing a bronchofiberscope includes the following steps: pre-cleaning, cleaning with an enzymatic detergent, rinsing, disinfecting, final rinsing, drying, and storing. However, when a bronchofiberscope was used emergently and frequently, it was reprocessed directly and manually by a doctor in the ICU after each use. Neither a doctor nor a nurse was specifically appointed to reprocess the bronchofiberscope and no automatic reprocessing machine was used.

Epidemiological investigation

During the epidemic period from 5^th August 2009 to 30^th November 2009, 153 patients were admitted to the ICU. The period from 1^st January 2009 to 4^th August 2009 was considered the pre-epidemic period. Medical records including paper and electronic charts were reviewed. Microbiological records were carefully analyzed to screen the cases and define the baseline MDR-Ab rate before the outbreak. Any patient who had at least one clinical or screening sample that was positive for a MDR-Ab who had the corresponding clinical symptoms (e.g., pneumonia, bacteremia, peritonitis) detected at least 48 h after ICU admission was noted. Multidrug resistance was defined as resistance to ≥3 of the following classes of antibiotics: penicillins, cephalosporins, aminoglycosides, fluoroquinolones, and carbapenems [24]. Environmental sampling was performed on 15^th, 21^st, and 28^th October 2009. Samples were taken from the hands and nasal cavities of the ICU staff as well as multiple surfaces within the ICU environment including: bed sheets, bedrails, and bedside tables associated with cases and controls; healthcare workers’ clothes, computer keyboards, and calculators; and the surfaces of invigilators, ventilators, hemofiltration machines, bronchofiberscopes, electrocardiography machines, ultrasound machines, and laryngeal endoscopes.

Case–control study

The case–control study was conducted to investigate this outbreak’s risk factors. Blood, urine, sputum, wound, bile, and catheter cultures were processed. A case was defined as a patient with at least one isolate identified as the outbreak MDR-Ab strain in clinical culture (outbreak strain carrier) at least 48 h after ICU admission during the period of 1^st September 2009 to 31^th October 2009. A control was defined as a patient who stayed ≥ 48 h in the ICU during the same period without the identification of an outbreak strain in any clinical culture [25]. Patients who stayed in the ICU < 48 h, carried strains other than the outbreak strain as determined by repetitive extragenic palindromic polymerase chain reaction (REP-PCR), or harbored MDR-Ab before the ICU admission were excluded from the study. The ratio of controls to cases was 2.7:1. At least one clinical urine, sputum, wound, or blood culture was processed for each control during the study period.

For the case–control study, the presence of primary diseases or medication history including septic shock, multiple organ failure, pulmonary diseases, renal diseases, and surgical operation was determined at the time of ICU admission. For invasive and other bedside procedures such as blood transfusion, mechanical ventilation, bedside diagnostic ultrasonography, bedside chest X-ray, bronchofiberscopy, electrocardiography, venipuncture, gastric lavage, urinary catheterization, and hemodialysis, the presence of a central line, and antibiotic use, the observation period lasted from ICU admission to outbreak strain detection for the cases and from ICU admission to patient discharge for the controls. ICU stay and hospital stay were defined as the length of stay until outbreak strain detection for the cases, whereas they were defined as the length of stay until patient discharge in the controls.

Microbiological methods

Swabs from environmental samples and healthcare workers were inoculated on blood plates. Colonies resembling Acinetobacter spp. were then isolated and transferred onto a China-Blue lactose agar plate (Luqiao, Beijing, China). The A. baumannii isolates were further identified according to their morphological and growth characteristics using the oxidase, triple sugar iron, and citrate tests. The clinical samples were also inoculated in blood agar and the isolates were identified using the automated Microscan-Walkaway Microbiology Identification System (Becton Dickinson, Sparks, MD, USA). Antimicrobial susceptibility was determined using the disk diffusion method and interpreted according to Clinical Laboratory Standards Institute guidelines. DNA extractions of the isolates were typed by REP-PCR using REP1 (5^′-IIIICGICGICATCIGGC-3^′) and REP2 (5^′-ICGICTTATCIGGCCTAC-3^′) primer sequences [26]. Characteristic DNA patterns were analyzed using BioNumerics software (version 3.0; Applied Math, Sint-Martens-Latem, Belgium) to determine the distance matrices and the unweighted pair group method with arithmetic mean method to create a dendrogram.

Intervention

Three major interventions were implemented on 21^st October 2009. First, the bronchofiberscope reprocessing by doctors within the ICU was stopped and the bronchofiberscope was sent to the Center for Cleaning and Disinfection of the hospital for professional reprocessing. Another one or two bronchofiberscopes were prepared for use in emergent situations in the ICU. Second, surveillance culturing for MDR microorganisms from the bronchofiberscope was performed regularly after every reprocessing round. Third, the ICU environmental surfaces were cleaned thoroughly and disinfected with a solution containing electrolyzed acid water according to the manufacturer’s instructions. Fourth, education and training were enhanced for endoscopy reprocessing and general infection control procedures in this ICU. In addition, to investigate the effect of hand hygiene on this outbreak, healthcare workers’ hand hygiene compliance was observed as described previously [27]. Briefly, infection control professionals recorded opportunities for hand hygiene during 1-h observation periods distributed randomly between 8:00 a.m. and 5:00 p.m. every day during the investigation. Hand hygiene compliance prior to the outbreak was calculated by review of the video documentation in the ICU. Healthcare workers’ hand hygiene compliance rose from 30% to 90% after the intervention.

Statistical analysis

Statistical analyses were performed using SAS software (version 8.1; SAS Institute Inc., Cary, NC, USA). The Chi-square test was used to compare the differences of MDR-Ab incidences between the pre-epidemic and epidemic periods. In the case–control study, cases and controls were compared using the Mann–Whitney U-test or Student’s t-test for continuous variables and using Fisher’s exact test for categorical variables. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for binomial variables. Two-sided P values < 0.05 were considered statistically significant. Multivariate logistic regression analysis was not applied due to the small sample size.

Epidemiological investigation

Three patients acquired MDR-Ab infection or colonization (MDR-Ab carriers) in the ICU during the pre-epidemic period. A total of 12 patients (seven males and five females ages 39–97) were identified as MDR-Ab carriers at least 48 h after ICU admission from 5^th August to 30^th November 2009 (Figure 1, Figure 2). Compared to the pre-epidemic period, the increase of MDR-Ab incidence during the epidemic period was statistically significant (χ² = 13.82, P < 0.001). Two MDR-Ab carriers who were admitted into the ICU before 20^th September 2009 were identified by medical record review. Five MDR-Ab carriers were identified in the 18 days between 21^st September and 8^th October 2009, while three other carriers were found in the four days between 19^th October and 22^nd October 2009. Thereafter, only two MDR-Ab carriers were identified on 2^nd November and 30^th November, and no other MDR-Ab carriers were detected up to the end of the study in January 2010.

The patients had a variety of underlying conditions, including septic shock, organ transplantation, malignant tumor, respiratory failure, acute pancreatitis, chronic obstructive pulmonary disease, multiple organ dysfunction syndrome, and coronary heart disease. Seven of the 12 patients were identified when they were staying in beds 1, 2, 3, 6, 7, and 8 (Figure 3). The average interval between ICU admission and MDR-Ab identification was 6.3 ± 3.8 days. Eight of the 12 patients had received bronchofiberscopy and five had BSIs. Six patients (50%) died in the ICU and three patients’ deaths (B, D, E) were possibly related to MDR-Ab infection (Table 1).

A total of 22 MDR-Ab isolates were available from seven patients who underwent bronchofiberscopy and from four patients who did not undergo bronchofiberscopy. All of the MDR-Ab isolates were completely resistant to cefuroxime, cefotaxime, imipenem, and levofloxacin but were susceptible to colistin. REP-PCR analysis revealed that 16 MDR-Ab isolates from six patients who received bronchofiberscopy shared an identical outbreak strain of MDR-Ab (genotype A) that differed from the isolates (genotypes C, D, G) obtained from three of the patients who did not undergo bronchofiberscopy (Figure 4, Table 1). Four patients who received bronchofiberscopy during the epidemic period did not test positive for MDR-Ab.

Environmental sampling showed that 26 MDR-Ab isolates were identified from 78 environmental samples on 15^th October 2009, whereas no MDR-Ab isolates were detected in the later environmental samples on 21^th and 28^th October 2009. A total of 22 of 26 MDR-Ab isolates (84.6%) were identical to the outbreak strain. Of these 22 isolates, four were recovered from the biopsy forceps and the bronchofiberscope tip that were used to treat a patient carrying the outbreak strain and from the bronchofiberscope surface after reprocessing within the ICU. Of the 22 isolates, 13 were recovered from the bed sheets, bedrails, dispensing table, and invigilator or blood filtering machine keyboard of beds 1–3 and 6–8, findings that were in line with the distribution of the outbreak strains within the patients. Five of 22 isolates were detected in the scrub sink, the medical treatment room, and the nurses’ notebook, desk, and calculator (Figure 3). However, no MDR-Ab isolates were detected from the healthcare workers’ hand or nasal cavity samples.

Further investigation disclosed several potential administrative and technical problems in the bronchofiberscope reprocessing protocol. First, from the end of July 2009, bronchofiberscope was frequently reprocessed in the ICU by doctors after emergent patient examinations and treatments. Second, the bronchofiberscope reprocessing procedure was not strictly in accordance with the Chinese guidelines for endoscopy cleaning and disinfection [23]. For instance, the pre-cleaning time was not adequate and the specific enzyme-containing detergent was seldom used. In addition, the patients who received bronchofiberscopy were seldom covered during emergent treatment, and the potentially contaminated environmental surface was not disinfected immediately and thoroughly after the bronchofiberscopy procedure was performed.

Case–control study

Intervention

After the intervention on 21^st October 2009, the cases decreased gradually and no further evidence of the outbreak strain was detected in December 2009 or January 2010. Follow-up ICU environmental surface cultures were processed monthly for 6 months after February 2010 and none grew MDR-Ab.