ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1989, p. 442447 Vol. 33, No.

4
0066-4804/89/040442-06$02.00/0
Copyright C 1989, American Society for Microbiology

Effects of Microamperage, Medium, and Bacterial Concentration

on Iontophoretic Killing of Bacteria in Fluid
C. P. DAVIS,' Z* S. WEINBERG,3 M. D. ANDERSON,' G. M. RAO,3 AND M. M. WARREN4
Departments of Microbiology,' Internal Medicine,2 and Surgery (Division of Urology),4 University of Texas Medical Branch,
Galveston, Texas 77550, and Advanced Clinical Products, Houston, Texas 772583
Received 12 September 1988/Accepted 17 January 1989

Prevention of nosocomial urinary tract infections by iontophoresis is addressed. An iontophoretic generator

was used to provide microamperage (10 to 400 ILA) to vials containing either synthetic urine or supplemented
synthetic urine. Bacteria were added to vials, and parameters of growth, bacterial killing, and multiple
electrode materials were examined. Escherichia coli and Proteus species were both inhibited and killed at
various microamperages and with several electrode types, the most efficient being gold-gold as the anode-
cathode combination. KlebsieUa pneumoniae in supplemented synthetic urine was least inhibited in growth, and
higher microamperage (200 to 400 pA) was most effective in killing the bacteria. Bacterial growth reduction
and killing were directly related to increasing microamperage and were inversely related to bacterial
concentration.

Downloaded from aac.asm.org by on September 14, 2009

Prevention of nosocomial urinary tract infections (UTI) Laboratories, University of Texas Medical Branch), Proteus
has long been a goal of clinicians. Urinary catheters are the mirabilis, and Klebsiella pneumoniae (the latter two organ-
leading cause of nosocomial UTI, and UTI is the most isms were donated by Johnny Peterson). These bacteria
common predisposing factor for cases of fatal gram-negative were chosen because gram-negative bacteria are responsible
sepsis that originate in hospitals (5, 7). Consequently, pre- for 80 to 90% of all UTI, with E. coli being responsible for
vention of UTI related to urinary catheters could lead to a more infections than all other genera combined (7). Frozen
substantial reduction in morbidity and mortality from noso- (-70°C) stocks were inoculated into brain heart infusion
comial gram-negative sepsis. Such prevention could also broth (Difco Laboratories, Detroit, Mich.) and incubated at
decrease morbidity and mortality in geriatric patients, be- 37°C for 24 h before being used. Bacteria were washed three
cause those who require catheterization can have a high times by centrifugation in cold filter-sterilized synthetic
incidence of bacteriuria (9, 15). urine (6). Bacterial concentrations were determined with a
Advancements have been made in catheter design to Petroff-Hausser counting chamber and diluted with sterile
include a completely closed drainage system. Some have
suggested that silver oxide, coated onto catheters, may synthetic urine to obtain the desired concentration of bacte-
protect patients from catheter-associated bacteriuria (14). In ria (2 x 102 or 2 x 104/10 ml, depending on the experiment).
a previous study, we examined iontophoresis as a method to Viable counts were made by diluting 0.1-ml samples in
generate ions that would kill bacteria both in static fluid and sterile synthetic urine and plating onto brain heart infusion
in a model catheter system (4). In the study reported here, agar plates. Because K. pneumoniae grew poorly in syn-
we further define parameters that influence bacterial killing thetic urine, Trypticase soy broth (TSB; BBL Microbiology
and survival during iontophoresis in an artificial urine. Systems, Cockeysville, Md.) was added to synthetic urine as
previously described (6). Standard growth curves for the
MATERIALS AND METHODS three genera of bacteria (without applying current) were
Constant current-constant voltage generator. A microam- determined. Preliminary results showed no difference in
perage generator was built in the Medical Electronics Divi- growth, either with or without metal wires in the vials.
sion at the University of Texas Medical Branch by D. Arnett Model catheter system for static fluid. Bacterial inocula
(4). The instrument provided constant current (1 to 450 puA) were added to glass vials containing 10 ml of synthetic urine
or constant voltage (0.01 to 12.50 V) to 10 independently (5% TSB was added to synthetic urine for K. pneumoniae
controlled channels. (A wiring diagram is available upon only) to give the desired concentration (2 x 102 or 2 x 104
request.) Vented, stoppered vials were attached by metal cells per 10 ml). The vented rubber stopper containing the
connectors to the microamperage generator. A diagram of desired electrode combination was then placed into the
our system has been published elsewhere (4). Wires (gold, bacterial suspension. Constant current was applied and
silver, copper, nickel, platinum, or gold-palladium [60:40 monitored for any changes. Samples (0.1 ml) were with-
alloy]), all approximately 0.2 mm in diameter, were attached drawn from each vial at designated intervals (4 and 8 h and
to the base of the metal connectors so that 2 cm of each wire 1, 2, 3, 4, 5, 6, and 7 days). These samples were serially
or rod was covered by 10 ml of synthetic urine. Variations in diluted in sterile synthetic urine, plated onto brain heart
microamperage were small (average variation, +2 ILA) in infusion agar (E. coli and K. pneumoniae) or MacConkey
each channel when the generator was set to maintain con- agar (P. mirabilis), and incubated at 37°C for 24 h to
stant microamperage (10, 50, 200, or 400 ,uA per channel). determine the CFU. In the gold-gold electrode reinoculation
Bacteria and media. The following bacteria were used: assay, one channel was maintained at 50 ,uA and one was
Escherichia coli (clinical isolate; identified by the Clinical maintained at 400 p.A, and the vials were reinoculated every
other day with the desired concentration of bacteria. Con-
*
Corresponding author. trols consisted of bacterial suspensions in contact with the
442
VOL. 33, 1989 IONTOPHORETIC KILLING OF BACTERIA 443

io 6
U.
E

104 - Proteus

lo
3 * ~~~~Klebsiella

- Kteb./no TSB

102
0 1 0 20 30 40 50 60 70 80

TIME (hrs)

Downloaded from aac.asm.org by on September 14, 2009

FIG. 1. Logarithmic growth of E. coli, P. mirabilis, and K. pneumoniae in synthetic urine. All organisms were cultivated in static vials
containing synthetic urine, except for K. pneumoniae, which was grown both in synthetic urine with no TSB and in synthetic urine containing
TSB.

desired electrode combination without application of con- determine whether they also were capable of reducing or
stant current. eliminating bacterial growth. Several electrode combinations
with gold as a component of the anode or cathode were
RESULTS effective in eliminating growth of E. coli and P. mirabilis
(Fig. 4a and b). The most effective electrodes were carbon-
Growth of bacteria in artificial urine. All three organisms gold, platinum-gold, and a gold-palladium alloy. Although
used in this study grew in artificial urine (Fig. 1). However, these electrodes were effective initially in slowing K. pneu-
K. pneumoniae grew at a very low rate and reached a low moniae growth, the effect was not sustained (Fig. 4c). By
total-population level. Consequently, TSB-enriched artificial day 4, either the electrodes had broken (by day 2) or they
urine was used, and K. pneumoniae then grew very well were no longer able to suppress or eliminate K. pneumoniae
(Fig. 1). In the remainder of the experiments, described growth (Fig. 4c). Of the electrode combinations silver-gold,
below, the basic artificial urine was used as a growth medium nickel-gold, and copper-gold, only the anodes (silver, nickel,
for E. coli and P. mirabilis, but for K. pneumoniae only and copper) broke.
enriched artificial urine was used.
Prevention of bacterial growth with iontophoresis. Gold DISCUSSION
iontophoresis was effective at preventing growth of E. coli,
P. mirabilis, and K. pneumoniae at the inoculum size of 2 x The study confirms and extends our previous work and the
102 cells per 10 ml. In general, iontophoresis at 10, 50, 200, work of others showing that bacterial populations of dif-
and 400 [LA killed the bacterial inocula within 2 days. ferent genera can be reduced or eliminated by iontophoresis
Usually, the organisms were undetectable within 4 h (Fig. 2a (1, 3, 4, 16, 17). Our earlier study (4) utilized a worst-case
and b), except for K. pneumoniae, for which survivors were situation which allowed optimal growth conditions for the
detected after 1 to 2 days (Fig. 2c). Investigations of bacteria bacteria. However, in this study, synthetic urine provided
inoculated at 2 x 102/10 ml did not continue after 2 days, as less than optimal conditions for the growth of bacteria.
pilot studies showed that no survivors were cultured at Synthetic urine was chosen because it more closely mimics
periods greater than 2 days. the growth conditions found in human urine than do micro-
Bacteria in the denser inoculum (2 x 104/10 ml) survived biologic growth media. Synthetic urine can be consistently
better than bacteria in the less-dense inoculum. However, if synthesized with the same compounds and avoids the vari-
the microamperage was kept at 200 to 400 [LA, bacterial ation in constituents typical of human urine (6). However, K.
growth was reduced, slowed, or eliminated. Although pneumoniae grew very poorly under these conditions, and,
growth of E. coli and P. mirabilis was frequently eliminated as has been done by others (6), TSB was added to the
by 200 to 400 pA, only 400 VaA killed K. pneumoniae (Fig. 2). synthetic urine (Fig. 1). Growth of K. pneumoniae and the
Reinoculation of bacteria into vials undergoing gold ion- other genera should be related to the results given in Fig. 1
tophoresis showed again that 400 p.A was somewhat more because even those electrodes that we describe as being less
.effective than 50 isA at reducing or eliminating the organisms effective than gold-gold electrodes in suppressing or elim-
at either inoculum size (Fig. 3). K. pneumoniae was elimi- inating bacterial growth were still able to significantly sup-
nated more rapidly in this series of experiments than in the press growth. For example, Fig. 4c shows that K. pneumo-
previous series. niae growth is only approximately 103 to 105 CFU/ml by day
Prevention of bacterial growth with multiple-electrode ion- 3, while Fig. 1 shows that normal growth would be approx-
tophoresis. Other electrodes were combined with gold to imately 108 CFU/ml, a highly significant difference.
444 DAVIS ET AL. ANTIMICROB. AGENTS CHEMOTHER.

17500

_ 15000
E

M. 125W
0

tacos,

TIME
2s000 B AO1 tA Cx200 vA

Downloaded from aac.asm.org by on September 14, 2009

B=MO A D400FA

O 2x0o 4
* 2xl0 2

ml

I Ohr | 41w T| r
T Id | 2d | 3d | 4d T S Sd
I7d
TPA

00

10

Ohr 4 hr hr Id 2d 3d 4d SdC d 7
TIME
FIG. 2. Effect of gold electrodes on E. coli (A), P. mirabilis (B), and K. pneumoniae (C). Inocula were placed into 10-ml vials at 0 h, and
measurements of CFU were made at the indicated intervals. For K. pneumoniae (C), the CFU per milliliter represent logarithmic growth.
VOL. 33, 1989 IONTOPHORETIC KILLING OF BACTERIA 445

IL

ioLIam
t

am
I
A CA C A c*m
C C AC A C| A CjAS Cj A8 C A B C
0h3
41w 51 3d d 2d d4d
|d |d¶ lid
TIME

twofft.noculati
Ax E. coN

ta ~: BXlil
P.
.

Downloaded from aac.asm.org by on September 14, 2009

amo

20. C. K. .wwwt B

IL~~~~~~~~~~~~TM

Obt 41w |kh I|d 3d' 3d dd4d 3d Gd' 7d

FIG. 3. Growth of bacteria reinoculated into vials undergoing gold iontophoresis at 50 ,uA (A) and 400 ,uA (B). Inocula were placed into
10-ml vials at 0 h, and measurements of CFU were made at the indicated intervals. Asterisks indicate days on which vials were reinoculated
with organisms, and the CFU for those days represent inocula placed into 10-ml vials.

This work suggests that even very low microamperage can or terminate bacterial growth by generating metallic salts or
be effective in reducing or eliminating bacterial growth. This chloride-containing salts (2, 10, 11). In our system and in
effect is strongly dependent on at least two parameters. If the others, multiple types of bacterial inhibition may occur
inoculum size is above approximately 2 x 104/10 ml, 10 to 50 simultaneously. Lower microamperage (<400 ,iA) may not
,uA generally will not control the population regardless of generate enough ions or salts to overcome bacterial growth.
electrode composition; consequently, one parameter is inoc- Included in the experimental design was an attempt to
ulum size. This is in agreement with the results of our determine whether other electrodes could function as well as
previous work (4). Although we do not claim to know or better than gold. In general, gold wire made the best
exactly how many microamperes will kill a given bacterial electrode pairs, as it had excellent killing characteristics and
population, we do know that the trend of increasing micro- gave a relatively long electrode life. This study included
amperage can reduce or eliminate denser bacterial inocula. trials of other electrodes that were apparently not well suited
Iontophoresis of metals or their salts may be related to for iontophoresis because they broke early in the experimen-
interference with DNA replication or alteration of proteins tal procedures (copper and nickel). Another electrode, sil-
associated with cell membranes, as suggested by previous ver, did fairly well when it was utilized as the anode. Early
investigators (12, 13), but this remains to be determined. The breakage of electrodes in the solutions is probably related to
second parameter that affects bacterial killing is the rate of rapid oxidation, as the usual place for electrode breakage
bacterial growth. This is shown by the relative increase in was at the junction of electrolyte solution (synthetic urine)
the rate of growth of K. pneumoniae in TSB-enriched and metal with the atmosphere. We are investigating elec-
artificial urine. We speculate that other parameters (e.g., trode life span in relation to variable voltage and microam-
growth medium composition or innate genetic resistance to perage (G. M. Rao, S. Weinberg, C. P. Davis, M. D.
ions) play a role in bacterial killing. Other investigators have Anderson, and M. M. Warren, unpublished results); prelim-
provided some evidence that electrical treatment may inhibit inary results show that decreasing microamperage increases
446 DAVIS ET AL. ANTIMICROB. AGENTS CHEMOTHER.

25000 ANODE/CATHODE:
A * = A/Au
22500 -

I
I
,c O - NI/Au
20000 I
I
-= Cu/Au
I
I

17500
I 13xC/Au
I
I
I
o3 P/Au
I
15000 I
I
0 = Au-PdVAu
I
E I
I

IL 12500 I
I
I
I
I
10000 I
I
I
I
I
7500 I
I
I
I

5000 I
I
I

Ih1ISI
I
I
I
2500 I
I

_,'W04A VAN p
a.jj.m: mhr
I
I
V

l
I-
° 1 4 hr I 8 hr T

TIME
I
z z z 10,

d |2 d 13 d|4 d 1 5 d1|6d 7d

Downloaded from aac.asm.org by on September 14, 2009

E
IL
C.

TIME

TIME
FIG. 4. Effect of mixed electrodes on E. coli (A), P. mirabilis (B), and K. pneumoniae (C). Inocula were placed into 10-ml vials at 0 h,
and measurements of CFU were made at the indicated intervals. For K. pneumoniae (C) the CFU per milliliter represent logarithmic growth.
The electrode combinations Ag-Au, Ni-Au, and Cu-Au broke in solutions after days 1 to 2; consequently, these electrodes were not evaluated
after day 1.
VOL. 33, 1989 IONTOPHORETIC KILLING OF BACTERIA 447

electrode life span. None of the mixed electrodes did as well ated metallic ions. Antimicrob. Agents Chemother. 10:856-860.
as gold-gold, however, in slowing growth of or killing K. 3. Berger, T. J., J. A. Spadaro, S. E. Chapin, and R. 0. Becker.
pneumoniae. The reason for this is not clear, although we 1976. Electrically generated silver ions: quantitative effects on
suspect that the synthetic urine or the products generated bacterial and mammalian cells. Antimicrob. Agents Chemother.
with added TSB broth allowed such rapid growth that the 9:357-358.
electrodes could not overcome the organisms. In addition, 4. Davis, C. P., D. Arnett, and M. M. Warren. 1982. lontophoretic
200 ,A was used; perhaps a higher amperage would have killing of Escherichia coli in static fluid and in a model catheter
system. J. Clin. Microbiol. 15:891-894.
produced a better reduction of growth. 5. Dixon, R. E. 1978. Effect of infections on hospital care. Ann.
We believe that this work is important because it provides Intern. Med. 89:749-753.
evidence that bacteria can be eliminated under conditions 6. Griffith, D. D., D. M. Musher, and C. Itein. 1976. Urease, the
similar to those found in catheterized patients. Our design primary cause of infection-induced urinary stones. Invest. Urol.
would have the most influence on bacteria in the bladder 13:346-350.
lumen but could be effective against bacteria entering the 7. Kunin, C. M. 1987. Detection, prevention and management of
bladder lumen in relatively low numbers from either the urinary tract infections, 4th ed., p. 245-297. Lea & Febiger,
catheter lumen or the catheter exterior. As discussed re- Philadelphia.
cently by Kunin (8), investigation of ways to prevent these 8. Kunin, C. M. 1988. Can we build a better urinary catheter? N.
infections has been limited. We are investigating the mech- Engl. J. Med. 319:365-366.
anism(s) of iontophoretic killing of bacteria and bacterial 9. Nordenstam, G. R., C. A. Brandberg, A. S. Oden, C. M.
potential for developing resistance to such killing. In addi- Svanborg-Eden, and A. Svanberg. 1986. Bacteriuria and mortal-
tion, we plan to synthesize catheters that can be used in ity in an elderly population. N. Engl. J. Med. 314:1152-1156.
experimental trials with large animals. The potential risks of 10. Pareilleux, A., and N. Sicard. 1970. Lethal effects of electrical
using such catheters in humans are not known, but risks to current on Escherichia coli. Appl. Microbiol. 19:421-424.
11. Rosenberg, B., L. Van Camp, and T. Krigas. 1965. Inhibition of

Downloaded from aac.asm.org by on September 14, 2009

the bladder epithelium and to catheter drainage by obstruc- cell division in Escherichia coli by electrolysis products from a
tion seem low. However, individuals who have been sensi- platinum electrode. Nature (London) 205:698-699.
tized to gold or gold compounds would probably have an 12. Rosenberg, H. S., and S. Rosenkranz. 1972. Silver sulfadiazine:
increased risk for allergic reactions. Nevertheless, our data interaction with isolated deoxyribonucleic acid. Antimicrob.
help provide a basis for the design of a catheter that may Agents Chemother. 2:373-383.
reduce or eliminate morbidity and mortality due to catheter- 13. Rosenkranz, H. S., and H. S. Carr. 1972. Silver sulfadiazine:
related nosocomial UTI. effect on the growth and metabolism of bacteria. Antimicrob.
Agents Chemother. 2:367-372.
ACKNOWLEDGMENTS 14. Schaeffer, A. J., K. 0. Story, and S. M. Johnson. 1988. Effect of
silver oxide/trichloroisocyanuric acid antimicrobial urinary
This work was supported in part by Public Health Service grant drainage system on catheter-associated bacteriuria. J. Urol.
1-KY3-A-124850-01 from the National Institutes of Health; Ad- 139:69-73.
vanced Clinical Products; and the University of Texas Medical 15. Setia, U., I. Serventi, and P. Lorenz. 1984. Bacteremia in a
Branch, Galveston. long-term care facility. Arch. Intern. Med. 144:1633-1635.
16. Spadaro, J. A., and R. 0. Becker. 1976. Some specific cellular
LITERATURE CITED effects of electrically injected silver and gold ions. Bioelectro-
1. Becker, R. O., and J. A. Spadaro. 1978. Treatment of ortho- chem. Bioeng. 3:49-57.
paedic infections with electrically generated silver ions. J. Bone 17. Spadaro, J. A., T. J. Berger, S. D. Barranco, S. E. Chapin, and
Jt. Surg. 60:871-881. R. 0. Becker. 1974. Antibacterial effects of silver electrodes
2. Berger, T. J., J. A. Spadaro, R. Bierman, S. E. Chapin, and with weak direct current. Antimicrob. Agents Chemother. 6:
R. 0. Becker. 1976. Antifungal properties of electrically gener- 637-642.