- Email: [email protected]
Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting
Accepted Manuscript Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting B. Lytsy, R.P.F. Lindblom, U. Ransjö, C. Leo-Swenne PII:
S0195-6701(15)00343-6
DOI:
10.1016/j.jhin.2015.08.021
Reference:
YJHIN 4626
To appear in:
Journal of Hospital Infection
Received Date: 10 June 2015 Accepted Date: 17 August 2015
Please cite this article as: Lytsy B, Lindblom RPF, Ransjö U, Leo-Swenne C, Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting, Journal of Hospital Infection (2015), doi: 10.1016/j.jhin.2015.08.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
B. Lytsy et al.
Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting B. Lytsya,*, R.P.F. Lindblomb, U. Ransjöa, C. Leo-Swenneb,c a
Department of Medical Sciences, Unit for Clinical Microbiology and Infectious Medicine,
Uppsala University, Uppsala, Sweden Department of Cardiothoracic Surgery and Anesthesiology, Uppsala University Hospital,
RI PT
b
Uppsala, Sweden c
Department of Public Health and Caring Sciences, Uppsala University, Uppsala, Sweden
________________________ *
SC
Corresponding author. Address: Uppsala University Hospital (Akademiska sjukhuset),
Department of Clinical Microbiology and Infection Control, Dag Hammarsolds vag 17,
M AN U
Uppsala, S-751 85, Sweden. Tel.: +46 186113903. E-mail address: [email protected] (B. Lytsy). SUMMARY
Background: The department of Cardiothoracic Surgery at Uppsala University Hospital has 25 beds in one to four patient rooms and an operating suite consisting of five operating rooms with ultraclean air. Around 700 open heart (250 isolated coronary artery bypass grafting,
TE D
CABG) operations are performed annually. In 2009, the numbers of deep sternal wound infections (DSWIs) increased to unacceptable rates despite existing hygienic guidelines. Aim: To show how root cause analysis (RCA) followed by quality improvement interventions reduced the rate of DSWI after CABG surgery.
EP
Methods: Only isolated CABG patients requiring surgical revision due to DSWI were included. Swabs and tissue biopsies were taken during surgical revision and analysed with
AC C
standard methods. DSWIs were registered prospectively according to US Centers for Disease Control and Prevention definitions. RCA for infection was performed between September 2009 and April 2010. Interventions based on results of the RCA and on nationally recommended practices were concluded in April 2010, and thought to have taken full effect by July 1st, 2010. Air was actively sampled at ≤0.5 m from the sternal incision. Findings: DSWI incidence rates per CABG operations decreased from 5.1% pre intervention to 0.9% post intervention. Wound cultures pre intervention grew Staphylococcus aureus 27.1% and coagulase negative staphylococcus (CoNS) 47.1%, post intervention S. aureus 23.1% and CoNS 30.8%. Air counts did not exceed 5 cfu/m3. Conclusion: When the aetiology of an error is multifactorial, RCA engaging both the medical professions and the infection control team is a potential tool to map causes leading to adverse
ACCEPTED MANUSCRIPT
events such as healthcare-associated infections. A systematic quality improvement intervention based on the RCA may reduce the number of deep sternal wound infections after CABG surgery. Keywords: Coronary artery bypass graft Deep sternal wound infections
RI PT
Hygiene policy Microbiology Root cause analysis Introduction
SC
Deep sternal wound infection (DSWI) is a major contributing factor to mortality,
morbidity, and costs following CABG.1 Many DSWIs after CABG are due to preventable causes, and risk factors associated with increased risk of DSWI after CABG may be divided
M AN U
into pre-, intra-, and postoperative factors.1,2
In 2009 the infection control team of Uppsala University Hospital (UUH) was contacted by leaders from the Department of Cardiothoracic Surgery and Anesthesiology because the incidence of DSWI during the first six months of 2009 was estimated to be >10% and the majority were caused by Staphylococcus aureus.
TE D
When the aetiology of an error is multifactorial, such as for DSWIs, root cause analysis (RCA) is a useful tool to systematically map causes leading to a specific serious adverse event in healthcare.3,4 The RCA is a structured process for retrospective investigation of causes for a specific outcome and finding solutions to eliminate these causes.
EP
Our aim was to show how RCA followed by quality improvement interventions could reduce the rate of DSWI after CABG surgery.
AC C
Methods Ethics
This study was a quality improvement project approved by the head of the department of cardiothoracic surgery and performed according to the Helsinki Declaration.5 Setting
The Department of Cardiothoracic Surgery and Anesthesiology at UUH performs ~700 open heart surgeries annually; 250‒300 are isolated CABG operations. Patients were admitted one day before surgery and spent the first 12‒24 postoperative hours in the intensive care unit before being transferred to a ward with 25 beds (one to four beds per room) where they stayed for four to seven days postoperatively.
ACCEPTED MANUSCRIPT
Study design
The study included all patients with DSWI after isolated CABG surgery in the department between 2006 and 2012 who were reoperated for retrosternal infection with or without sternal instability. Patients who had undergone CABG with another procedure, e.g. valve replacement, were excluded. Deep and organ/space infections as defined by the US Centers for Disease Control and
RI PT
Prevention was registered prospectively in a text file when a patient was reoperated for
DSWI.6 Data were validated through review of the files of every third isolated CABG patient operated on during the same time-period. The patients were divided into those operated on between January 1st, 2006 and June 30th, 2010 (period I) and those operated on between July
SC
1st, 2010 and December 31st, 2012 (period II). Hygiene policy
M AN U
The hospital dress code included short-sleeved working clothes for all categories of staff; rings and wristwatches were not allowed. An alcohol-based hand disinfectant was to be used before and after patient care, aseptic procedures, donning and removing gloves, and when entering or leaving a patient room. Soap was used before disinfection only when hands were visibly dirty. Gloves were to be used when at risk of touching blood, secretions, excretions, and contaminated or dirty equipment. Gowns or aprons were used for close
TE D
contact with a patient or a patient’s bed. The precautions were the same for patients nursed in single- or four-bedded rooms.
In the operating suite, the dress code was the same as in the wards, but the working clothes were standard clean air suits. Single-use caps or helmets and clean shoes were worn.
EP
Visitors wore non-sterile gowns and shoe covers. Surgical scrub consisted of liquid soap plus isopropanol/N-propanol hand disinfection, or chlorhexidine soap. Masks, sterile gowns, and
AC C
sterile non-powdered gloves were used by surgeons and theatre nurses. The patient performed two whole-body washes during the preoperative 24 h with a 4% chlorhexidine detergent solution, since 2007 one shower in the patient’s home and one shower at the hospital. Hair removal from the chest and leg was done by clipping with a clipping machine with single-use shears on the night before surgery. The surgical field was disinfected with chlorhexidine 5 mg/mL in 70% ethanol by the operating room nurse. The operating rooms were large and equipped with downflow ultraclean air at ~80 air changes/h. Instruments were cleaned in validated washer-disinfectors and steam-sterilized in a separate disinfection room within the operating suite.
ACCEPTED MANUSCRIPT
Root cause analysis
For conducting the RCA, three groups were formed, each consisting of head nurse, ward registrar, hygiene link nurse, and at least two other members of staff from the preoperative, operative, and intensive care wards, respectively.7 All categories of staff were engaged: nurses, nurses’ aides, perfusionists, surgeons, and anaesthesiologists. Representatives from the infection control team attended all meetings. The groups were
RI PT
chaired by the deputy hospital director and met at regular intervals between September 2009 and April 2010. All activities of pre-, intra-, and postoperative care were systematically analysed. Possible errors were identified and judged according to severity (1‒4 points),
probability that the error would occur (1‒4 points), and probability of the error to be detected
SC
(1‒4 points). An error could score a minimum of 1 point (1×1×1 = 1) and a maximum of 64 points (4×4×4 = 64). Errors exceeding 8 points were analysed in detail and possible reasons
M AN U
for the error to occur were examined. Possible actions to prevent these errors were selected for the action plan. Microbiology
Swabs and up to six tissue biopsies were taken during operations for suspected infection.8 Biopsies were grown in anaerobic broth and analysed after seven days. Alginate swabs were cultured aerobically and anaerobically, and analysed after two days. Bacteria
TE D
belonging to the groups Staphylococcus, Enterobacteriaceae, Streptococcus, and Propionibacterium were identified according to standard methods. Airborne bacteria-carrying particles (colony-forming units, cfu) were monitored during CABG surgery with a Sartorius™ MD8 membrane impactor (gelatin membrane filter pore
EP
size ≤3 µm, flow rate 6 m3/h, sampling time 10 min per filter) at ≤0.5 m from the sternal incision. The total number of colony-forming units (cfu) per plate was counted and the
AC C
concentration stated as the number of cfu/m3. The micro-organisms were phenotyped and identified according to standard methods. Patients and patient risk factors Patients undergoing isolated CABG were identified from the computerized operating lists of the department.
In a parallel study (R.P.F. Lindblom et al., unpublished data), the files of every third patient operated on with isolated CABG were analysed for a large number of risk factors, of which only diabetes was identified as important for infection. Their non-infected patients were used as our controls and pre- and intraoperative risk factors for infection [age, sex M/F, body mass index, diabetes, haemoglobin, serum creatinine, chronic obstructive pulmonary
ACCEPTED MANUSCRIPT
disease (COPD), corticosteroid treatment, smoking, number of coronary artery grafts] were analysed for all DSWI cases. Statistical analysis Microbiology data and demographic data in the two periods were analysed with nonparametric methods by Statistica 12 (www.statsoft.com). Results
RI PT
Root cause analysis
Major errors and corresponding actions resulting in permanent changes of routines and procedures are summarized below. The initial brainstorming phase of the preoperative group identified 131 preoperative errors; 37 actions to correct theses errors were adopted for the
SC
action plan. The intraoperative group identified 134 errors and 80 actions. The postoperative group identified 104 errors and 47 actions. In total, 369 errors were identified and the final
M AN U
action plan contained 164 actions. The action plan was launched with support from the leaders of the department; persons were named for each action and an endpoint was set for all interventions. Interventions
A new temporary leader of the department introduced the concept ‘zero-tolerance towards DSWIs’ and did not tolerate any surgeon or nurse to go on with ‘business as usual’.
TE D
All surgeons and staff members had to accept the rules, change behaviour, and comply with hygiene policies. The leader made it very clear that individual solutions deviating from these routines were unacceptable. All planned surgery was paused for one week in April 2010 to give time for the changes to settle.
EP
All categories of staff including doctors were educated by the infection control team about the causes and prevention of surgical site infections with focus on DSWI after CABG.
AC C
An infection control audit was undertaken in the pre- and postoperative wards during December 2009.
Preoperative procedures
In the preoperative ward, routines for hand disinfection, glove use, and dress code had deteriorated. Training in hand hygiene and in the use of gloves and protective aprons was introduced for all categories including doctors. Clocks were put up in all rooms and alcohol hand-rub dispensers were placed at every patient bed. Housekeeping routines were improved in the ward cleaning and storage rooms. From 2010 special attention was paid to patients with increased risks for DSWI (COPD, obesity, diabetes mellitus, heart and renal failure, steroid medication) with careful monitoring of pre-, intra-, and postoperative procedures.
ACCEPTED MANUSCRIPT
All patients were enrolled to a named surgeon responsible for the patient from admission to discharge, including the CABG operation. All cases of DSWI were reported at staff meetings for surgeons and anaesthesiologists, and the responsible surgeon performed the RCA for each case. From 2010 the preoperative hospital stay was shortened to less than one day. Routines for patient preoperative washing procedures were revised and communicated to all
RI PT
personnel at the preoperative ward. The first wash, performed in the patient’s home, was
repeated twice and included a chlorhexidine shampoo. A checklist, signed by the nurses’ aide responsible for helping the patient, was reintroduced.
During 2006‒2010 (period I) cloxacillin, 2 g was administered three times a day by
SC
intravenous infusion for two days, starting at the induction of anaesthesia. During 2010‒2012 (period II) cloxacillin 2 g was given every fourth hour for 24 h, with the first dose given
Intraoperative procedures
M AN U
30‒60 min before the skin incision.
The surgical procedures were performed in an operating room with ultraclean air (<10 cfu/m3). The microbial cleanliness of the operating room air was examined with Sartorius filtration and found to be adequate.
Intraoperative routines were revised and included prophylaxis procedures, dress code,
TE D
hand hygiene, and the introduction of double gloves for all surgeons. The World Health Organization checklist for safe surgery was introduced. In the operating theatre, door-openings were restricted and the number of people was minimized. Students were not allowed.
EP
Mechanical skin disinfection of the operating site with 0.5% chlorhexidine in 70% ethanol followed by evaporation was emphasized. After the closure of the wound, the skin
AC C
was again disinfected with chlorhexidine in ethanol. Cleaning procedures for the operating room were revised. Disinfectable keyboards were purchased. Function controls of the instrument washer-disinfector were introduced, and instruments that could not be adequately cleaned were replaced by single-use items. Postoperative procedures
Routines for blood glucose control were improved. Registration of wound infections through self-reporting by patients was enhanced and results reported at staff meetings. In both periods dressings were allowed to stay in place for three days if dry. On day 4 all dressings were removed and the patient showered. In period I, dressings and swabs were stored on a trolley which was wheeled in and out between patients, whereas in period II a
ACCEPTED MANUSCRIPT
fresh tray was prepared for each patient. Clean dry wounds were then covered with a polyurethane film which stayed in place for one week. Infections The results are divided into period I (1 January 2006 to 30 June 2010) and period II (1 July 2010 to 31 December 2012) (Figure 1). The sample of patients whose files were reviewed (N = 503; age 67 ± 9 years; M/F: 410/93) were comparable to the others operated on
RI PT
during the study period (N = 1137; age 67 ± 9 years; M/F: 923/241).
The number of infections in period I was 80 out of 1695 CABG operations (4.72%) and in period II 13 out of 674 (1.9%). The risk of infection was thus 2.44 times higher in period I than in period II (confidence interval: 1.37‒4.37, P = 0.002). Infection rates per year (six-
SC
month period) as reported to the operating unit are shown in Figure 1. Numbers of isolated CABG operations declined from 429 in 2006 to 235 in 2012 due to changes in operating
M AN U
policy from CABG to percutaneous coronary interventions. Infection rates were lowered from 5.1% in 2006 to 0.9% in 2012. The decline was continuous over time (Figure 1). Microbiology Wound cultures
In wound cultures, S. aureus and coagulase-negative staphylococci (CoNS) were the most frequently isolated bacteria, both from swabs and from tissue biopsies. The CoNS were
meticillin resistant.
TE D
not further typed. All S. aureus were meticillin susceptible, and the majority of CoNS were
In period I (January 2006 to June 2010) 67 of 77 infected patients were cultured (52 with tissue biopsy). Fifty-seven swabs were positive, 17 (25.4%) had S. aureus and 32 (47.8%)
EP
had CoNS. Forty-six tissue biopsies were positive, 15 (29.6%) grew S. aureus, and 18 (34.6%) grew CoNS (Table I).
AC C
In period II (July 2010 to December 2012) 12 of 12 patients were cultured (eight with tissue biopsy). Eleven swabs were positive, three (25%) had S. aureus and four (33%) had CoNS. Five tissue biopsies were positive, two (25%) grew S. aureus, and five (62.5%) grew CoNS.
In both periods, Gram-negative organisms were few. Differences between periods I and II in culture results from tissue biopsies or swabs were not statistically significant for S. aureus, CoNS or other organisms when compared by Kruskal‒Wallis analysis of variance by ranks. Air sampling
ACCEPTED MANUSCRIPT
Air sampling was carried out as part of the audit in period I. As an example, in one of the operating rooms with a downflow of ~0.4 m3/s in the central zone; median air counts during five operations were 0–3.5 cfu/m3. During period II, air was sampled in two of the six operating rooms. In both rooms median air counts did not exceed 1 cfu/m3. Patient risk factors
RI PT
Significant risk factors by Spearman rank correlations for DSWI as compared to
controls were high body mass index, smoking, and COPD, but not diabetes. The DSWI
patients did not differ significantly in Kruskal‒Wallis analysis between the two periods with respect to age, sex, body mass index, COPD, diabetes, or smoking.
SC
Discussion
Multiple factors, both patient- and procedural-related, predispose to acquiring postoperative wound infection after open heart surgery.9,10 Several measures to minimize the
M AN U
incidence of DSWI have been tried, with variable results.10 Implementation of bundled interventions against surgical site infections has now been reported to decrease DSWI following cardiovascular surgery.11
Most deep surgical site infections originate in the operating room. The epidemiology and control of S. aureus and CoNS, the main pathogens in thoracic surgery, has been
TE D
extensively studied in Sweden.12‒14 Infection rates following CABG had been lowered in the department before this study, but not sustainably. Routines for hygienic and metabolic control had slackened, for several reasons including lack of attention to detail, inadequate adherence to hygiene guidelines, and poor awareness of risks.
EP
A structured RCA, followed by interventions, more than halved infection rates. Skin micro-organisms, mainly staphylococci, were the dominant pathogens in periods I and II. One
AC C
week of ceased operative activity marked the endpoint of the interventions. This break in daily productivity also marked the importance of the intervention programme to the leadership.
Why did the infection rates for DSWI decrease? Major changes in routines were few. The environment in the operating room was optimal throughout both periods. More attention was paid to hand hygiene and the use of gloves, to patients’ skin disinfection with chlorhexidine soap and chlorhexidine alcohol, and to the timing of the isoxazoyl‒penicillin prophylaxis. Preoperative optimizing of the patient was put into focus, but the hospitals referring the patients were not reached, so all groups were similar concerning smoking, obesity, and COPD (R.P.F. Lindblom et al., unpublished data). Preoperative diabetes control seems to have been adequate, as diagnosed diabetes was not a risk factor for DSWI. There
ACCEPTED MANUSCRIPT
may have been a relative decrease in S. aureus infections due to more efficient skin disinfection, but this could not be demonstrated as patients in period II were so few. The main reason for the decline was that all categories of staff were actively involved, whereby they realized that the problems were multifactorial.12 The interventions were produced by a democratic process and strong leadership. Surgical technique and competence was crucial, and enough time had to be allotted for necessary procedures and controls. To
RI PT
achieve persistent positive effects, a monitoring programme with regular feedback to all levels of staff is needed.15
Good facilities and ventilation in the operating room cannot guarantee low infection rates. Guidelines are necessary, but experience from our study shows that a systematic
SC
approach to identify and adjust possible errors, a teamwork approach during the process and strong leadership among all professions is needed to ensure compliance to guidelines.
M AN U
In conclusion, when the aetiology of an error is multifactorial, RCA engaging both the medical professions and the infection control team is a potential tool to map causes leading to adverse events such as healthcare-associated infections. A systematic quality improvement intervention based on the RCA can reduce the number of DSWIs after CABG surgery. Conflict of interest statement None declared.
TE D
Funding sources None. References 1.
Toumpoulis IK, Anagnostopoulos CE, Derose JJ, Jr, Swistel DG. The impact of deep
EP
sternal wound infection on long-term survival after coronary artery bypass grafting. Chest 2005;127:464‒471.
Kubota H, Miyata H, Motomura N, et al. Deep sternal wound infection after cardiac
AC C
2.
surgery. J Cardiothorac Surg 2013;8:132. 3.
Iedema RA, Jorm C, Braithwaite J, Travaglia J, Lum M. A root cause analysis of clinical error: confronting the disjunction between formal rules and situated clinical activity. Social Sci Med 2006;63:1201‒1212.
4.
Nicolini D, Waring J, Mengis J. Policy and practice in the use of root cause analysis to investigate clinical adverse events: mind the gap. Social Sci Med 2011;73:217‒225.
5.
World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 2013;310:2191‒2194.
6.
ACCEPTED MANUSCRIPT
Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health careassociated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309‒332.
7.
Bagian JP, Lee C, Gosbee J, et al. Developing and deploying a patient safety program in a large health care delivery system: you can’t fix what you don’t know about. J Comm J Qual Improvmt 2001;27:522‒532. Tammelin A, Hambraeus A, Stahle E. Mediastinitis after cardiac surgery: improvement
RI PT
8.
of bacteriological diagnosis by use of multiple tissue samples and strain typing. J Clin Microbiol 2002;40:2936‒2941. 9.
Roy MC, Herwaldt LA, Embrey R, Kuhns K, Wenzel RP, Perl TM. Does the Centers for
SC
Disease Control’s NNIS system risk index stratify patients undergoing cardiothoracic operations by their risk of surgical-site infection? Infect Control Hosp Epidemiol
M AN U
2000;21:186‒190.
10. Bryan CS, Yarbrough WM. Preventing deep wound infection after coronary artery bypass grafting: a review. Texas Heart Inst J 2013;40:125‒139. 11. Miyahara K, Matsuura A, Takemura H, Mizutani S, Saito S, Toyama M. Implementation of bundled interventions greatly decreases deep sternal wound infection following cardiovascular surgery. J Thoracic Cardiovasc Surg 2014;148:2381‒2388.
TE D
12. Gardlund B, Bitkover CY, Vaage J. Postoperative mediastinitis in cardiac surgery ‒ microbiology and pathogenesis. Eur J Cardiothoracic Surg 2002;21:825‒830. 13. Bitkover CY, Marcusson E, Ransjo U. Spread of coagulase-negative staphylococci during cardiac operations in a modern operating room. Ann Thorac Surg
EP
2000;69:1110‒1115.
14. Tammelin A, Klotz F, Hambraeus A, Stahle E, Ransjo U. Nasal and hand carriage of
AC C
Staphylococcus aureus in staff at a department for thoracic and cardiovascular surgery: endogenous or exogenous source? Infect Control Hosp Epidemiol 2003;24:686‒689. 15. Brandt C, Sohr D, Behnke M, Daschner F, Ruden H, Gastmeier P. Reduction of surgical site infection rates associated with active surveillance. Infect Control Hosp Epidemiol 2006;27:1347‒1351.
ACCEPTED MANUSCRIPT
Table I
Period II
(January 2006 to June 2009)
(July 2009 to December 2012)
Tissue finding 1 Swab finding 1
Tissue finding 1 Swab finding 1
15 (29%)
17 (25%)
2 (25%)
Coagulase-negative staphylococcus 18 (35%)
32 (48%)
5 (62%)
Anaerobe Gram-positive rods
7 (13%)
3 (4%)
1 (12%)
Aerobe Gram-positive rods
0
0
0
1 (8%)
Enterobacteriaceae
3 (6%)
5 (7%)
0
Enterococcus faecalis
1 (2%)
0
0
Enterococcus faecium
0
0
0
Serratia spp.
0
0
0
0
Candida spp.
2 (4%)
0
0
0
Negative
6 (12%)
10 (15%)
0
1 (8%)
No. of patients cultured
52
67
8
12
Total positive
46 (88%)
57 (85%)
8 (100%)
11 (92%)
3 (25%) 4 (33%) 2 (17%)
M AN U
TE D
Staphylococcus aureus
EP
Micro-organsims
RI PT
Period I
SC
Cultures from wounds of patients with deep sternal wound infections after isolated coronary artery bypass grafting during reoperation
1 (8%) 0 0
AC C
Finding 1 is the dominant and finding 2 is the second most frequent organism type in each culture. Selective media for staphylococci, enterococci, and Gram-negatives were used for each specimen.
ACCEPTED MANUSCRIPT Figure 1. Rates of deep sternal wound infections (red lines) and numbers of coronary artery bypass graft (CABG) operations per half-year (blue bars) (gap inserted between periods I and
AC C
EP
TE D
M AN U
SC
RI PT
II).
S0195-6701(15)00343-6
DOI:
10.1016/j.jhin.2015.08.021
Reference:
YJHIN 4626
To appear in:
Journal of Hospital Infection
Received Date: 10 June 2015 Accepted Date: 17 August 2015
Please cite this article as: Lytsy B, Lindblom RPF, Ransjö U, Leo-Swenne C, Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting, Journal of Hospital Infection (2015), doi: 10.1016/j.jhin.2015.08.021. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
B. Lytsy et al.
Hygienic interventions to decrease deep sternal wound infections following coronary artery bypass grafting B. Lytsya,*, R.P.F. Lindblomb, U. Ransjöa, C. Leo-Swenneb,c a
Department of Medical Sciences, Unit for Clinical Microbiology and Infectious Medicine,
Uppsala University, Uppsala, Sweden Department of Cardiothoracic Surgery and Anesthesiology, Uppsala University Hospital,
RI PT
b
Uppsala, Sweden c
Department of Public Health and Caring Sciences, Uppsala University, Uppsala, Sweden
________________________ *
SC
Corresponding author. Address: Uppsala University Hospital (Akademiska sjukhuset),
Department of Clinical Microbiology and Infection Control, Dag Hammarsolds vag 17,
M AN U
Uppsala, S-751 85, Sweden. Tel.: +46 186113903. E-mail address: [email protected] (B. Lytsy). SUMMARY
Background: The department of Cardiothoracic Surgery at Uppsala University Hospital has 25 beds in one to four patient rooms and an operating suite consisting of five operating rooms with ultraclean air. Around 700 open heart (250 isolated coronary artery bypass grafting,
TE D
CABG) operations are performed annually. In 2009, the numbers of deep sternal wound infections (DSWIs) increased to unacceptable rates despite existing hygienic guidelines. Aim: To show how root cause analysis (RCA) followed by quality improvement interventions reduced the rate of DSWI after CABG surgery.
EP
Methods: Only isolated CABG patients requiring surgical revision due to DSWI were included. Swabs and tissue biopsies were taken during surgical revision and analysed with
AC C
standard methods. DSWIs were registered prospectively according to US Centers for Disease Control and Prevention definitions. RCA for infection was performed between September 2009 and April 2010. Interventions based on results of the RCA and on nationally recommended practices were concluded in April 2010, and thought to have taken full effect by July 1st, 2010. Air was actively sampled at ≤0.5 m from the sternal incision. Findings: DSWI incidence rates per CABG operations decreased from 5.1% pre intervention to 0.9% post intervention. Wound cultures pre intervention grew Staphylococcus aureus 27.1% and coagulase negative staphylococcus (CoNS) 47.1%, post intervention S. aureus 23.1% and CoNS 30.8%. Air counts did not exceed 5 cfu/m3. Conclusion: When the aetiology of an error is multifactorial, RCA engaging both the medical professions and the infection control team is a potential tool to map causes leading to adverse
ACCEPTED MANUSCRIPT
events such as healthcare-associated infections. A systematic quality improvement intervention based on the RCA may reduce the number of deep sternal wound infections after CABG surgery. Keywords: Coronary artery bypass graft Deep sternal wound infections
RI PT
Hygiene policy Microbiology Root cause analysis Introduction
SC
Deep sternal wound infection (DSWI) is a major contributing factor to mortality,
morbidity, and costs following CABG.1 Many DSWIs after CABG are due to preventable causes, and risk factors associated with increased risk of DSWI after CABG may be divided
M AN U
into pre-, intra-, and postoperative factors.1,2
In 2009 the infection control team of Uppsala University Hospital (UUH) was contacted by leaders from the Department of Cardiothoracic Surgery and Anesthesiology because the incidence of DSWI during the first six months of 2009 was estimated to be >10% and the majority were caused by Staphylococcus aureus.
TE D
When the aetiology of an error is multifactorial, such as for DSWIs, root cause analysis (RCA) is a useful tool to systematically map causes leading to a specific serious adverse event in healthcare.3,4 The RCA is a structured process for retrospective investigation of causes for a specific outcome and finding solutions to eliminate these causes.
EP
Our aim was to show how RCA followed by quality improvement interventions could reduce the rate of DSWI after CABG surgery.
AC C
Methods Ethics
This study was a quality improvement project approved by the head of the department of cardiothoracic surgery and performed according to the Helsinki Declaration.5 Setting
The Department of Cardiothoracic Surgery and Anesthesiology at UUH performs ~700 open heart surgeries annually; 250‒300 are isolated CABG operations. Patients were admitted one day before surgery and spent the first 12‒24 postoperative hours in the intensive care unit before being transferred to a ward with 25 beds (one to four beds per room) where they stayed for four to seven days postoperatively.
ACCEPTED MANUSCRIPT
Study design
The study included all patients with DSWI after isolated CABG surgery in the department between 2006 and 2012 who were reoperated for retrosternal infection with or without sternal instability. Patients who had undergone CABG with another procedure, e.g. valve replacement, were excluded. Deep and organ/space infections as defined by the US Centers for Disease Control and
RI PT
Prevention was registered prospectively in a text file when a patient was reoperated for
DSWI.6 Data were validated through review of the files of every third isolated CABG patient operated on during the same time-period. The patients were divided into those operated on between January 1st, 2006 and June 30th, 2010 (period I) and those operated on between July
SC
1st, 2010 and December 31st, 2012 (period II). Hygiene policy
M AN U
The hospital dress code included short-sleeved working clothes for all categories of staff; rings and wristwatches were not allowed. An alcohol-based hand disinfectant was to be used before and after patient care, aseptic procedures, donning and removing gloves, and when entering or leaving a patient room. Soap was used before disinfection only when hands were visibly dirty. Gloves were to be used when at risk of touching blood, secretions, excretions, and contaminated or dirty equipment. Gowns or aprons were used for close
TE D
contact with a patient or a patient’s bed. The precautions were the same for patients nursed in single- or four-bedded rooms.
In the operating suite, the dress code was the same as in the wards, but the working clothes were standard clean air suits. Single-use caps or helmets and clean shoes were worn.
EP
Visitors wore non-sterile gowns and shoe covers. Surgical scrub consisted of liquid soap plus isopropanol/N-propanol hand disinfection, or chlorhexidine soap. Masks, sterile gowns, and
AC C
sterile non-powdered gloves were used by surgeons and theatre nurses. The patient performed two whole-body washes during the preoperative 24 h with a 4% chlorhexidine detergent solution, since 2007 one shower in the patient’s home and one shower at the hospital. Hair removal from the chest and leg was done by clipping with a clipping machine with single-use shears on the night before surgery. The surgical field was disinfected with chlorhexidine 5 mg/mL in 70% ethanol by the operating room nurse. The operating rooms were large and equipped with downflow ultraclean air at ~80 air changes/h. Instruments were cleaned in validated washer-disinfectors and steam-sterilized in a separate disinfection room within the operating suite.
ACCEPTED MANUSCRIPT
Root cause analysis
For conducting the RCA, three groups were formed, each consisting of head nurse, ward registrar, hygiene link nurse, and at least two other members of staff from the preoperative, operative, and intensive care wards, respectively.7 All categories of staff were engaged: nurses, nurses’ aides, perfusionists, surgeons, and anaesthesiologists. Representatives from the infection control team attended all meetings. The groups were
RI PT
chaired by the deputy hospital director and met at regular intervals between September 2009 and April 2010. All activities of pre-, intra-, and postoperative care were systematically analysed. Possible errors were identified and judged according to severity (1‒4 points),
probability that the error would occur (1‒4 points), and probability of the error to be detected
SC
(1‒4 points). An error could score a minimum of 1 point (1×1×1 = 1) and a maximum of 64 points (4×4×4 = 64). Errors exceeding 8 points were analysed in detail and possible reasons
M AN U
for the error to occur were examined. Possible actions to prevent these errors were selected for the action plan. Microbiology
Swabs and up to six tissue biopsies were taken during operations for suspected infection.8 Biopsies were grown in anaerobic broth and analysed after seven days. Alginate swabs were cultured aerobically and anaerobically, and analysed after two days. Bacteria
TE D
belonging to the groups Staphylococcus, Enterobacteriaceae, Streptococcus, and Propionibacterium were identified according to standard methods. Airborne bacteria-carrying particles (colony-forming units, cfu) were monitored during CABG surgery with a Sartorius™ MD8 membrane impactor (gelatin membrane filter pore
EP
size ≤3 µm, flow rate 6 m3/h, sampling time 10 min per filter) at ≤0.5 m from the sternal incision. The total number of colony-forming units (cfu) per plate was counted and the
AC C
concentration stated as the number of cfu/m3. The micro-organisms were phenotyped and identified according to standard methods. Patients and patient risk factors Patients undergoing isolated CABG were identified from the computerized operating lists of the department.
In a parallel study (R.P.F. Lindblom et al., unpublished data), the files of every third patient operated on with isolated CABG were analysed for a large number of risk factors, of which only diabetes was identified as important for infection. Their non-infected patients were used as our controls and pre- and intraoperative risk factors for infection [age, sex M/F, body mass index, diabetes, haemoglobin, serum creatinine, chronic obstructive pulmonary
ACCEPTED MANUSCRIPT
disease (COPD), corticosteroid treatment, smoking, number of coronary artery grafts] were analysed for all DSWI cases. Statistical analysis Microbiology data and demographic data in the two periods were analysed with nonparametric methods by Statistica 12 (www.statsoft.com). Results
RI PT
Root cause analysis
Major errors and corresponding actions resulting in permanent changes of routines and procedures are summarized below. The initial brainstorming phase of the preoperative group identified 131 preoperative errors; 37 actions to correct theses errors were adopted for the
SC
action plan. The intraoperative group identified 134 errors and 80 actions. The postoperative group identified 104 errors and 47 actions. In total, 369 errors were identified and the final
M AN U
action plan contained 164 actions. The action plan was launched with support from the leaders of the department; persons were named for each action and an endpoint was set for all interventions. Interventions
A new temporary leader of the department introduced the concept ‘zero-tolerance towards DSWIs’ and did not tolerate any surgeon or nurse to go on with ‘business as usual’.
TE D
All surgeons and staff members had to accept the rules, change behaviour, and comply with hygiene policies. The leader made it very clear that individual solutions deviating from these routines were unacceptable. All planned surgery was paused for one week in April 2010 to give time for the changes to settle.
EP
All categories of staff including doctors were educated by the infection control team about the causes and prevention of surgical site infections with focus on DSWI after CABG.
AC C
An infection control audit was undertaken in the pre- and postoperative wards during December 2009.
Preoperative procedures
In the preoperative ward, routines for hand disinfection, glove use, and dress code had deteriorated. Training in hand hygiene and in the use of gloves and protective aprons was introduced for all categories including doctors. Clocks were put up in all rooms and alcohol hand-rub dispensers were placed at every patient bed. Housekeeping routines were improved in the ward cleaning and storage rooms. From 2010 special attention was paid to patients with increased risks for DSWI (COPD, obesity, diabetes mellitus, heart and renal failure, steroid medication) with careful monitoring of pre-, intra-, and postoperative procedures.
ACCEPTED MANUSCRIPT
All patients were enrolled to a named surgeon responsible for the patient from admission to discharge, including the CABG operation. All cases of DSWI were reported at staff meetings for surgeons and anaesthesiologists, and the responsible surgeon performed the RCA for each case. From 2010 the preoperative hospital stay was shortened to less than one day. Routines for patient preoperative washing procedures were revised and communicated to all
RI PT
personnel at the preoperative ward. The first wash, performed in the patient’s home, was
repeated twice and included a chlorhexidine shampoo. A checklist, signed by the nurses’ aide responsible for helping the patient, was reintroduced.
During 2006‒2010 (period I) cloxacillin, 2 g was administered three times a day by
SC
intravenous infusion for two days, starting at the induction of anaesthesia. During 2010‒2012 (period II) cloxacillin 2 g was given every fourth hour for 24 h, with the first dose given
Intraoperative procedures
M AN U
30‒60 min before the skin incision.
The surgical procedures were performed in an operating room with ultraclean air (<10 cfu/m3). The microbial cleanliness of the operating room air was examined with Sartorius filtration and found to be adequate.
Intraoperative routines were revised and included prophylaxis procedures, dress code,
TE D
hand hygiene, and the introduction of double gloves for all surgeons. The World Health Organization checklist for safe surgery was introduced. In the operating theatre, door-openings were restricted and the number of people was minimized. Students were not allowed.
EP
Mechanical skin disinfection of the operating site with 0.5% chlorhexidine in 70% ethanol followed by evaporation was emphasized. After the closure of the wound, the skin
AC C
was again disinfected with chlorhexidine in ethanol. Cleaning procedures for the operating room were revised. Disinfectable keyboards were purchased. Function controls of the instrument washer-disinfector were introduced, and instruments that could not be adequately cleaned were replaced by single-use items. Postoperative procedures
Routines for blood glucose control were improved. Registration of wound infections through self-reporting by patients was enhanced and results reported at staff meetings. In both periods dressings were allowed to stay in place for three days if dry. On day 4 all dressings were removed and the patient showered. In period I, dressings and swabs were stored on a trolley which was wheeled in and out between patients, whereas in period II a
ACCEPTED MANUSCRIPT
fresh tray was prepared for each patient. Clean dry wounds were then covered with a polyurethane film which stayed in place for one week. Infections The results are divided into period I (1 January 2006 to 30 June 2010) and period II (1 July 2010 to 31 December 2012) (Figure 1). The sample of patients whose files were reviewed (N = 503; age 67 ± 9 years; M/F: 410/93) were comparable to the others operated on
RI PT
during the study period (N = 1137; age 67 ± 9 years; M/F: 923/241).
The number of infections in period I was 80 out of 1695 CABG operations (4.72%) and in period II 13 out of 674 (1.9%). The risk of infection was thus 2.44 times higher in period I than in period II (confidence interval: 1.37‒4.37, P = 0.002). Infection rates per year (six-
SC
month period) as reported to the operating unit are shown in Figure 1. Numbers of isolated CABG operations declined from 429 in 2006 to 235 in 2012 due to changes in operating
M AN U
policy from CABG to percutaneous coronary interventions. Infection rates were lowered from 5.1% in 2006 to 0.9% in 2012. The decline was continuous over time (Figure 1). Microbiology Wound cultures
In wound cultures, S. aureus and coagulase-negative staphylococci (CoNS) were the most frequently isolated bacteria, both from swabs and from tissue biopsies. The CoNS were
meticillin resistant.
TE D
not further typed. All S. aureus were meticillin susceptible, and the majority of CoNS were
In period I (January 2006 to June 2010) 67 of 77 infected patients were cultured (52 with tissue biopsy). Fifty-seven swabs were positive, 17 (25.4%) had S. aureus and 32 (47.8%)
EP
had CoNS. Forty-six tissue biopsies were positive, 15 (29.6%) grew S. aureus, and 18 (34.6%) grew CoNS (Table I).
AC C
In period II (July 2010 to December 2012) 12 of 12 patients were cultured (eight with tissue biopsy). Eleven swabs were positive, three (25%) had S. aureus and four (33%) had CoNS. Five tissue biopsies were positive, two (25%) grew S. aureus, and five (62.5%) grew CoNS.
In both periods, Gram-negative organisms were few. Differences between periods I and II in culture results from tissue biopsies or swabs were not statistically significant for S. aureus, CoNS or other organisms when compared by Kruskal‒Wallis analysis of variance by ranks. Air sampling
ACCEPTED MANUSCRIPT
Air sampling was carried out as part of the audit in period I. As an example, in one of the operating rooms with a downflow of ~0.4 m3/s in the central zone; median air counts during five operations were 0–3.5 cfu/m3. During period II, air was sampled in two of the six operating rooms. In both rooms median air counts did not exceed 1 cfu/m3. Patient risk factors
RI PT
Significant risk factors by Spearman rank correlations for DSWI as compared to
controls were high body mass index, smoking, and COPD, but not diabetes. The DSWI
patients did not differ significantly in Kruskal‒Wallis analysis between the two periods with respect to age, sex, body mass index, COPD, diabetes, or smoking.
SC
Discussion
Multiple factors, both patient- and procedural-related, predispose to acquiring postoperative wound infection after open heart surgery.9,10 Several measures to minimize the
M AN U
incidence of DSWI have been tried, with variable results.10 Implementation of bundled interventions against surgical site infections has now been reported to decrease DSWI following cardiovascular surgery.11
Most deep surgical site infections originate in the operating room. The epidemiology and control of S. aureus and CoNS, the main pathogens in thoracic surgery, has been
TE D
extensively studied in Sweden.12‒14 Infection rates following CABG had been lowered in the department before this study, but not sustainably. Routines for hygienic and metabolic control had slackened, for several reasons including lack of attention to detail, inadequate adherence to hygiene guidelines, and poor awareness of risks.
EP
A structured RCA, followed by interventions, more than halved infection rates. Skin micro-organisms, mainly staphylococci, were the dominant pathogens in periods I and II. One
AC C
week of ceased operative activity marked the endpoint of the interventions. This break in daily productivity also marked the importance of the intervention programme to the leadership.
Why did the infection rates for DSWI decrease? Major changes in routines were few. The environment in the operating room was optimal throughout both periods. More attention was paid to hand hygiene and the use of gloves, to patients’ skin disinfection with chlorhexidine soap and chlorhexidine alcohol, and to the timing of the isoxazoyl‒penicillin prophylaxis. Preoperative optimizing of the patient was put into focus, but the hospitals referring the patients were not reached, so all groups were similar concerning smoking, obesity, and COPD (R.P.F. Lindblom et al., unpublished data). Preoperative diabetes control seems to have been adequate, as diagnosed diabetes was not a risk factor for DSWI. There
ACCEPTED MANUSCRIPT
may have been a relative decrease in S. aureus infections due to more efficient skin disinfection, but this could not be demonstrated as patients in period II were so few. The main reason for the decline was that all categories of staff were actively involved, whereby they realized that the problems were multifactorial.12 The interventions were produced by a democratic process and strong leadership. Surgical technique and competence was crucial, and enough time had to be allotted for necessary procedures and controls. To
RI PT
achieve persistent positive effects, a monitoring programme with regular feedback to all levels of staff is needed.15
Good facilities and ventilation in the operating room cannot guarantee low infection rates. Guidelines are necessary, but experience from our study shows that a systematic
SC
approach to identify and adjust possible errors, a teamwork approach during the process and strong leadership among all professions is needed to ensure compliance to guidelines.
M AN U
In conclusion, when the aetiology of an error is multifactorial, RCA engaging both the medical professions and the infection control team is a potential tool to map causes leading to adverse events such as healthcare-associated infections. A systematic quality improvement intervention based on the RCA can reduce the number of DSWIs after CABG surgery. Conflict of interest statement None declared.
TE D
Funding sources None. References 1.
Toumpoulis IK, Anagnostopoulos CE, Derose JJ, Jr, Swistel DG. The impact of deep
EP
sternal wound infection on long-term survival after coronary artery bypass grafting. Chest 2005;127:464‒471.
Kubota H, Miyata H, Motomura N, et al. Deep sternal wound infection after cardiac
AC C
2.
surgery. J Cardiothorac Surg 2013;8:132. 3.
Iedema RA, Jorm C, Braithwaite J, Travaglia J, Lum M. A root cause analysis of clinical error: confronting the disjunction between formal rules and situated clinical activity. Social Sci Med 2006;63:1201‒1212.
4.
Nicolini D, Waring J, Mengis J. Policy and practice in the use of root cause analysis to investigate clinical adverse events: mind the gap. Social Sci Med 2011;73:217‒225.
5.
World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 2013;310:2191‒2194.
6.
ACCEPTED MANUSCRIPT
Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health careassociated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309‒332.
7.
Bagian JP, Lee C, Gosbee J, et al. Developing and deploying a patient safety program in a large health care delivery system: you can’t fix what you don’t know about. J Comm J Qual Improvmt 2001;27:522‒532. Tammelin A, Hambraeus A, Stahle E. Mediastinitis after cardiac surgery: improvement
RI PT
8.
of bacteriological diagnosis by use of multiple tissue samples and strain typing. J Clin Microbiol 2002;40:2936‒2941. 9.
Roy MC, Herwaldt LA, Embrey R, Kuhns K, Wenzel RP, Perl TM. Does the Centers for
SC
Disease Control’s NNIS system risk index stratify patients undergoing cardiothoracic operations by their risk of surgical-site infection? Infect Control Hosp Epidemiol
M AN U
2000;21:186‒190.
10. Bryan CS, Yarbrough WM. Preventing deep wound infection after coronary artery bypass grafting: a review. Texas Heart Inst J 2013;40:125‒139. 11. Miyahara K, Matsuura A, Takemura H, Mizutani S, Saito S, Toyama M. Implementation of bundled interventions greatly decreases deep sternal wound infection following cardiovascular surgery. J Thoracic Cardiovasc Surg 2014;148:2381‒2388.
TE D
12. Gardlund B, Bitkover CY, Vaage J. Postoperative mediastinitis in cardiac surgery ‒ microbiology and pathogenesis. Eur J Cardiothoracic Surg 2002;21:825‒830. 13. Bitkover CY, Marcusson E, Ransjo U. Spread of coagulase-negative staphylococci during cardiac operations in a modern operating room. Ann Thorac Surg
EP
2000;69:1110‒1115.
14. Tammelin A, Klotz F, Hambraeus A, Stahle E, Ransjo U. Nasal and hand carriage of
AC C
Staphylococcus aureus in staff at a department for thoracic and cardiovascular surgery: endogenous or exogenous source? Infect Control Hosp Epidemiol 2003;24:686‒689. 15. Brandt C, Sohr D, Behnke M, Daschner F, Ruden H, Gastmeier P. Reduction of surgical site infection rates associated with active surveillance. Infect Control Hosp Epidemiol 2006;27:1347‒1351.
ACCEPTED MANUSCRIPT
Table I
Period II
(January 2006 to June 2009)
(July 2009 to December 2012)
Tissue finding 1 Swab finding 1
Tissue finding 1 Swab finding 1
15 (29%)
17 (25%)
2 (25%)
Coagulase-negative staphylococcus 18 (35%)
32 (48%)
5 (62%)
Anaerobe Gram-positive rods
7 (13%)
3 (4%)
1 (12%)
Aerobe Gram-positive rods
0
0
0
1 (8%)
Enterobacteriaceae
3 (6%)
5 (7%)
0
Enterococcus faecalis
1 (2%)
0
0
Enterococcus faecium
0
0
0
Serratia spp.
0
0
0
0
Candida spp.
2 (4%)
0
0
0
Negative
6 (12%)
10 (15%)
0
1 (8%)
No. of patients cultured
52
67
8
12
Total positive
46 (88%)
57 (85%)
8 (100%)
11 (92%)
3 (25%) 4 (33%) 2 (17%)
M AN U
TE D
Staphylococcus aureus
EP
Micro-organsims
RI PT
Period I
SC
Cultures from wounds of patients with deep sternal wound infections after isolated coronary artery bypass grafting during reoperation
1 (8%) 0 0
AC C
Finding 1 is the dominant and finding 2 is the second most frequent organism type in each culture. Selective media for staphylococci, enterococci, and Gram-negatives were used for each specimen.
ACCEPTED MANUSCRIPT Figure 1. Rates of deep sternal wound infections (red lines) and numbers of coronary artery bypass graft (CABG) operations per half-year (blue bars) (gap inserted between periods I and
AC C
EP
TE D
M AN U
SC
RI PT
II).