- Email: [email protected]
The effects of electric current on bacteria colonising intravenous catheters
ffournaloflnfection (1993) 27, 261-269
T h e effects o f e l e c t r i c c u r r e n t on b a c t e r i a c o l o n i s i n g intravenous catheters Wai-Kin Liu, Sarah E. Tebbs, Philip O. Byrne and T h o m a s S. J. Elliott Department of Clinical Microbiology, Queen Elizabeth Hospital, Edgbaston, Birmingham B I 5 2TH, U.K. Accepted for publication 17 May 1993
Summary The effect of a direct electric current (IO #A) on the growth of Staphylococcus aureus, Staphylococcus epidermidis, Escherichia coli, Klebsiella pneurnoniae and Proteus mirabilis was investigated. When the ends of negatively-charged intravascular catheters were placed in nutrient agar seeded with bacteria, circular zones of inhibition of bacterial growth were observed around the catheters. The zones ranged from 6 to 16 mm in diameter according to the organism under test. Zones of inhibition were not produced around positively-charged catheters. Bacteria colonising the surfaces of catheters were similarly affected by the application of a IO #A electric current. A negative electric current applied to colonised catheters for 4 to 24 h significantly reduced the number of adherent viable organisms as compared to controls. The results demonstrated that a constant electric current of low amperage might be used to reduce bacterial colonisation of intravascular catheters. This may offer a novel means of protecting catheters and other prosthetic devices from associated sepsis in vivo.
Introduction Between 4 to 18 % all central venous catheters (CVC) are associated with sepsis. 1 Coagulase-negative staphylococci are the most f r e q u e n t cause. In addition to Staphylococcus aureus, t h e y account for m o r e t h a n 5 o % all catheter-related infections. A wide range o f other organisms, including G r a m negative aerobic bacilli a n d yeasts have also been associated with these infections. 1'2 T h e organisms are derived primarily f r o m patients' skin, some gaining access to the catheter at the time of insertion, whereas others migrate along the outer surface o f the catheter2 Colonisation o f the internal surface o f the catheter h u b also leads to infection following intraluminal migration. 4'5 After adhesion to a catheter surface, m a n y bacteria, including S. aureus and coagulase-negative staphylococci, p r o d u c e a slime layer or glycocalyx6 which facilitates their a t t a c h m e n t 7 and inhibits cellular i m m u n i t y . 8'9 Slime also provides a d y n a m i c ecological niche which attracts essential nutrients and protects organisms f r o m antimicrobial agents. 1° Successful therapy o f C V C related infections m a y therefore be difficult to attain. Several approaches to the prevention o f C V C infection have been attempted. T h e y include i m p r o v e d catheter care, 11 modification of the intrinsic properties o f polymers 12 a n d the b o n d i n g of antimicrobial agents to catheter material.13' la Address correspondence to: Dr T. S. J. Elliott, Department of Clinical Microbiology, Queen Elizabeth Hospital, Edgbaston, Birmingham BI5 2TH, U.K. oi63-4453/93/o6o26I + 0 9 $08.00/0 14
© 1993 The British Society for the Study of Infection JIN 27
262
W.-K. L I U E T A L .
Despite these approaches, some of which are still in developmental stages, CVC-associated infections still occur. We recently reported a novel m e t h o d for preventing CVC-related infection in which a negative electric current repelled organisms from the surface of a catheter. 15 F u r t h e r investigations demonstrated that catheters acting as a cathode prevented both internal and external migration of organisms along the catheter surfaces./6 In this present study, the effects of a direct electric current on the growth and viability of bacteria that have already colonised intravascular catheter surfaces was investigated. Since electric currents not only repel micro-organisms but are also bactericidal, this novel application may prove both prophylactically and therapeutically useful for infections associated with catheters and other indwelling medical prostheses. If successful, it would constitute a major advance. Methods Catheters
Intravascular catheters made of 15 % carbon-impregnated polyetherurethane, with an external diameter of 2"3 m m and which conducted electricity, were kindly supplied by Viggo-Spectramed (Swindon, U.K.). T h e y were steamsterilised at Ioo °C for 3 o m i n u t e s . T h i s did not disrupt the molecular configuration of the plastic polymers. Electrical device
T h e electrical device consisted of a constant current (io #A) generator with a miniature I2V alkaline battery (Duracell M N 2 I ; Farnell components L S I 22TU), in series with a 1 . 2 M ~ , 0"5 W high stability carbon film resistor (RS C o m p o n e n t s , N N I 7 9RS). T h e resistor and battery were enclosed with general-purpose epoxy resin in a plastic container of dimensions 28 m m x I8 m m x I4 m m and weighing Io g.16 Flexible external electrical leads were used to attach the device to both cathodal and anodal catheters. Organisms
T e n strains of bacteria were tested. T h e y included reference strains and fresh clinical isolates obtained from patients at the Queen Elizabeth Hospital, Birmingham, U . K . Reference strains included S. aureus N C T C 657I, S. aureus A T C C 6538, Staphylococcus epidermidis N C T C I IO47, Escherichia coli N C T C IO418, and Klebsiella pneumoniae A T C C 4352. Clinical isolates were S. epidermidis 023 and 462 (non-slime producers), S. epidermidis 8 I I and 983 (slime producers) and Proteus mirabilis. All clinical isolates were identified by standard microbiological techniques. Slime production was detected by the m e t h o d of Christensen. 17 Culture m e d i a
N u t r i e n t agar was prepared by adding I.o % agar to nutrient broth No. 2 (Unipath L t d , Basingstoke, U.K.). I n the experiment designed to investigate the effect of an electric current on bacteria growing on catheter surfaces in nutrient agar, triphenyl tetrazolium salt was added to nutrient agar at a final concentration of o.oi % w / v (BDH Chemical Ltd, Poole, U.K.). T h e salt is
Effects of electric current on bacteria
263
r e d u c e d b y bacteria to f o r m a red metabolite, formazan, which facilitated observation o f bacterial growth. 18 T h e n u m b e r of viable bacteria on a catheter surface was d e t e r m i n e d b y means o f the roll plate culture m e t h o d 19 and the use o f columbia agar ( B B L Becton Dickinson, Cockeysville, U . S . A . ) supplem e n t e d with 7 % sterile defibrinated horse blood. In experiments designed to investigate the bactericidal effect o f an electric current on bacteria attached to a catheter surface, the catheters were s u s p e n d e d in a simple defined m e d i u m ( S D M ) while an electric current was applied. T h e m e d i u m , which was modified Stuart's t r a n s p o r t m e d i u m , 2° maintained the viability o f fastidious organisms for over 48 h. It consisted of s o d i u m glycerolphosphate Io.o g/l, s o d i u m thioglycollate o'5 g/l, cysteine h y d r o c h l o r i d e o'5 g/l, and calcium chloride o.I g/1 (Sigma Chemical Co. L t d , U . K . ) . It was sterilised b y autoclaving at I 2 I °C for I5 min. Zone of inhibition tests
T h r e e bacterial colonies obtained from cultures on b l o o d agar were inoculated into 5 ml of either nutrient broth, for staphylococci, or p e p t o n e water for G r a m negative bacilli. T h e resulting bacterial suspensions were i n c u b a t e d at 37 °C for 2 h in air. T h e n , 4 #1 of the staphylococcal suspension or 3 #1 of the suspension o f G r a m - n e g a t i v e bacilli were spread on the surface of nutrient agar in Petri dishes. T w o carbon catheters (2"5 cm lengths), were placed perpendicularly, 5 cm apart, in the nutrient agar. T h e two catheters were c o n n e c t e d to the electrical device via external leads, one catheter acting as the cathode and the other as the anode. T h e cultures were then i n c u b a t e d at 37 °C for I6 h in air before the diameters of zones o f bacterial inhibition a r o u n d the catheters were m e a s u r e d b y means of vernier calipers. All the bacterial strains were tested on six occasions. Cultures of staphylococci were further examined for the presence o f viable bacteria within the inhibition zones. A block of agar containing the inhibition zone was r e m o v e d and an impression smear prepared, G r a m - s t a i n e d and examined microscopically for bacteria. Also, in order to determine w h e t h e r or not viable bacteria were present in the zones of inhibition, some o f the agar blocks were inoculated directly on b l o o d agar in plates which were then i n c u b a t e d at 37 °C in air for I6 h before being examined for bacterial growth. T h e effect of an electric current on the s u b s e q u e n t ability of nutrient agar to s u p p o r t bacterial g r o w t h was investigated. A IO # A current was applied via the carbon catheters to n u t r i e n t agar in Petri dishes, as described earlier, for I6 h in air at 37 °C. After the electric current had been applied and the catheters r e m o v e d , 4 #1 of either S. epidermidis N C T C I IO47, 811 or 983 were inoculated on the agar. T h e cultures were incubated at 37 °C, in air, for I6 h and then examined for bacterial g r o w t h and any zones of inhibition. T h e test was p e r f o r m e d in triplicate. Catheter colonisation
Catheters (2"5 cm lengths) were colonised with bacteria b y immersion in Io ml bacterial suspensions, each o f which was p r e p a r e d b y diluting an overnight b r o t h culture 2oo times in fresh nutrient broth. T h e catheters were incubated in the suspensions at 37 °C for 2"5 h in a shaking incubator ( U n i p a t h L t d 14-2
264
W.-K. LIU ET AL.
Basingstoke, U.K.). T h e y were then removed from the suspensions and washed in 25 ml phosphate buffered saline (PBS). T h e washing was repeated twice with fresh PBS so as to remove loosely attached organisms. Effect of the electric current on bacterial growth
T h r e e colonised catheters were placed through apertures in the side wall of the base of a sterile Petri dish. Molten nutrient agar at 46 °C containing tetrazolium salt was slowly poured over the catheters. After the agar had set, the catheters were connected to the electrical device via external leads and clips. One catheter acted as the cathode, one as the anode and the third as the control. T h e plate was then incubated at 37 °C in air for 48 h. T h e surfaces of the catheters were examined daily for bacterial growth with a stereomicroscope (SV8, Zeiss, Germany). Each strain of bacteria was tested in triplicate. T h e b a c t e r i c i d a l effect o f a n e l e c t r i c c u r r e n t o n t h e n u m b e r o f v i a b l e b a c t e r i a a t t a c h e d to c a t h e t e r s u r f a c e s
Catheters (2"5 cm lengths) were colonised with bacteria for 2"5 h as described above. T h e y were then connected to the cathode of the electrical device and completely immersed vertically in 15 ml SDM. A further catheter, colonised with bacteria and similarly placed in 15 ml S D M without the electric current being applied, served as a control. After incubation at 37 °C for 16 h in air, the catheters were removed and the n u m b e r of bacteria attached to their surfaces determined by the roll plate technique. T h e plates were incubated for 16 h at 37 °C in air and the n u m b e r of colony-forming units (CFU) determined. S. epidermidis N C T C 11o47, S. epidermidis 023 and 983 were each tested individually on six occasions. Susceptibility of bacteria after various catheter colonisation times
Catheters were colonised with S. epidermidis N C T C 1 lO47 for 2"5 or 16 h as described earlier. T h e y were then placed into 15 ml S D M and connected to the cathode of the electrical device. Control catheters without an electric current were prepared also. Five identical sets of catheters were prepared simultaneously. After 2, 4, 6, 8, or 24 h incubation, the catheters were examined by means of the roll plate technique for the n u m b e r of bacteria colonising their surfaces. This experiment was done in duplicate. Statistics
Statistical analysis was performed by means of the unpaired t-test (two-tailed) applied to log~0 C F U values. Results Zone of inhibition tests
A zone of inhibition was present around the cathodal catheter for each of the bacterial strains tested (Table I). A zone was not observed around any of the anodal or control catheters. Staphylococci had larger zones of inhibition than the Gram-negative bacilli. When impression smears, prepared from material within the zones of inhibition, were Gram-stained, staphylococci were evident.
Effects of electric current on bacteria
26 5
T a b l e I Zones of inhibition produced at the cathodal catheter after the application of a Io # A electric current Organism
Zone size (mm) Mean+ S.D., n = 6
Staphylococcus epidermidis NCTC I Io47 Staphylococcus epidermidis 023 Staphylococcus epidermidis 462 Staphylococcus epidermidis 983 Staphylococcus epidermidis 8I I Staphylococcus aureus ATCC 6538 Staphylococcus aureus NCTC 657I Proteus mirabilis Escherichia coli NCTC IO418 Klebsiella pneumoniae ATCC 4352
I4+3
I5_+o'5 13__+2 II+2 I2_.+2 II__I I6_2
8_+1 8+2 6+0-5
Bacterial g r o w t h was not obtained, however, f r o m material within the zone in any of the experiments, t h e r e b y confirming that the organisms were n o n viable. Bacteria were able to grow on agar which h a d the electric current applied to the surface for 16 h b u t a zone of inhibition was n o t detected a r o u n d the area where the cathodal catheter h a d been placed. Effect of an electric current on growth of bacteria
Catheters were colonised for 2"5 h. After incubation for I6 h in n u t r i e n t agar, the bacteria were observed, by stereo-microscopy, to have grown on the surface o f the anodal and control catheters. F u r t h e r bacterial g r o w t h was not detected, however, on any of the cathodal catheters attached to the electrical device for up to 48 h of incubation. Similar results were observed for all the strains of bacteria tested. E f f e c t o f e l e c t r i c c u r r e n t o n t h e n u m b e r o f v i a b l e b a c t e r i a a t t a c h e d to t h e catheter surfaces
T h e n u m b e r o f bacteria attached to cathodal catheters, colonised for 2"5 h in n u t r i e n t broth, as d e t e r m i n e d by the roll plate m e t h o d , was significantly r e d u c e d after application of an electrical current ( t o #A) for I6 h (P < o'oI). By comparison, > 200o C F U were detected on the surface of each control catheter. Similar results were obtained for the three strains of staphylococci tested (Table 2). S u s c e p t i b i l i t y o f b a c t e r i a to t h e e l e c t r i c c u r r e n t a f t e r c o l o n i s a t i o n o f catheters for various periods of time
T h e n u m b e r s o f viable bacteria present on the surfaces of both the control and cathodal catheters w h i c h were colonised for 2"5 and I6 h are shown in T a b l e 3. W h e n an electric c u r r e n t was applied to bacteria that had colonised the catheters for 2"5 h, a significant reduction in bacterial count was detected as early as 6 h (P < 0"05). T h e bacteria that h a d colonised the catheters for I6 h,
W . - K . L I U E T AL.
266
Table II Numbers of colony forming units (CFU) of bacteria on the catheter surface after application of an electric current for I6 hours, as determined by the roll plate method Numbers of CFU/cm M e a n (range) N = 6 Organism
Control catheter
Cathodal catheter
> 2000
24 (0-77)*
> 2000 > 2ooo
I0 (o--45)* 95 (o-440)*
Staphylococcus epidermidis N C T C i IO47 Staphylococcus epidermidis 023 Staphylococcus epidermidis 983
*P < 0 ' 0 I c o m p a r e d to control catheters.
Table III Numbers of colony forming units (CFU) of Staphylococcus epidermidis N C T C I Io47 attached to a catheter surface after either 2"5 or I6 h colonisation, followed by exposure to negative electric currents for various periods of time. The numbers of C F U were determined by the roll plate method Number of CFU/cm catheter M e a n (range) N = 4 2"5 h c o l o n i s a t i o n Duration of electrical a p p l i c a t i o n (h) 2 4 6 8
24
Control catheter
Cathodal catheter
75 (60-82) 226 (4o-5 I4) 341
72 (37-I00) 56
(Io0-69o)
(0-I3o)
Control catheter >
2000
Cathodal catheter >
2000
> 2000
> 2000
> 2ooo
> 2ooo
36*
> 2000
> 2ooo
(I 1--60) ND
> 2000
(4-I IO) 42*
I38
(46-220) ND
Number of CFU/cm catheter M e a n (range) N = 4 r6 h c o l o n i s a t i o n
76*
(25--282) * P < 0.05 c o m p a r e d to t h e c o n t r o l catheter. N D = n o t done.
however, appeared to be more resistant to the killing effect of the electric current. After 8 h of electrical application, confluent growth of bacteria was still obtained from the cathodal catheter, although bacterial growth was clearly reduced. However, after 24 h of electrical application, the viable count related to the cathodal catheter was significantly reduced when compared with that of the control catheter (P < o'o5).
Effects of electric current on bacteria
z67
Discussion
T h e effects of physical agents such as radiation and heat on bacteria have been studied extensively and their application to controlling bacterial growth and to sterilisation are well established. T h e effect of electrical currents on bacteria and its possible application to modifying bacterial growth, however, has not been thoroughly investigated nor previously applied clinically. T h e r e are reports of the lethal effect of high voltage (2o kV) electric pulses on bacteria and yeasts 21'22 reaching a 99"9 % reduction of total viable counts. Electric currents when applied to silver electrodes have also been shown to be synergistic with the release of silver ions being enhanced by the c u r r e n t ) 3,24 More recently, Davis 25 showed that iontophoresis with gold, carbon and platinum electrodes effectively reduces or eliminates bacteria and yeasts in synthetic urine. T h e electrodes used in these reports have all been metallic. While electric currents enhance release of metallic ions, the effect observed has been related to the action of these ions rather than the electrical potential. T h e present study demonstrated that application of an electric current, via a polyurethane catheter, has bactericidal activity. A zone of inhibition was detected around a cathodal catheter for both staphylococci and Gram-negative bacilli. Bacteria were observed in the Gram-stained smear prepared from material within the zone of inhibition. Organisms from the same zone, however, were not viable when subcultured on fresh medium. T o test whether or not the electric current was affecting the medium, either chemically or by contamination from leached products which may have subsequently inhibited bacterial growth, current was applied to some of the plates which were then inoculated with organisms. Bacteria were able to grow on the entire surface of the agar in these plates, thereby indicating that the electric current did not have an indirect bactericidal effect. After microbial attachment to the surface of a catheter, production of slime is an important step in subsequent colonisation. T h e low amperage current, however, was shown in the roll plate experiment to have bactericidal activity on organisms that had already attached to catheter surfaces. T h e electrical charge required a longer period to reduce significantly the numbers of organisms that had attached to catheters after I6 h colonisation as compared to 2"5 h colonisation. This may have reflected the size and n u m b e r of adherent colonies. T h e mechanism of the lethal effect of an electric current can only be speculative. T r e a t m e n t with high voltage electric pulses has not resulted in any morphological changes being observed by electron microscopy. 21 It would therefore appear unlikely that a low amperage electric current would induce morphological changes that lead to cell death. It is, however, known that biological membranes carry numerous ionised or polarised groups and that their function and structural integrity rely on the balance of electrostatic charges and potentials. 26 Electric currents may disrupt these important physiological functions including membrane transport. This explanation is supported by use of electroporation to transfer large molecules such as D N A by opening the pores in a membrane temporarily) 7 An electric current of up to Io #A can be applied in the h u m a n heart without
268
W.-K. LIU E T A L .
causing side-effects. Furthermore, the probability of such application resulting in e i t h e r v e n t r i c u l a r f i b r i l l a t i o n o r p u m p f a i l u r e is u n l i k e l y . ~8 A t t h e c e l l u l a r level, 75 # A f o r 4 h h a d n o effect in vitro o n cells in c u l t u r e . 24 T h e r e f o r e , t h e a p p l i c a t i o n o f a l o w v o l t a g e e l e c t r i c c u r r e n t in p a t i e n t s a p p e a r s to b e safe. If, as o u r r e s u l t s s u g g e s t , it b o t h r e p e l s a n d kills b a c t e r i a a t t a c h e d to s u r f a c e s , it may offer an exciting way forward for preventing infection from prosthetic d e v i c e s . F u r t h e r in vivo s t u d i e s will b e n e e d e d to c o n f i r m o u r f i n d i n g s . (We thank Professor M. R. W. Brown for his advice and helpful discussion, Miss H. Bailey for typing and preparing the m a n u s c r i p t as well as the British T e c h n o l o g y G r o u p for financial support.) References
I. Elliott T SJ, Faroqui MH. Infections and intravascular devices. BritJ Hosp Med I992; 48: 496-5o3 . 2. Decker MD, Edwards KM. Central venous catheter infections. Paed Clin North Am I988; 35: 579-612. 3. Maki DG. Pathogenesis, prevention, and management of infections due to intravascular devices used for infusion therapy. In: Bisno AL, Waldvogel FA, Eds. Infections associated with indwelling medical devices. Washington: American Society for Microbiology, I99I: I8I-I77. 4. Sitges-Serra A, Puig P, Linares J, Perez JL, Farrero N, Jaurrieta E, Garan J. Hub colonisation as the initial step in an outbreak of catheter-related sepsis due to coagulasenegative staphylococci during parenteral nutrition. Parent Ent Nutr I984; 8: 668-672. 5. Linares J, Sitges-Serra A, Garau J, Perez JL, Martin R. Pathogenesis of catheter sepsis: a prospective study and semi-quantitative culture of catheter hub and segments. J Clin Microbiol I985; 2I: 357-360. 6. Costerton JW, Irvin RT, Cheng KJ. The bacterial glycocalyx in nature and disease. Annu Rev Microbiol I98I ; 35: 299-324. 7. EUiott TSJ. Intravascular-device Infections. J Med Microbiol x988; 27: I6I-I67. 8. Johnson GM, Lee PA, Regelmann WE, Gray ED, Peters G, Quie PG. Interference with granulocyte function by Staphylococcus epidermidis slime. Infect Immun I986; 54: I3-2o. 9- Gray El), Regelmann, WE, Peters G. Staphylococcal slime host defences: effects on lymphocytes and immune function. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and clinical significance of eoagulase-negative staphylococci. New York: Gustav Fischer, I987: 45-53. Io. Sheth NK, Franson TR, Sohule PG. Influence of bacterial adhesion to intravascular catheters on in-vitro antibiotic susceptibility. Lancet I985; ii: i266-I268. I I. Elliott TSJ. A guide to peripherial I.V. cannulation for medical and nursing staff. Helsingborg: Viggo-Spectramed AB, I99O. I2. Jansen B, Peters G. Modern strategies in the prevention of polymer associated infections. ~Hosp Infect I99I; I9: 83-88. I3. Jansen B, Jansen S, Peters G, Pulverer G. In vitro efficacy of a central venous catheter (Hydrocath) loaded with teicoplanin to prevent bacterial colonisation. J Hosp Infect I992; 22 : 93-IO7. I4. Kamal GD, Pfallar MA, Rempe LE, Jebson PJ. Reduced intravascular catheter infection by antibiotic bonding: a prospective, randomised controlled trial. J A M A I99I; 265: 2364-2368. I5. Elliott TSJ, Holford J, Sisson P, Byme P. A novel method to prevent catheter-associated infections. J Med Microbiol I99O; 33: 2. I6. Crocker IC, Liu W-K, Byme P, Elliott TSJ. A novel electrical method for the prevention of microbial colonisation of intravascular cannulae. J Hosp Infect I992; 22: 7-I7. I7. Christensen GD, Simpson WA, Bisno AL, Beachey EH. Adherence of slime producing strains of Staphylococcus epidermidis to smooth surfaces. Infect Immun I982; 37: 318-326.
Effects o f electric current on bacteria
269
I8. Cowan ST. Manual for the identification of medical bacteria. Cambridge: Cambridge University Press, I977. 19. Maki DG, Weise CE, Sarafin HW. A semi-quantitative culture method for identifying intravenous catheter-related infections. New Engl J Med 1977; 296: I3O5-I3O9. 20. The Oxoid Manual. Basingstoke: Unipath Ltd, 199o: 2-203. 21. Htilsheger H, Potel J, Niemann E-G. Electric effects on bacteria and yeast cells. Radiat Environ Biophys 1983; 22: 149-162. 22. Hfilsheger H, Potel J, Niemann E-G. Killing of bacteria with electric pulses of high field strength. Radiat Environ Biophys I98I ; 20: 53-65. 23. Chu C-C, Tasi W-C, Yao J-Y, Chiu S-S. Newly made antibacterial braided nylon sutures. I : In vitro qualitative and in vitro preliminary biocompatibility study. J Biomed Mater Res 1987; 2,: Z28I--Z300. 24. Berger TJ, Spadaro JA, Chapin SE, Becker RO. Electrically charged silver ions: quantitative effects on bacterial and mammalian cells. Antimicrob Agents Chemother 1976; 9: 357-358. 25. Davis CP, Wagle N, Anderson MD, Warren MM. Bacterial and fungal killing by iontophoresis with long-lived electrodes. Antimicrob Agents Chemother I99I; 35: 2131-2134. 26. McLaughlin S. The electrostatic properties of membranes. Annu Rev Biophys Chem 1989;
I8:II3-36. 27. Kilbane JJ, Bielaga BA. Instantaneous gene transfer from donor to recipient microorganisms via electroporation. Biotechniques I99I ; IO: 354-365. 28. British Standard. Medical Electrical Equipment. BS 5724 part I. General requirements for safety. London, 1989.
T h e effects o f e l e c t r i c c u r r e n t on b a c t e r i a c o l o n i s i n g intravenous catheters Wai-Kin Liu, Sarah E. Tebbs, Philip O. Byrne and T h o m a s S. J. Elliott Department of Clinical Microbiology, Queen Elizabeth Hospital, Edgbaston, Birmingham B I 5 2TH, U.K. Accepted for publication 17 May 1993
Summary The effect of a direct electric current (IO #A) on the growth of Staphylococcus aureus, Staphylococcus epidermidis, Escherichia coli, Klebsiella pneurnoniae and Proteus mirabilis was investigated. When the ends of negatively-charged intravascular catheters were placed in nutrient agar seeded with bacteria, circular zones of inhibition of bacterial growth were observed around the catheters. The zones ranged from 6 to 16 mm in diameter according to the organism under test. Zones of inhibition were not produced around positively-charged catheters. Bacteria colonising the surfaces of catheters were similarly affected by the application of a IO #A electric current. A negative electric current applied to colonised catheters for 4 to 24 h significantly reduced the number of adherent viable organisms as compared to controls. The results demonstrated that a constant electric current of low amperage might be used to reduce bacterial colonisation of intravascular catheters. This may offer a novel means of protecting catheters and other prosthetic devices from associated sepsis in vivo.
Introduction Between 4 to 18 % all central venous catheters (CVC) are associated with sepsis. 1 Coagulase-negative staphylococci are the most f r e q u e n t cause. In addition to Staphylococcus aureus, t h e y account for m o r e t h a n 5 o % all catheter-related infections. A wide range o f other organisms, including G r a m negative aerobic bacilli a n d yeasts have also been associated with these infections. 1'2 T h e organisms are derived primarily f r o m patients' skin, some gaining access to the catheter at the time of insertion, whereas others migrate along the outer surface o f the catheter2 Colonisation o f the internal surface o f the catheter h u b also leads to infection following intraluminal migration. 4'5 After adhesion to a catheter surface, m a n y bacteria, including S. aureus and coagulase-negative staphylococci, p r o d u c e a slime layer or glycocalyx6 which facilitates their a t t a c h m e n t 7 and inhibits cellular i m m u n i t y . 8'9 Slime also provides a d y n a m i c ecological niche which attracts essential nutrients and protects organisms f r o m antimicrobial agents. 1° Successful therapy o f C V C related infections m a y therefore be difficult to attain. Several approaches to the prevention o f C V C infection have been attempted. T h e y include i m p r o v e d catheter care, 11 modification of the intrinsic properties o f polymers 12 a n d the b o n d i n g of antimicrobial agents to catheter material.13' la Address correspondence to: Dr T. S. J. Elliott, Department of Clinical Microbiology, Queen Elizabeth Hospital, Edgbaston, Birmingham BI5 2TH, U.K. oi63-4453/93/o6o26I + 0 9 $08.00/0 14
© 1993 The British Society for the Study of Infection JIN 27
262
W.-K. L I U E T A L .
Despite these approaches, some of which are still in developmental stages, CVC-associated infections still occur. We recently reported a novel m e t h o d for preventing CVC-related infection in which a negative electric current repelled organisms from the surface of a catheter. 15 F u r t h e r investigations demonstrated that catheters acting as a cathode prevented both internal and external migration of organisms along the catheter surfaces./6 In this present study, the effects of a direct electric current on the growth and viability of bacteria that have already colonised intravascular catheter surfaces was investigated. Since electric currents not only repel micro-organisms but are also bactericidal, this novel application may prove both prophylactically and therapeutically useful for infections associated with catheters and other indwelling medical prostheses. If successful, it would constitute a major advance. Methods Catheters
Intravascular catheters made of 15 % carbon-impregnated polyetherurethane, with an external diameter of 2"3 m m and which conducted electricity, were kindly supplied by Viggo-Spectramed (Swindon, U.K.). T h e y were steamsterilised at Ioo °C for 3 o m i n u t e s . T h i s did not disrupt the molecular configuration of the plastic polymers. Electrical device
T h e electrical device consisted of a constant current (io #A) generator with a miniature I2V alkaline battery (Duracell M N 2 I ; Farnell components L S I 22TU), in series with a 1 . 2 M ~ , 0"5 W high stability carbon film resistor (RS C o m p o n e n t s , N N I 7 9RS). T h e resistor and battery were enclosed with general-purpose epoxy resin in a plastic container of dimensions 28 m m x I8 m m x I4 m m and weighing Io g.16 Flexible external electrical leads were used to attach the device to both cathodal and anodal catheters. Organisms
T e n strains of bacteria were tested. T h e y included reference strains and fresh clinical isolates obtained from patients at the Queen Elizabeth Hospital, Birmingham, U . K . Reference strains included S. aureus N C T C 657I, S. aureus A T C C 6538, Staphylococcus epidermidis N C T C I IO47, Escherichia coli N C T C IO418, and Klebsiella pneumoniae A T C C 4352. Clinical isolates were S. epidermidis 023 and 462 (non-slime producers), S. epidermidis 8 I I and 983 (slime producers) and Proteus mirabilis. All clinical isolates were identified by standard microbiological techniques. Slime production was detected by the m e t h o d of Christensen. 17 Culture m e d i a
N u t r i e n t agar was prepared by adding I.o % agar to nutrient broth No. 2 (Unipath L t d , Basingstoke, U.K.). I n the experiment designed to investigate the effect of an electric current on bacteria growing on catheter surfaces in nutrient agar, triphenyl tetrazolium salt was added to nutrient agar at a final concentration of o.oi % w / v (BDH Chemical Ltd, Poole, U.K.). T h e salt is
Effects of electric current on bacteria
263
r e d u c e d b y bacteria to f o r m a red metabolite, formazan, which facilitated observation o f bacterial growth. 18 T h e n u m b e r of viable bacteria on a catheter surface was d e t e r m i n e d b y means o f the roll plate culture m e t h o d 19 and the use o f columbia agar ( B B L Becton Dickinson, Cockeysville, U . S . A . ) supplem e n t e d with 7 % sterile defibrinated horse blood. In experiments designed to investigate the bactericidal effect o f an electric current on bacteria attached to a catheter surface, the catheters were s u s p e n d e d in a simple defined m e d i u m ( S D M ) while an electric current was applied. T h e m e d i u m , which was modified Stuart's t r a n s p o r t m e d i u m , 2° maintained the viability o f fastidious organisms for over 48 h. It consisted of s o d i u m glycerolphosphate Io.o g/l, s o d i u m thioglycollate o'5 g/l, cysteine h y d r o c h l o r i d e o'5 g/l, and calcium chloride o.I g/1 (Sigma Chemical Co. L t d , U . K . ) . It was sterilised b y autoclaving at I 2 I °C for I5 min. Zone of inhibition tests
T h r e e bacterial colonies obtained from cultures on b l o o d agar were inoculated into 5 ml of either nutrient broth, for staphylococci, or p e p t o n e water for G r a m negative bacilli. T h e resulting bacterial suspensions were i n c u b a t e d at 37 °C for 2 h in air. T h e n , 4 #1 of the staphylococcal suspension or 3 #1 of the suspension o f G r a m - n e g a t i v e bacilli were spread on the surface of nutrient agar in Petri dishes. T w o carbon catheters (2"5 cm lengths), were placed perpendicularly, 5 cm apart, in the nutrient agar. T h e two catheters were c o n n e c t e d to the electrical device via external leads, one catheter acting as the cathode and the other as the anode. T h e cultures were then i n c u b a t e d at 37 °C for I6 h in air before the diameters of zones o f bacterial inhibition a r o u n d the catheters were m e a s u r e d b y means of vernier calipers. All the bacterial strains were tested on six occasions. Cultures of staphylococci were further examined for the presence o f viable bacteria within the inhibition zones. A block of agar containing the inhibition zone was r e m o v e d and an impression smear prepared, G r a m - s t a i n e d and examined microscopically for bacteria. Also, in order to determine w h e t h e r or not viable bacteria were present in the zones of inhibition, some o f the agar blocks were inoculated directly on b l o o d agar in plates which were then i n c u b a t e d at 37 °C in air for I6 h before being examined for bacterial growth. T h e effect of an electric current on the s u b s e q u e n t ability of nutrient agar to s u p p o r t bacterial g r o w t h was investigated. A IO # A current was applied via the carbon catheters to n u t r i e n t agar in Petri dishes, as described earlier, for I6 h in air at 37 °C. After the electric current had been applied and the catheters r e m o v e d , 4 #1 of either S. epidermidis N C T C I IO47, 811 or 983 were inoculated on the agar. T h e cultures were incubated at 37 °C, in air, for I6 h and then examined for bacterial g r o w t h and any zones of inhibition. T h e test was p e r f o r m e d in triplicate. Catheter colonisation
Catheters (2"5 cm lengths) were colonised with bacteria b y immersion in Io ml bacterial suspensions, each o f which was p r e p a r e d b y diluting an overnight b r o t h culture 2oo times in fresh nutrient broth. T h e catheters were incubated in the suspensions at 37 °C for 2"5 h in a shaking incubator ( U n i p a t h L t d 14-2
264
W.-K. LIU ET AL.
Basingstoke, U.K.). T h e y were then removed from the suspensions and washed in 25 ml phosphate buffered saline (PBS). T h e washing was repeated twice with fresh PBS so as to remove loosely attached organisms. Effect of the electric current on bacterial growth
T h r e e colonised catheters were placed through apertures in the side wall of the base of a sterile Petri dish. Molten nutrient agar at 46 °C containing tetrazolium salt was slowly poured over the catheters. After the agar had set, the catheters were connected to the electrical device via external leads and clips. One catheter acted as the cathode, one as the anode and the third as the control. T h e plate was then incubated at 37 °C in air for 48 h. T h e surfaces of the catheters were examined daily for bacterial growth with a stereomicroscope (SV8, Zeiss, Germany). Each strain of bacteria was tested in triplicate. T h e b a c t e r i c i d a l effect o f a n e l e c t r i c c u r r e n t o n t h e n u m b e r o f v i a b l e b a c t e r i a a t t a c h e d to c a t h e t e r s u r f a c e s
Catheters (2"5 cm lengths) were colonised with bacteria for 2"5 h as described above. T h e y were then connected to the cathode of the electrical device and completely immersed vertically in 15 ml SDM. A further catheter, colonised with bacteria and similarly placed in 15 ml S D M without the electric current being applied, served as a control. After incubation at 37 °C for 16 h in air, the catheters were removed and the n u m b e r of bacteria attached to their surfaces determined by the roll plate technique. T h e plates were incubated for 16 h at 37 °C in air and the n u m b e r of colony-forming units (CFU) determined. S. epidermidis N C T C 11o47, S. epidermidis 023 and 983 were each tested individually on six occasions. Susceptibility of bacteria after various catheter colonisation times
Catheters were colonised with S. epidermidis N C T C 1 lO47 for 2"5 or 16 h as described earlier. T h e y were then placed into 15 ml S D M and connected to the cathode of the electrical device. Control catheters without an electric current were prepared also. Five identical sets of catheters were prepared simultaneously. After 2, 4, 6, 8, or 24 h incubation, the catheters were examined by means of the roll plate technique for the n u m b e r of bacteria colonising their surfaces. This experiment was done in duplicate. Statistics
Statistical analysis was performed by means of the unpaired t-test (two-tailed) applied to log~0 C F U values. Results Zone of inhibition tests
A zone of inhibition was present around the cathodal catheter for each of the bacterial strains tested (Table I). A zone was not observed around any of the anodal or control catheters. Staphylococci had larger zones of inhibition than the Gram-negative bacilli. When impression smears, prepared from material within the zones of inhibition, were Gram-stained, staphylococci were evident.
Effects of electric current on bacteria
26 5
T a b l e I Zones of inhibition produced at the cathodal catheter after the application of a Io # A electric current Organism
Zone size (mm) Mean+ S.D., n = 6
Staphylococcus epidermidis NCTC I Io47 Staphylococcus epidermidis 023 Staphylococcus epidermidis 462 Staphylococcus epidermidis 983 Staphylococcus epidermidis 8I I Staphylococcus aureus ATCC 6538 Staphylococcus aureus NCTC 657I Proteus mirabilis Escherichia coli NCTC IO418 Klebsiella pneumoniae ATCC 4352
I4+3
I5_+o'5 13__+2 II+2 I2_.+2 II__I I6_2
8_+1 8+2 6+0-5
Bacterial g r o w t h was not obtained, however, f r o m material within the zone in any of the experiments, t h e r e b y confirming that the organisms were n o n viable. Bacteria were able to grow on agar which h a d the electric current applied to the surface for 16 h b u t a zone of inhibition was n o t detected a r o u n d the area where the cathodal catheter h a d been placed. Effect of an electric current on growth of bacteria
Catheters were colonised for 2"5 h. After incubation for I6 h in n u t r i e n t agar, the bacteria were observed, by stereo-microscopy, to have grown on the surface o f the anodal and control catheters. F u r t h e r bacterial g r o w t h was not detected, however, on any of the cathodal catheters attached to the electrical device for up to 48 h of incubation. Similar results were observed for all the strains of bacteria tested. E f f e c t o f e l e c t r i c c u r r e n t o n t h e n u m b e r o f v i a b l e b a c t e r i a a t t a c h e d to t h e catheter surfaces
T h e n u m b e r o f bacteria attached to cathodal catheters, colonised for 2"5 h in n u t r i e n t broth, as d e t e r m i n e d by the roll plate m e t h o d , was significantly r e d u c e d after application of an electrical current ( t o #A) for I6 h (P < o'oI). By comparison, > 200o C F U were detected on the surface of each control catheter. Similar results were obtained for the three strains of staphylococci tested (Table 2). S u s c e p t i b i l i t y o f b a c t e r i a to t h e e l e c t r i c c u r r e n t a f t e r c o l o n i s a t i o n o f catheters for various periods of time
T h e n u m b e r s o f viable bacteria present on the surfaces of both the control and cathodal catheters w h i c h were colonised for 2"5 and I6 h are shown in T a b l e 3. W h e n an electric c u r r e n t was applied to bacteria that had colonised the catheters for 2"5 h, a significant reduction in bacterial count was detected as early as 6 h (P < 0"05). T h e bacteria that h a d colonised the catheters for I6 h,
W . - K . L I U E T AL.
266
Table II Numbers of colony forming units (CFU) of bacteria on the catheter surface after application of an electric current for I6 hours, as determined by the roll plate method Numbers of CFU/cm M e a n (range) N = 6 Organism
Control catheter
Cathodal catheter
> 2000
24 (0-77)*
> 2000 > 2ooo
I0 (o--45)* 95 (o-440)*
Staphylococcus epidermidis N C T C i IO47 Staphylococcus epidermidis 023 Staphylococcus epidermidis 983
*P < 0 ' 0 I c o m p a r e d to control catheters.
Table III Numbers of colony forming units (CFU) of Staphylococcus epidermidis N C T C I Io47 attached to a catheter surface after either 2"5 or I6 h colonisation, followed by exposure to negative electric currents for various periods of time. The numbers of C F U were determined by the roll plate method Number of CFU/cm catheter M e a n (range) N = 4 2"5 h c o l o n i s a t i o n Duration of electrical a p p l i c a t i o n (h) 2 4 6 8
24
Control catheter
Cathodal catheter
75 (60-82) 226 (4o-5 I4) 341
72 (37-I00) 56
(Io0-69o)
(0-I3o)
Control catheter >
2000
Cathodal catheter >
2000
> 2000
> 2000
> 2ooo
> 2ooo
36*
> 2000
> 2ooo
(I 1--60) ND
> 2000
(4-I IO) 42*
I38
(46-220) ND
Number of CFU/cm catheter M e a n (range) N = 4 r6 h c o l o n i s a t i o n
76*
(25--282) * P < 0.05 c o m p a r e d to t h e c o n t r o l catheter. N D = n o t done.
however, appeared to be more resistant to the killing effect of the electric current. After 8 h of electrical application, confluent growth of bacteria was still obtained from the cathodal catheter, although bacterial growth was clearly reduced. However, after 24 h of electrical application, the viable count related to the cathodal catheter was significantly reduced when compared with that of the control catheter (P < o'o5).
Effects of electric current on bacteria
z67
Discussion
T h e effects of physical agents such as radiation and heat on bacteria have been studied extensively and their application to controlling bacterial growth and to sterilisation are well established. T h e effect of electrical currents on bacteria and its possible application to modifying bacterial growth, however, has not been thoroughly investigated nor previously applied clinically. T h e r e are reports of the lethal effect of high voltage (2o kV) electric pulses on bacteria and yeasts 21'22 reaching a 99"9 % reduction of total viable counts. Electric currents when applied to silver electrodes have also been shown to be synergistic with the release of silver ions being enhanced by the c u r r e n t ) 3,24 More recently, Davis 25 showed that iontophoresis with gold, carbon and platinum electrodes effectively reduces or eliminates bacteria and yeasts in synthetic urine. T h e electrodes used in these reports have all been metallic. While electric currents enhance release of metallic ions, the effect observed has been related to the action of these ions rather than the electrical potential. T h e present study demonstrated that application of an electric current, via a polyurethane catheter, has bactericidal activity. A zone of inhibition was detected around a cathodal catheter for both staphylococci and Gram-negative bacilli. Bacteria were observed in the Gram-stained smear prepared from material within the zone of inhibition. Organisms from the same zone, however, were not viable when subcultured on fresh medium. T o test whether or not the electric current was affecting the medium, either chemically or by contamination from leached products which may have subsequently inhibited bacterial growth, current was applied to some of the plates which were then inoculated with organisms. Bacteria were able to grow on the entire surface of the agar in these plates, thereby indicating that the electric current did not have an indirect bactericidal effect. After microbial attachment to the surface of a catheter, production of slime is an important step in subsequent colonisation. T h e low amperage current, however, was shown in the roll plate experiment to have bactericidal activity on organisms that had already attached to catheter surfaces. T h e electrical charge required a longer period to reduce significantly the numbers of organisms that had attached to catheters after I6 h colonisation as compared to 2"5 h colonisation. This may have reflected the size and n u m b e r of adherent colonies. T h e mechanism of the lethal effect of an electric current can only be speculative. T r e a t m e n t with high voltage electric pulses has not resulted in any morphological changes being observed by electron microscopy. 21 It would therefore appear unlikely that a low amperage electric current would induce morphological changes that lead to cell death. It is, however, known that biological membranes carry numerous ionised or polarised groups and that their function and structural integrity rely on the balance of electrostatic charges and potentials. 26 Electric currents may disrupt these important physiological functions including membrane transport. This explanation is supported by use of electroporation to transfer large molecules such as D N A by opening the pores in a membrane temporarily) 7 An electric current of up to Io #A can be applied in the h u m a n heart without
268
W.-K. LIU E T A L .
causing side-effects. Furthermore, the probability of such application resulting in e i t h e r v e n t r i c u l a r f i b r i l l a t i o n o r p u m p f a i l u r e is u n l i k e l y . ~8 A t t h e c e l l u l a r level, 75 # A f o r 4 h h a d n o effect in vitro o n cells in c u l t u r e . 24 T h e r e f o r e , t h e a p p l i c a t i o n o f a l o w v o l t a g e e l e c t r i c c u r r e n t in p a t i e n t s a p p e a r s to b e safe. If, as o u r r e s u l t s s u g g e s t , it b o t h r e p e l s a n d kills b a c t e r i a a t t a c h e d to s u r f a c e s , it may offer an exciting way forward for preventing infection from prosthetic d e v i c e s . F u r t h e r in vivo s t u d i e s will b e n e e d e d to c o n f i r m o u r f i n d i n g s . (We thank Professor M. R. W. Brown for his advice and helpful discussion, Miss H. Bailey for typing and preparing the m a n u s c r i p t as well as the British T e c h n o l o g y G r o u p for financial support.) References
I. Elliott T SJ, Faroqui MH. Infections and intravascular devices. BritJ Hosp Med I992; 48: 496-5o3 . 2. Decker MD, Edwards KM. Central venous catheter infections. Paed Clin North Am I988; 35: 579-612. 3. Maki DG. Pathogenesis, prevention, and management of infections due to intravascular devices used for infusion therapy. In: Bisno AL, Waldvogel FA, Eds. Infections associated with indwelling medical devices. Washington: American Society for Microbiology, I99I: I8I-I77. 4. Sitges-Serra A, Puig P, Linares J, Perez JL, Farrero N, Jaurrieta E, Garan J. Hub colonisation as the initial step in an outbreak of catheter-related sepsis due to coagulasenegative staphylococci during parenteral nutrition. Parent Ent Nutr I984; 8: 668-672. 5. Linares J, Sitges-Serra A, Garau J, Perez JL, Martin R. Pathogenesis of catheter sepsis: a prospective study and semi-quantitative culture of catheter hub and segments. J Clin Microbiol I985; 2I: 357-360. 6. Costerton JW, Irvin RT, Cheng KJ. The bacterial glycocalyx in nature and disease. Annu Rev Microbiol I98I ; 35: 299-324. 7. EUiott TSJ. Intravascular-device Infections. J Med Microbiol x988; 27: I6I-I67. 8. Johnson GM, Lee PA, Regelmann WE, Gray ED, Peters G, Quie PG. Interference with granulocyte function by Staphylococcus epidermidis slime. Infect Immun I986; 54: I3-2o. 9- Gray El), Regelmann, WE, Peters G. Staphylococcal slime host defences: effects on lymphocytes and immune function. In: Pulverer G, Quie PG, Peters G, Eds. Pathogenicity and clinical significance of eoagulase-negative staphylococci. New York: Gustav Fischer, I987: 45-53. Io. Sheth NK, Franson TR, Sohule PG. Influence of bacterial adhesion to intravascular catheters on in-vitro antibiotic susceptibility. Lancet I985; ii: i266-I268. I I. Elliott TSJ. A guide to peripherial I.V. cannulation for medical and nursing staff. Helsingborg: Viggo-Spectramed AB, I99O. I2. Jansen B, Peters G. Modern strategies in the prevention of polymer associated infections. ~Hosp Infect I99I; I9: 83-88. I3. Jansen B, Jansen S, Peters G, Pulverer G. In vitro efficacy of a central venous catheter (Hydrocath) loaded with teicoplanin to prevent bacterial colonisation. J Hosp Infect I992; 22 : 93-IO7. I4. Kamal GD, Pfallar MA, Rempe LE, Jebson PJ. Reduced intravascular catheter infection by antibiotic bonding: a prospective, randomised controlled trial. J A M A I99I; 265: 2364-2368. I5. Elliott TSJ, Holford J, Sisson P, Byme P. A novel method to prevent catheter-associated infections. J Med Microbiol I99O; 33: 2. I6. Crocker IC, Liu W-K, Byme P, Elliott TSJ. A novel electrical method for the prevention of microbial colonisation of intravascular cannulae. J Hosp Infect I992; 22: 7-I7. I7. Christensen GD, Simpson WA, Bisno AL, Beachey EH. Adherence of slime producing strains of Staphylococcus epidermidis to smooth surfaces. Infect Immun I982; 37: 318-326.
Effects o f electric current on bacteria
269
I8. Cowan ST. Manual for the identification of medical bacteria. Cambridge: Cambridge University Press, I977. 19. Maki DG, Weise CE, Sarafin HW. A semi-quantitative culture method for identifying intravenous catheter-related infections. New Engl J Med 1977; 296: I3O5-I3O9. 20. The Oxoid Manual. Basingstoke: Unipath Ltd, 199o: 2-203. 21. Htilsheger H, Potel J, Niemann E-G. Electric effects on bacteria and yeast cells. Radiat Environ Biophys 1983; 22: 149-162. 22. Hfilsheger H, Potel J, Niemann E-G. Killing of bacteria with electric pulses of high field strength. Radiat Environ Biophys I98I ; 20: 53-65. 23. Chu C-C, Tasi W-C, Yao J-Y, Chiu S-S. Newly made antibacterial braided nylon sutures. I : In vitro qualitative and in vitro preliminary biocompatibility study. J Biomed Mater Res 1987; 2,: Z28I--Z300. 24. Berger TJ, Spadaro JA, Chapin SE, Becker RO. Electrically charged silver ions: quantitative effects on bacterial and mammalian cells. Antimicrob Agents Chemother 1976; 9: 357-358. 25. Davis CP, Wagle N, Anderson MD, Warren MM. Bacterial and fungal killing by iontophoresis with long-lived electrodes. Antimicrob Agents Chemother I99I; 35: 2131-2134. 26. McLaughlin S. The electrostatic properties of membranes. Annu Rev Biophys Chem 1989;
I8:II3-36. 27. Kilbane JJ, Bielaga BA. Instantaneous gene transfer from donor to recipient microorganisms via electroporation. Biotechniques I99I ; IO: 354-365. 28. British Standard. Medical Electrical Equipment. BS 5724 part I. General requirements for safety. London, 1989.