British Journal of Anaesthesia 1992; 68: 216-220
COMMENTARIES
INFECTION CONTROL IN INTENSIVE CARE UNITS: U.K. NATIONAL SURVEY T. J. J. INGLIS, L. J. SPROAT, P. M. HAWKEY AND P. KNAPPETT SUMMARY
KEY WORDS Infect/on: control practices. Intensive care.
The growing complexity of intensive care medicine during the past three decades has been accompanied by an increased potential for nosocomial infection in patients undergoing artificial ventilation of the lungs. A combination of invasive monitoring, multiple therapies and impaired host defences renders these patients uniquely susceptible to nosocomial wound, device-related, respiratory and urinary tract infections [1]. As a result of common-source outbreaks in patients in intensive care units (ICU), strategies have been developed to interrupt cross-infection, including rigorous handwashing [2]. More recently, the role of the patient's own bacterial flora has been recognized as an important source of subsequent infection [3, 4]. Local experience suggested wide variations in the infection control practice of ICU. We therefore undertook a survey of local and other U.K. adult ICU to assess the extent of this variation, identify
METHODS
A questionnaire was designed by microbiologists and anaesthetist members of the Yorkshire Regional Intensive Care Committee. After a pilot survey (the results of which were included in the overall analysis), the confidential questionnaire was circulated to all other adult ICU listed in the Directory of Emergency and Special Care Units, 1989 (CMA Medical Data Ltd, Cambridge), during April-June 1990. Medical Directors were asked to complete the questionnaire, or to arrange for its completion by a senior medical or nursing member of the ICU staff. Non-responding units were followed up by post in July 1990 and by telephone in August 1990. The questionnaire dealt with workload, staffing, architecture and specific infection control practices. RESULTS
Completed questionnaires were received from 246 of 290 (85 %) ICU in the survey. One paediatric ICU had been classified incorrectly in the source manual. Six of the 246 responding units had closed or changed status and therefore were excluded from further analysis. Of the remaining 240 units, 10 were surgical only and 10 were specialist ICU. The size range was two to 16 beds per unit with a most frequent response (m.f.r.) of six beds. Occupancy ranged from < 10 to > 9 0 % (m.f.r. = 61-70%); patients per year from < 100 to 2432 (m.f.r. = 201-300); nursing staff one to 15 per shift (m.f.r. = 4) with a nurse: patient ratio of 1:2-2:1 (m.f.r. = 1:1), and the percentage admissions who underwent tracheal intubation < 10 % to > 90 % (m.f.r. = 71-80%). Facilities There were 33 ICU with no beds in single side wards (range 0-8, m.f.r. = 1). In the majority of units, less than 30% of the beds were in single
TIMOTHY J. J. INGLIS, D.M. ; LOUISE J. SPROAT, R.G.N. ; PETER M.
HAWKEY, M.D. ; Department of Microbiology, University of Leeds, Leeds LS2 9JT. P. KNAPPETT, F.C.ANAES., Bradford Royal
Infirmary, Bradford, W. Yorks. Accepted for Publication: July 10, 1991. Correspondence to T.J.J.I.
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
A completed questionnaire was returned by 246 (85%) intensive care units participating in a national survey of infection control practice in intensive care. Thirty-three units had no provision for isolating patients in single side wards. Sixty percent of responding ICU had fewer than one washbasin per bedspace. Several units reported using ventilator filters or tubing more frequently than is currently recommended. Excessive numbers of catheter urine specimens were sent for laboratory examination by some units. A small number of units used open urinary drainage systems. A significant proportion of ICU had no formally recognized policy on the management of intravascular cannulae. Only 8% of ICU in the U.K. were using a selective decontamination regimen, and nine of these (50 %) had no full time consultant microbiologist available to supervise the recommended microbiological management. The majority of ICU received a regular visit (^ one per week) from a microbiologist. Proposals are made on the development of a specialized infection control service in order to reduce the risk of nosocomial infection in intensive care, and to improve on existing resource management.
inappropriate procedures and suggest priorities for future infection control practice in the ICU.
INFECTION CONTROL IN THE ICU
217
TABLE I. Visits to intensive care units by medical microbiology staff- *Full and part-time medical microbiologist on hospital staff listed in 1990 Hospital Directory [5]
Visit frequency
Number (%) of ICU responding
Number (%) with consultant microbiologist*
Daily 2-4 per week Weekly Less often Only on request Never
53 (22) 48 (20) 22(9) 28 (12) 82 (34) 6(2)
32/53 (60) 20/48 (42) 7/22 (32) 4/28 (14) 14/82 (17) 1/6 (17)
Infection control and microbiology support In response to examples given (the examples used were a patient with meningitis, and an outbreak of salmonellosis), the microbiologist was stated to be the most frequent source of advice by 203 (84 %) and 214 (89%) ICU, respectively. The advice of an infection control nurse was sought also by 192 (80 %) and 189 (79%) units, and infection control laboratory staff were involved in 32 (13%) and 38 (16%) instances. In only six (2.5%) cases did the ICU report no visit from a microbiologist under any circumstance. Five of these hospitals had no permanent consultant microbiologist and the remaining centre had one [5]. Another 82 received a visit only on request (table I). Infections associated with intravascular cannulae The reported use of policies to specify aspects of management for different cannula sites is given in table II. Responses on strategies used to prevent intravascular cannula infections were too varied to allow generalized comment. When intravascular pressures (arterial, pulmonary arterial) were monitored, non-disposable transducers were used in 72/240 (31 %) units and non-disposable domes in 11 units. Water was used between the cannula and transducer membrane in 32 units, using non-disposable transducers, domes, or both. In response to the question of how an i.v. catheter tip culture result of Staphylococcus epidermidis would
Number (%) ICU responding
Interval (days) 1 2 3
73 (32) 62 (27) 20(9) 12(5) 61 (27) 228 (100)
Not at all Other Total
be dealt with, 169/224 (75%) units replied that no treatment would be given; 17 (8%) would use antibiotics and 23 (10%) would treat with reservations. Respiratory support The time interval between ventilator tubing changes reported by units with a specific policy is shown in table III. Ventilator filters were used by 223/237 (94%) units, 165 (70%) of which reported their use in all circuits. Another 36 (16%) used filters in most circuits. The most commonly reported device combined heat-moisture exchange (HME) properties and filtration (used by 168 ICU). This device was used by 161 (80%) of the 201 units using filters on all or most ventilator circuits. Of 154 units using the filter/HME device, 118 (77%) changed them daily. Thirty-three of 159 (21 %) units reported using the device on both inspiratory and expiratory limbs of the ventilator circuit. Urinary tract Open urine drainage systems were reported to be in routine use by four units. Time to replacement of urinary drainage systems varied from 1 to 14 days (m.f.r. = 7 days), but there were 16 categories of qualified response. Meatal care practices varied considerably: 27 different responses were given, with no clear consensus. Routine catheter specimens were sent to the laboratory from "on admission only" to daily (m.f.r. = twice weekly), and the proportion of positive results used as an indication for treatment ranged between < 25 % and > 75 % (m.f.r. < 25%). Surgical wound care A large range of antiseptic agents were reported to be in common use, amongst which were Eusol, Milton, peroxide and alcohol. Agents reported most
TABLE II. Type of ICU policy for management of intravascular cannula
Total number Cannula type
responding
I.v. peripheral I.v. central Intra-arterial Pulmonary artery Parenteral nutrition
187 203 190 170 191
Policy type Written
Verbal
None
74(40%)" 90(44%) 72(38%) 72(42%) 105(55%)
78(42%) 83 (41 %) 88(46%) 72(42%) 66(35%)
35(19%) 30(15%) 30(16%) 26(15%) 20(10%)
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
occupant rooms. The number of washbasins varied from one to 19 per unit (m.f.r. = 5) and 145 (60%) had fewer than one basin per bed. The minimum distance from bed to basin varied from < 1 to > 10 m (m.f.r. > 2 m) and the maximum from > 1 to > 23 m (m.f.r. > 6 m).
TABLE III. Interval between changes of ventilator tubing
BRITISH JOURNAL OF ANAESTHESIA
218 TABLE IV. Strategies for prevention of gastric stress ulcer
Strategy Histamine antagonist Sucralfate Antacids None of above
Number (%) of ICU responding 197 (84) 62 (26) 34 (14) 7(3)
frequently were chlorhexidine-based (used by 80 units (36%)) and iodophor-based (79 (35%)). A total of 34 preparations was reported.
Selective decontamination of the digestive tract was in use in 18 units (8 %), only five of which were teaching hospitals. Nine of these centres had no full time consultant microbiologist. Strategies for prophylaxis of gastric stress ulceration are given in table IV. DISCUSSION
The response rate (85%) in this survey reflects the importance attached to infection control by British Intensive Care Units, and presents an opportunity to review current infection control arrangements. It is held widely that the risk of cross-infection from a given patient may be reduced by source isolation, often achieved by nursing the patient in a single-bedded room under conditions of scrupulous hygiene [6]. Although some authorities regard the provision of single cubicles for all intensive care patients as mandatory [7], a lower proportion of the total bed complement is probably satisfactory. When an intensive care patient presents a particular infection hazard to other patients or staff, source isolation can be achieved by nursing the patient in a side ward and institution of additional barrier-type precautions. It is therefore a matter of concern that 33 units reported having no single-bed side ward. If, as may easily happen in an ICU, handwashing and other barrier-type procedures are violated for reasons of convenience or emergency, the separation of a potentially infectious patient from other susceptible individuals by enclosure in a single side ward may be the only remaining barrier to cross-infection. Handwashing is probably the single most important procedure in the prevention of transmissible disease in hospital inpatients. In the ICU, where patients are more likely to be colonized with potential pathogens, and physical contact with hospital staff is more frequent, high standards of staff hand hygiene are necessary. Yet it has been shown that handwashing rates are poor amongst ICU staff, and medical staff are particularly resistant to educational programmes [8]. This level of compliance with a simple infection control procedure such as handwashing is alarming when the possible effect of a single violation on days or weeks of careful hygiene observance is considered. Nevertheless, ICU staff cannot be held entirely responsible for poor rates of compliance when a unit is poorly equipped for
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
Strategies for the prevention of endogenous source infection
handwashing. Small numbers of washbasins located far from the head end of bedspaces, or obstructed by respiratory support and monitoring equipment, act as a disincentive to regular handwashing. Moreover, washbasins shared between ICU bedspaces present a further route for indirect bacterial transmission between patients. The provision of alcohol hand rub dispensers may improve compliance, but they are not a substitute for washbasins which, unlike hand rub dispensers, cannot be removed. The infection control team (microbiologist, nurse and laboratory staff) have a major responsibility for the prevention and containment of infectious diseases throughout the hospital. It is notable that a majority of ICU receive active support from a medical microbiologist in the form of visits to the unit, and rely on the various components of the infection control team for advice on the management of infection. The very small number of responding units reporting a lack of microbiological support had no consultant medical microbiologist appointed to their hospital, with one exception. Intravascular cannula infections begin either with bacterial colonization of the insertion site, or contamination of the limb. As these devices give direct access to the vascular space, it is surprising that a substantial minority of units reported having no formally recognized policy on cannula management. Formal unit policies for cannula management would be a possible route towards establishing uniformity and quality in standards of intravascular cannula hygiene. Arterial pressure transducers present a specific infection risk, highlighted in a recent review [9]. Non-disposable domes may crack after repeated autoclaving and admit bacteria from the transducer side of the membrane, and have been recognized for some time as a risk to patients [10]. The use of nondisposable transducers with disposable domes may still assist bacterial access to the arterial circulation, because the fluid used in some units to improve membrane-to-transducer coupling may become contaminated. Subsequent dismantling of the transducer from the dome contaminates the fingertips of staff and, if followed by arterial blood sampling, may lead to contamination of the fluid path. Units using either non-disposable domes, or non-disposable transducers/disposable domes (particularly those using an aqueous coupling fluid) may need to review their practice. The current extent of filter use in ventilator circuits cannot be justified fully according to present view on the pathogenesis of ventilator-associated pneumonia—that is, that infecting bacteria are primarily endogenous in origin [11]. However, there is some evidence that Pseudomonas sp. infection may be reduced by use of filters [12]. Compelling evidence for the in vivo efficacy offiltersis noticeably lacking [13], and concern has been expressed over increases in resistance caused by the accumulation of condensate in some filters [14]. There is a need to establish a clear rationale for the use of ventilator filters in independent, controlled trials. The hygienic management of ventilator tubing has already been studied [15]. Daily changes may be
INFECTION CONTROL IN THE ICU
(2) Where there is no clear basis or consensus for a given practice, it should be the subject of independent evaluation, before a decision is made regarding its suitability for incorporation in unit policy or training schemes. (3) Many units require alterations to their architecture, fittings, or both, to enable adequate hand washing and source-isolation of infected patients to be carried out. (4) Given the current contribution of the infection control team to intensive care and the need for a proactive infection control programme in the ICU, a member of intensive care staff should be appointed to work as a member of the infection control team. Smaller units might second a senior member of nursing staff to the infection control team as a link worker, whereas larger ICU may prefer to appoint an infection control sister as a full time specialist.
ACKNOWLEDGEMENTS We thank the Intensive Care Society for providing financial support, and members of the Yorkshire Regional Intensive Care Committee for assistance with the survey in its early stages.
REFERENCES 1. Haley RW, Hooton TM, Culver DH, Stanley RC, Emori TG, Hardison CD, Quade D, Schachtman RH, Schaberg DR, Shah BV, Schatz GD. Nosocomial infections in U.S. hospitals 1975-76. Estimated frequency by selected characteristics of patients. American Journal of Medicine 1981; 70: 947-959. 2. Casewell M, Phillips I. Hands as route of transmission of Klebsiella sp. British Medical Journal 1977; 2: 1315-1317. 3. Atherton ST, White DJ. Stomach as a source of bacteria colonising respiratory tract during artificial ventilation. Lancet 1978; 2: 968-969. 4. du Moulin GC, Hedley-Whyte J, Paterson DG, Lisbon A. Aspiration of gastric bacteria in antacid-treated patients: a frequent cause of postoperative colonisation of the airway. Lancet 1982; 1: 242-245. 5. The Medical Directory. Harlow: Longman, 1990. 6. du Moulin GC. Minimising the potential for nosocomial pneumonia: architectural, engineering, and environmental considerations for the intensive care unit. European Journal of Clinical Microbiology and Infectious Diseases 1989; 8: 69-74. 7. Shirani KZ, McManus AT, Vaughan GM, McManus WF, Pruitt BA, Mason AD. Effects of environment on infection in burns patients. Archives of Surgery 1986; 121: 31-36. 8. Cowley JM, Hill S, Ross J, Lertzman J, Louie TJ. Handwashing practices in an intensive care unit: the effect of an educational program and its relationship to infection rates. American Journal of Infection Control 1989; 17: 330-339. 9. Kappstein I, Daschner FD. Nosocomial infections in intensive care units. Current Opinion in Infectious Diseases 1990; 3: 509-512. 10. Walton JR, Shapiro BA, Harrison RA, Davison R, Reisberg BE. Serratia bacteremia from mean arterial pressure monitors. Anesthesiology 1975; 1: 113-114. 11. Inglis TJJ. Pneumonia in intensive care patients. British - Journal of Anaesthesia 1990; 65: 94-106. 12. Gallagher j , Strangeways-JEM, Allt-Graham J. Contamination control in long-term ventilation. Anaesthesia 1987; 42: 476-481. 13. Garibaldi RA, Britt MR, Webster C, Pace NL. Failure of bacterial filters to reduce the incidence of pneumonia after inhalational anesthesia. Anesthesiology 1981; 54: 364-368. 14. Buckley PM. Increase in resistance of in-line breathing filters
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
reduced to 48-hourly regimens without a significant increase in bacterial contamination. Changes every 48 h have been recommended also in a recent study by the Hospital Infection Research laboratory [16]. Moreover, bacterial filter units placed at the Yconnector remove the risk of colonizing the ventilated airway with exogenous bacteria, and contamination of the tubing by the patient. Any cost-benefit analysis of filter use therefore has to include an assessment of impact on tubing changes. Substantial savings could be made in many units by reducing the frequency of tubing change and limiting the use of microbial filters to single units placed at the Y-connector. There is still uncertainty over the value of some aspects of urinary tract management, reflected in the range of responses given regarding frequency of changing systems, and meatal care. Until the efficacy of meatal toilet has been demonstrated conclusively, very frequent or complex toilet regimens seem inappropriate and may even be counterproductive. While many units send routine catheter specimens of urine for culture, many centres reported using antibiotics in less than 25 % of positive cases, suggesting a significant waste of diagnostic resources. The apparent mismatch between urine culture results and antibiotic treatment might be reduced if units were to establish a clearer rationale for the diagnosis and treatment of urinary tract infections in catheterized patients. The range of antiseptics used in a given unit depends in part on the hospital antiseptic policy. The large range of proprietary agents available is thus reflected in the range of responses in this survey. However, this degree of variation makes it difficult for ICU staff to become familiar with specific indications for use of specific agents. Agents such as peroxide and Eusol are poor antiseptics and, in leaving a wet surface, may encourage the further growth of bacteria. Units using this type of agent should review their practice, and consider the use of more effective antiseptic agents. In the past decade it has been recognized that the patient's bacterial flora comprises the most important source of bacteria colonizing the ventilated airway. Strategies aimed at reducing or preventing endogenous source infections (selective decontamination of the digestive tract (SDD)) and sucralfate substitution for H2-receptor antagonists, have not been shown to have unequivocal benefit in the prevention of deaths caused by ventilator-associated pneumonia [17, 18]. Moreover, selective decontamination requires intensive microbiological surveillance throughout its use, and the use of SDD by 18 ICU, of which only nine have a consultant-led microbiology service, runs contrary to the recommendations [19], On the basis of this survey, we wish to make the following proposals.:. (1) A given practice should be incorporated into the unit's infection control policy and staff training programme, if there is a rational basis for its implementation. It is essential that these practices are agreed between intensive care staff and the infection control team, and written in a clearly defined policy.
219
220 in humidified air. British Journal of Anaesthesia 1984; 56: 637-643. 15. Craven DE, Connolly MG jr, Lichtenberg DA, Primeau PJ, McCabe WR. Contamination of mechanical ventilators with tubing changes every 24 or 48 hours. New England Journal of Medicine 1983; 306: 1505-1509. 16. Cadwallader HL, Bradley CR, Ayliffe GAJ. Bacterial contamination and frequency of changing ventilator circuiting. Journal of Hospital Infection 1990; IS: 65-72. 17. Ledingham IMacA, Alcock SR, Eastaway ST, McDonald JC, McKay IC, Ramsay G. Triple regimen of selective decontamination of the digestive tract, systemic cefotaxime
BRITISH JOURNAL OF ANAESTHESIA and microbiological surveillance for prevention of acquired infection in intensive care. Lancet 1988; 1: 785-790. 18. Driks MR, Craven DE, Celli BR, Manning M, Burke RA, Garvin GM, Kunches LM, Farber HW, Wedel SA, McCabe WR. Nbsocomial pneumonia in intubated patients given sucralfate as compared with antacids or histamine type 2 blockers. New England Journal of Medicine 1987; 317: 1376-1382. 19. Alcock SR, Ledingham IMacA. Selective decontamination of the digestive tract and prevention of infection in intensive care units. Journal of Antimicrobial Chemotherapy 1988; 22: 97-101.
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
COMMENTARIES
INFECTION CONTROL IN INTENSIVE CARE UNITS: U.K. NATIONAL SURVEY T. J. J. INGLIS, L. J. SPROAT, P. M. HAWKEY AND P. KNAPPETT SUMMARY
KEY WORDS Infect/on: control practices. Intensive care.
The growing complexity of intensive care medicine during the past three decades has been accompanied by an increased potential for nosocomial infection in patients undergoing artificial ventilation of the lungs. A combination of invasive monitoring, multiple therapies and impaired host defences renders these patients uniquely susceptible to nosocomial wound, device-related, respiratory and urinary tract infections [1]. As a result of common-source outbreaks in patients in intensive care units (ICU), strategies have been developed to interrupt cross-infection, including rigorous handwashing [2]. More recently, the role of the patient's own bacterial flora has been recognized as an important source of subsequent infection [3, 4]. Local experience suggested wide variations in the infection control practice of ICU. We therefore undertook a survey of local and other U.K. adult ICU to assess the extent of this variation, identify
METHODS
A questionnaire was designed by microbiologists and anaesthetist members of the Yorkshire Regional Intensive Care Committee. After a pilot survey (the results of which were included in the overall analysis), the confidential questionnaire was circulated to all other adult ICU listed in the Directory of Emergency and Special Care Units, 1989 (CMA Medical Data Ltd, Cambridge), during April-June 1990. Medical Directors were asked to complete the questionnaire, or to arrange for its completion by a senior medical or nursing member of the ICU staff. Non-responding units were followed up by post in July 1990 and by telephone in August 1990. The questionnaire dealt with workload, staffing, architecture and specific infection control practices. RESULTS
Completed questionnaires were received from 246 of 290 (85 %) ICU in the survey. One paediatric ICU had been classified incorrectly in the source manual. Six of the 246 responding units had closed or changed status and therefore were excluded from further analysis. Of the remaining 240 units, 10 were surgical only and 10 were specialist ICU. The size range was two to 16 beds per unit with a most frequent response (m.f.r.) of six beds. Occupancy ranged from < 10 to > 9 0 % (m.f.r. = 61-70%); patients per year from < 100 to 2432 (m.f.r. = 201-300); nursing staff one to 15 per shift (m.f.r. = 4) with a nurse: patient ratio of 1:2-2:1 (m.f.r. = 1:1), and the percentage admissions who underwent tracheal intubation < 10 % to > 90 % (m.f.r. = 71-80%). Facilities There were 33 ICU with no beds in single side wards (range 0-8, m.f.r. = 1). In the majority of units, less than 30% of the beds were in single
TIMOTHY J. J. INGLIS, D.M. ; LOUISE J. SPROAT, R.G.N. ; PETER M.
HAWKEY, M.D. ; Department of Microbiology, University of Leeds, Leeds LS2 9JT. P. KNAPPETT, F.C.ANAES., Bradford Royal
Infirmary, Bradford, W. Yorks. Accepted for Publication: July 10, 1991. Correspondence to T.J.J.I.
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
A completed questionnaire was returned by 246 (85%) intensive care units participating in a national survey of infection control practice in intensive care. Thirty-three units had no provision for isolating patients in single side wards. Sixty percent of responding ICU had fewer than one washbasin per bedspace. Several units reported using ventilator filters or tubing more frequently than is currently recommended. Excessive numbers of catheter urine specimens were sent for laboratory examination by some units. A small number of units used open urinary drainage systems. A significant proportion of ICU had no formally recognized policy on the management of intravascular cannulae. Only 8% of ICU in the U.K. were using a selective decontamination regimen, and nine of these (50 %) had no full time consultant microbiologist available to supervise the recommended microbiological management. The majority of ICU received a regular visit (^ one per week) from a microbiologist. Proposals are made on the development of a specialized infection control service in order to reduce the risk of nosocomial infection in intensive care, and to improve on existing resource management.
inappropriate procedures and suggest priorities for future infection control practice in the ICU.
INFECTION CONTROL IN THE ICU
217
TABLE I. Visits to intensive care units by medical microbiology staff- *Full and part-time medical microbiologist on hospital staff listed in 1990 Hospital Directory [5]
Visit frequency
Number (%) of ICU responding
Number (%) with consultant microbiologist*
Daily 2-4 per week Weekly Less often Only on request Never
53 (22) 48 (20) 22(9) 28 (12) 82 (34) 6(2)
32/53 (60) 20/48 (42) 7/22 (32) 4/28 (14) 14/82 (17) 1/6 (17)
Infection control and microbiology support In response to examples given (the examples used were a patient with meningitis, and an outbreak of salmonellosis), the microbiologist was stated to be the most frequent source of advice by 203 (84 %) and 214 (89%) ICU, respectively. The advice of an infection control nurse was sought also by 192 (80 %) and 189 (79%) units, and infection control laboratory staff were involved in 32 (13%) and 38 (16%) instances. In only six (2.5%) cases did the ICU report no visit from a microbiologist under any circumstance. Five of these hospitals had no permanent consultant microbiologist and the remaining centre had one [5]. Another 82 received a visit only on request (table I). Infections associated with intravascular cannulae The reported use of policies to specify aspects of management for different cannula sites is given in table II. Responses on strategies used to prevent intravascular cannula infections were too varied to allow generalized comment. When intravascular pressures (arterial, pulmonary arterial) were monitored, non-disposable transducers were used in 72/240 (31 %) units and non-disposable domes in 11 units. Water was used between the cannula and transducer membrane in 32 units, using non-disposable transducers, domes, or both. In response to the question of how an i.v. catheter tip culture result of Staphylococcus epidermidis would
Number (%) ICU responding
Interval (days) 1 2 3
73 (32) 62 (27) 20(9) 12(5) 61 (27) 228 (100)
Not at all Other Total
be dealt with, 169/224 (75%) units replied that no treatment would be given; 17 (8%) would use antibiotics and 23 (10%) would treat with reservations. Respiratory support The time interval between ventilator tubing changes reported by units with a specific policy is shown in table III. Ventilator filters were used by 223/237 (94%) units, 165 (70%) of which reported their use in all circuits. Another 36 (16%) used filters in most circuits. The most commonly reported device combined heat-moisture exchange (HME) properties and filtration (used by 168 ICU). This device was used by 161 (80%) of the 201 units using filters on all or most ventilator circuits. Of 154 units using the filter/HME device, 118 (77%) changed them daily. Thirty-three of 159 (21 %) units reported using the device on both inspiratory and expiratory limbs of the ventilator circuit. Urinary tract Open urine drainage systems were reported to be in routine use by four units. Time to replacement of urinary drainage systems varied from 1 to 14 days (m.f.r. = 7 days), but there were 16 categories of qualified response. Meatal care practices varied considerably: 27 different responses were given, with no clear consensus. Routine catheter specimens were sent to the laboratory from "on admission only" to daily (m.f.r. = twice weekly), and the proportion of positive results used as an indication for treatment ranged between < 25 % and > 75 % (m.f.r. < 25%). Surgical wound care A large range of antiseptic agents were reported to be in common use, amongst which were Eusol, Milton, peroxide and alcohol. Agents reported most
TABLE II. Type of ICU policy for management of intravascular cannula
Total number Cannula type
responding
I.v. peripheral I.v. central Intra-arterial Pulmonary artery Parenteral nutrition
187 203 190 170 191
Policy type Written
Verbal
None
74(40%)" 90(44%) 72(38%) 72(42%) 105(55%)
78(42%) 83 (41 %) 88(46%) 72(42%) 66(35%)
35(19%) 30(15%) 30(16%) 26(15%) 20(10%)
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
occupant rooms. The number of washbasins varied from one to 19 per unit (m.f.r. = 5) and 145 (60%) had fewer than one basin per bed. The minimum distance from bed to basin varied from < 1 to > 10 m (m.f.r. > 2 m) and the maximum from > 1 to > 23 m (m.f.r. > 6 m).
TABLE III. Interval between changes of ventilator tubing
BRITISH JOURNAL OF ANAESTHESIA
218 TABLE IV. Strategies for prevention of gastric stress ulcer
Strategy Histamine antagonist Sucralfate Antacids None of above
Number (%) of ICU responding 197 (84) 62 (26) 34 (14) 7(3)
frequently were chlorhexidine-based (used by 80 units (36%)) and iodophor-based (79 (35%)). A total of 34 preparations was reported.
Selective decontamination of the digestive tract was in use in 18 units (8 %), only five of which were teaching hospitals. Nine of these centres had no full time consultant microbiologist. Strategies for prophylaxis of gastric stress ulceration are given in table IV. DISCUSSION
The response rate (85%) in this survey reflects the importance attached to infection control by British Intensive Care Units, and presents an opportunity to review current infection control arrangements. It is held widely that the risk of cross-infection from a given patient may be reduced by source isolation, often achieved by nursing the patient in a single-bedded room under conditions of scrupulous hygiene [6]. Although some authorities regard the provision of single cubicles for all intensive care patients as mandatory [7], a lower proportion of the total bed complement is probably satisfactory. When an intensive care patient presents a particular infection hazard to other patients or staff, source isolation can be achieved by nursing the patient in a side ward and institution of additional barrier-type precautions. It is therefore a matter of concern that 33 units reported having no single-bed side ward. If, as may easily happen in an ICU, handwashing and other barrier-type procedures are violated for reasons of convenience or emergency, the separation of a potentially infectious patient from other susceptible individuals by enclosure in a single side ward may be the only remaining barrier to cross-infection. Handwashing is probably the single most important procedure in the prevention of transmissible disease in hospital inpatients. In the ICU, where patients are more likely to be colonized with potential pathogens, and physical contact with hospital staff is more frequent, high standards of staff hand hygiene are necessary. Yet it has been shown that handwashing rates are poor amongst ICU staff, and medical staff are particularly resistant to educational programmes [8]. This level of compliance with a simple infection control procedure such as handwashing is alarming when the possible effect of a single violation on days or weeks of careful hygiene observance is considered. Nevertheless, ICU staff cannot be held entirely responsible for poor rates of compliance when a unit is poorly equipped for
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
Strategies for the prevention of endogenous source infection
handwashing. Small numbers of washbasins located far from the head end of bedspaces, or obstructed by respiratory support and monitoring equipment, act as a disincentive to regular handwashing. Moreover, washbasins shared between ICU bedspaces present a further route for indirect bacterial transmission between patients. The provision of alcohol hand rub dispensers may improve compliance, but they are not a substitute for washbasins which, unlike hand rub dispensers, cannot be removed. The infection control team (microbiologist, nurse and laboratory staff) have a major responsibility for the prevention and containment of infectious diseases throughout the hospital. It is notable that a majority of ICU receive active support from a medical microbiologist in the form of visits to the unit, and rely on the various components of the infection control team for advice on the management of infection. The very small number of responding units reporting a lack of microbiological support had no consultant medical microbiologist appointed to their hospital, with one exception. Intravascular cannula infections begin either with bacterial colonization of the insertion site, or contamination of the limb. As these devices give direct access to the vascular space, it is surprising that a substantial minority of units reported having no formally recognized policy on cannula management. Formal unit policies for cannula management would be a possible route towards establishing uniformity and quality in standards of intravascular cannula hygiene. Arterial pressure transducers present a specific infection risk, highlighted in a recent review [9]. Non-disposable domes may crack after repeated autoclaving and admit bacteria from the transducer side of the membrane, and have been recognized for some time as a risk to patients [10]. The use of nondisposable transducers with disposable domes may still assist bacterial access to the arterial circulation, because the fluid used in some units to improve membrane-to-transducer coupling may become contaminated. Subsequent dismantling of the transducer from the dome contaminates the fingertips of staff and, if followed by arterial blood sampling, may lead to contamination of the fluid path. Units using either non-disposable domes, or non-disposable transducers/disposable domes (particularly those using an aqueous coupling fluid) may need to review their practice. The current extent of filter use in ventilator circuits cannot be justified fully according to present view on the pathogenesis of ventilator-associated pneumonia—that is, that infecting bacteria are primarily endogenous in origin [11]. However, there is some evidence that Pseudomonas sp. infection may be reduced by use of filters [12]. Compelling evidence for the in vivo efficacy offiltersis noticeably lacking [13], and concern has been expressed over increases in resistance caused by the accumulation of condensate in some filters [14]. There is a need to establish a clear rationale for the use of ventilator filters in independent, controlled trials. The hygienic management of ventilator tubing has already been studied [15]. Daily changes may be
INFECTION CONTROL IN THE ICU
(2) Where there is no clear basis or consensus for a given practice, it should be the subject of independent evaluation, before a decision is made regarding its suitability for incorporation in unit policy or training schemes. (3) Many units require alterations to their architecture, fittings, or both, to enable adequate hand washing and source-isolation of infected patients to be carried out. (4) Given the current contribution of the infection control team to intensive care and the need for a proactive infection control programme in the ICU, a member of intensive care staff should be appointed to work as a member of the infection control team. Smaller units might second a senior member of nursing staff to the infection control team as a link worker, whereas larger ICU may prefer to appoint an infection control sister as a full time specialist.
ACKNOWLEDGEMENTS We thank the Intensive Care Society for providing financial support, and members of the Yorkshire Regional Intensive Care Committee for assistance with the survey in its early stages.
REFERENCES 1. Haley RW, Hooton TM, Culver DH, Stanley RC, Emori TG, Hardison CD, Quade D, Schachtman RH, Schaberg DR, Shah BV, Schatz GD. Nosocomial infections in U.S. hospitals 1975-76. Estimated frequency by selected characteristics of patients. American Journal of Medicine 1981; 70: 947-959. 2. Casewell M, Phillips I. Hands as route of transmission of Klebsiella sp. British Medical Journal 1977; 2: 1315-1317. 3. Atherton ST, White DJ. Stomach as a source of bacteria colonising respiratory tract during artificial ventilation. Lancet 1978; 2: 968-969. 4. du Moulin GC, Hedley-Whyte J, Paterson DG, Lisbon A. Aspiration of gastric bacteria in antacid-treated patients: a frequent cause of postoperative colonisation of the airway. Lancet 1982; 1: 242-245. 5. The Medical Directory. Harlow: Longman, 1990. 6. du Moulin GC. Minimising the potential for nosocomial pneumonia: architectural, engineering, and environmental considerations for the intensive care unit. European Journal of Clinical Microbiology and Infectious Diseases 1989; 8: 69-74. 7. Shirani KZ, McManus AT, Vaughan GM, McManus WF, Pruitt BA, Mason AD. Effects of environment on infection in burns patients. Archives of Surgery 1986; 121: 31-36. 8. Cowley JM, Hill S, Ross J, Lertzman J, Louie TJ. Handwashing practices in an intensive care unit: the effect of an educational program and its relationship to infection rates. American Journal of Infection Control 1989; 17: 330-339. 9. Kappstein I, Daschner FD. Nosocomial infections in intensive care units. Current Opinion in Infectious Diseases 1990; 3: 509-512. 10. Walton JR, Shapiro BA, Harrison RA, Davison R, Reisberg BE. Serratia bacteremia from mean arterial pressure monitors. Anesthesiology 1975; 1: 113-114. 11. Inglis TJJ. Pneumonia in intensive care patients. British - Journal of Anaesthesia 1990; 65: 94-106. 12. Gallagher j , Strangeways-JEM, Allt-Graham J. Contamination control in long-term ventilation. Anaesthesia 1987; 42: 476-481. 13. Garibaldi RA, Britt MR, Webster C, Pace NL. Failure of bacterial filters to reduce the incidence of pneumonia after inhalational anesthesia. Anesthesiology 1981; 54: 364-368. 14. Buckley PM. Increase in resistance of in-line breathing filters
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
reduced to 48-hourly regimens without a significant increase in bacterial contamination. Changes every 48 h have been recommended also in a recent study by the Hospital Infection Research laboratory [16]. Moreover, bacterial filter units placed at the Yconnector remove the risk of colonizing the ventilated airway with exogenous bacteria, and contamination of the tubing by the patient. Any cost-benefit analysis of filter use therefore has to include an assessment of impact on tubing changes. Substantial savings could be made in many units by reducing the frequency of tubing change and limiting the use of microbial filters to single units placed at the Y-connector. There is still uncertainty over the value of some aspects of urinary tract management, reflected in the range of responses given regarding frequency of changing systems, and meatal care. Until the efficacy of meatal toilet has been demonstrated conclusively, very frequent or complex toilet regimens seem inappropriate and may even be counterproductive. While many units send routine catheter specimens of urine for culture, many centres reported using antibiotics in less than 25 % of positive cases, suggesting a significant waste of diagnostic resources. The apparent mismatch between urine culture results and antibiotic treatment might be reduced if units were to establish a clearer rationale for the diagnosis and treatment of urinary tract infections in catheterized patients. The range of antiseptics used in a given unit depends in part on the hospital antiseptic policy. The large range of proprietary agents available is thus reflected in the range of responses in this survey. However, this degree of variation makes it difficult for ICU staff to become familiar with specific indications for use of specific agents. Agents such as peroxide and Eusol are poor antiseptics and, in leaving a wet surface, may encourage the further growth of bacteria. Units using this type of agent should review their practice, and consider the use of more effective antiseptic agents. In the past decade it has been recognized that the patient's bacterial flora comprises the most important source of bacteria colonizing the ventilated airway. Strategies aimed at reducing or preventing endogenous source infections (selective decontamination of the digestive tract (SDD)) and sucralfate substitution for H2-receptor antagonists, have not been shown to have unequivocal benefit in the prevention of deaths caused by ventilator-associated pneumonia [17, 18]. Moreover, selective decontamination requires intensive microbiological surveillance throughout its use, and the use of SDD by 18 ICU, of which only nine have a consultant-led microbiology service, runs contrary to the recommendations [19], On the basis of this survey, we wish to make the following proposals.:. (1) A given practice should be incorporated into the unit's infection control policy and staff training programme, if there is a rational basis for its implementation. It is essential that these practices are agreed between intensive care staff and the infection control team, and written in a clearly defined policy.
219
220 in humidified air. British Journal of Anaesthesia 1984; 56: 637-643. 15. Craven DE, Connolly MG jr, Lichtenberg DA, Primeau PJ, McCabe WR. Contamination of mechanical ventilators with tubing changes every 24 or 48 hours. New England Journal of Medicine 1983; 306: 1505-1509. 16. Cadwallader HL, Bradley CR, Ayliffe GAJ. Bacterial contamination and frequency of changing ventilator circuiting. Journal of Hospital Infection 1990; IS: 65-72. 17. Ledingham IMacA, Alcock SR, Eastaway ST, McDonald JC, McKay IC, Ramsay G. Triple regimen of selective decontamination of the digestive tract, systemic cefotaxime
BRITISH JOURNAL OF ANAESTHESIA and microbiological surveillance for prevention of acquired infection in intensive care. Lancet 1988; 1: 785-790. 18. Driks MR, Craven DE, Celli BR, Manning M, Burke RA, Garvin GM, Kunches LM, Farber HW, Wedel SA, McCabe WR. Nbsocomial pneumonia in intubated patients given sucralfate as compared with antacids or histamine type 2 blockers. New England Journal of Medicine 1987; 317: 1376-1382. 19. Alcock SR, Ledingham IMacA. Selective decontamination of the digestive tract and prevention of infection in intensive care units. Journal of Antimicrobial Chemotherapy 1988; 22: 97-101.
Downloaded from http://bja.oxfordjournals.org/ at University of Illinois at Urbana-Champaign on March 9, 2015
