ANTiMICROBIAL AGENTS AND CHEMOTHERAPY, Oct. 1974, p. 379-386Copyright 0 1974 American Society for Microbiology

Vol. 6, No. 4Printed in U.S.A.

Evaluation of Aquarium Antibiotic FormulationsT. J. TRUST AND DIANE C. CHIPMAN

Department of Bacteriology and Biochemistry, University of Victoria, Victoria, British Columbia, Canada

Received for publication 17 June 1974

The antibacterial activity of eight products marketed for the therapy andprophylaxis of diseases of ornamental fishes was tested. The products containederythromycin, neomycin, a nitrofuran, penicillin, sulfa compounds, strep-tomycin, or tetracycline. When used at the concentration recommended by themanufacturer, the products failed to inhibit the growth of bacterial speciesknown to be potential pathogens of ornamental fishes and failed to reducesignificantly the bacterial numbers in water containing fish. Concentrations ofthe products that were bacteriostatic were markedly higher than the in-use con-centration recommended by the manufacturer. The danger presented by theunrestricted availability of antibiotic compounds frequently used in the treat-ment of human and animal disease is discussed.

Large numbers of a wide variety of species ofornamental fishes are sold throughout the worldto hobbyists for home aquaria and are also usedextensively in a number of areas of biologicalresearch (5). Many of these fish subsequentlysuccumb to bacterial diseases, the most com-mon being fin and tail rot, furunculosis, gillinfections, piscine mycobacteriosis, columnarisdisease, and infectious hemorrhagic septicemia(6). A number of products are manufacturedand marketed for the therapy and prophylaxisof these diseases. Some of these products areliquid formulations containing such compoundsas the acridines, formaldehyde, malachitegreen, methylene blue, potassium permanga-nate, quaternary ammonium compounds, orsilver oxide, compounds that do not exhibitselective toxicity against bacterial cells. Theantibacterial activity of 14 of these liquid for-mulations was tested, and it was shown thatwhen used at the dilutions recommended by themanufacturers, the products failed to inhibitthe growth of bacterial species known to bepotential pathogens of freshwater fishes andfailed to reduce significantly the bacterial num-bers in water containing fish (18). Anothergroup of products is available to treat diseasesof ornamental fishes. These solid formulationsare reputed to contain compounds known toexhibit selective toxicity against bacteria.These include erythromycin, neomycin, a ni-trofuran, penicillin, streptomycin, the sulfacompounds, and tetracycline. Because theprevious study had demonstrated the liquidformulations to be inadequate, it was thoughtdesirable to continue the previous study to as-

certain the effectiveness of these solid formula-tions.

MATERIALS AND METHODSProducts tested. Products 1, 2, 3, 4, 5, 6, and 8

were purchased from commercial outlets in Victoria,British Columbia; product 7 was supplied by the manu-facturer. This latter product contained 6-hydroxy-methyl-2-[2-(5-nitro-2-furyl)vinyl]pyridine (NFP).Thiscompound has recently been marketed as an antibac-terial specifically useful for fish culture. The prod-ucts were marketed by three manufacturers in theUnited States for the therapy and prophylaxis ofinfections of aquarium fishes. The active antibacterialingredients in each formulation, the weight of eachantibacterial ingredient per capsule or tablet, andthe manufacturers' recommended usage concentra-tions expressed in terms of the active antibacterialcomponents are shown in Table 1. Additional infor-mation on the presence or absence of other activeingredients in product 7 was not available. Puresamples of the active, selectively toxic compoundspresent in the above formulations were also tested forcomparative purposes. Erythromycin was purchasedfrom Sigma Chemicals Co., and potassium penicillin,neomycin sulfate, sodium sulfamerazine, sodium sul-famethazine, sodium sulfathiazole, streptomycin sul-fate, and tetracycline hydrochloride were purchasedfrom Nutritional Biochemicals Corp. Pure samples ofNFP were not available. All solids were dried toconstant weight before use. Standard solutions wereprepared according to the recommendations of theAssociation of Official Analytical Chemists (4) andKavanagh (11, 12). Product 1 and neomycin sulfatesolutions were both prepared in 0.05 M tris(hydroxy-methyl)aminomethane buffer (pH 8.0); product 2and tetracycline hydrochloride solutions were bothprepared in 0.048 M KH2PO4 (pH 4.5). A 0.14MKH2PO4-K2HPO4 buffer (pH 7.0) was used to prepare

379

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

380 TRUST AND CHIPMAN

TABLE 1. Details of products tested

Wt of antibacterial Manufacturers'Producta Antibacterial ingredients" ingredients/capsule recommended usage

or tablet concn (,ug of activecomponent/ml)

1 Neomycin sulfate 10 0.3Copper sulfate 27 0.8

2 Tetracycline hydrochloride 250 6.53 Streptomycin sulfate 10 0.3

Merbromin 10 0.54 Sodium sulfathiazole 65 3.2

Acriflavine 22 1.15 Sodium sulfathiazole 166 8.1

Sodium sulfamerazine 166 8.1Sodium sulfamethazine 166 8.1Neomycin sulfate 10 0.3

6 Erythromycin 200 10.67 NFPc 7.5 0.28 Potassium penicillin 12 (200,000 U) 0.6

a Market name, control or lot number, and manufacturer are listed in laboratory protocol. Names availableon request.

b Information obtained from manufacturers' packages.c NFP, 6-Hydroxymethyl-2- [2-(5-nitro-2-furyl)vinyl]pyridine.

solutions of both erythromycin and product 6. Allother stock solutions were prepared in distilled water.

Test organisms. Aeromonas hydrophila ATCC9071, Aeromonas liquefaciens ATCC 14715, Aeromo-nas salmonicida ATCC 14174, a Cytophaga species,Mycobacterium fortuitum ATCC 9820, Mycobacte-rium marinum ATCC 927, Pseudomonas fluorescensATCC 13525, Pseudomonas putida ATCC 12633, andVibrio anguillarum ATCC 19264 were selected forstudy because these species cause infection in aqua-rium fishes (6). These species of Aeromonas, Myco-bacterium, and Pseudomonas have also been isolatedfrom a variety of clinical diseases in man (2, 10, 21).

Escherichia coli ATCC 11775, Klebsiella pneumo-niae ATCC 10031-1, Pseudomonas aeruginosa ATCC15442, Salmonella typhi ATCC 167-P, and Staphylo-coccus aureus ATCC 6538 and ATCC 6538 P werechosen for study because these species are recom-mended for assays of antibacterial compounds (4, 11,12). These species are also potential human pathogens(22). Species of Aeromonas, Pseudomonas, andEnterobacteriaceae are common in aquarium water(19). Bacillus subtilis ATCC 6633 was chosen becauseof its ability to form spores and because Bacillusspecies are common in fish feeds (20). A sporesuspension was used in assays.

Preparation of inocula. Different cultural condi-tions were required to prepare inocula of the testspecies for the antibacterial assays. Cultures of E.coli, K. pneumoniae, P. aeruginosa, Salmonella typhi,and Staphylococcus aureus were incubated at 37 C for24 h in antibiotic assay broth (BBL); A. hydrophila,A. liquefaciens, P. putida, P. fluorescens, and V.anguillarum were incubated at 30 C for 24 h. A.salmonicida was grown at 30 C for 48 h. The Cyto-phaga species, M. fortuitum, and M. marinum weregrown on Penassay agar slants at 30 C for 48 h, and

the growth on each slant was suspended in antibioticassay broth. The B. subtilis spore suspension wasprepared by the method of the Association of OfficialAnalytical Chemists (4). This suspension contained 3x 108 viable spores per ml. All other cultures wereadjusted to contain approximately 108 cells per ml.These cultures were then used to assay the bacterio-static properties of the test formulations.

Bacteriostatic activity of products. The dilutionof each test product that prevented the growth of eachtest species was determined by serial tube dilution,using the method described by Kavanagh (11). Dupli-cate twofold dilutions of the test product were pre-pared in broth. Compounds containing erythromycin,NFP, neomycin sulfate, penicillin, streptomycin sul-fate, or tetracycline hydrochloride were assayed inantibiotic assay broth (pH 7.0). Mueller-Hinton broth(BBL) (pH 7.4) was used for assays of productscontaining sulfa compounds. Each 5-ml dilution wasthen inoculated with 0.03 ml of the appropriateinoculum culture and incubated at 30 C for 96 h.Assays with E. coli, K. pneumoniae, P. aeruginosa,Salmonella typhi, and Staphylococcus aureus wereincubated at 37 C for 96 h. Tubes were then examinedfor the presence or absence of growth of the testspecies.

Antibacterial activity of products in aquariaholding fish. The ability of the products to reduce theviable aerobic bacterial count in the aquarium waterin the presence of goldfish (Carassius auratus) wasgenerally tested at one, two, and four times theconcentration recommended by the manufacturers(18). The effect of the control antibacterial com-pounds on the viable bacterial numbers was alsotested. The concentrations chosen were based on thehighest concentrations added to commercial test disksused to determine the antibiotic susceptibility of

ANTIMICROB. AG. CHEMOTHER.

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

AQUARIUM ANTIBIOTIC FORMULATIONS 381

gram-negative species. Hence, assays of the controlantibacterial compounds were generally performed atone, two, and four times this concentration. In severalexperiments,, concentrations higher than these werealso tested. The fish were acclimatized in a holdingaquarium for at least 7 days before the assays. Thebioassays were performed as recommended by theAmerican Public Health Association (3) in 6-literround glass jars, each containing 5 liters of distilledwater (adjusted to pH 7). The jars were aerated at 1liter/min and held at 22 C. Ten goldfish (meanweight, 1.2 g; range of weights, 0.9 to 1.6 g) wereplaced in each of four assay jars. The fish were placedin the test squarium 48 h before the addition of thetest formulation and were not fed until the end of thebioassay. The pH of the water in the aquaria wasbetween pH 6.5 and pH 6.8 immediately beforeaddition of the test formulation. The formulationunder test was then added at the appropriate concen-tration to each of three aquaria. The fourth wasmaintained as a control. In assays with NFP ortetracycline, the jars were covered with aluminum foilto minimize photodecomposition of the antibacterialcompound. Each jar was sampled immediately beforeaddition of the product (zero time) as well as 0.5, 1, 3,6, 12, 24, 48, and 72 h after addition of the product.The viable bacterial load present in these watersamples was determined by the drop pipette methodof Miles and Misra (13). The count procedure usedstandard methods agar (BBL) plates and duplicateserial dilutions in sterile buffered water. The plateswere incubated at 20 C for 72 h, and the colonies werethen counted.

RESULTSThe concentration of active ingredients of

each formulation required to inhibit growth ofthe test species is shown in Table 2. None of theproducts inhibited growth of all test specieswhen used at the concentration recommendedby the manufacturers. The concentrations ofthe active ingredients of product 1 required toinhibit growth of all the test species were 25 ugof neomycin per ml and 67.6 ,g of CuSO4 perml. The recommended usage concentration oftetracycline in product 2 inhibited the growth ofall species except B. subtilis, M. marinum, andP. aeruginosa. These species were inhibited bythe tetracycline in product 2 at 100 ,ug/ml.Product 3 required more than 25 ug of strep-tomycin and 50 gg of merbromin per ml toinhibit the test species, whereas product 4required 125 lAg of sulfathiazole and 50 ,ug ofacriflavine per ml. With this assay technique,product 5 failed to inhibit the test species evenwhen used at 225 ,g of each sulfa compound perml and 9.3 ug of neomycin per ml. Product 6produced bacteriostasis with 500 ,ug of erythro-mycin per ml. In product 7 more than 1.6 ,ug ofNFP per ml was required to inhibit all the test

COD13

So

cJ00.

C.

0

.sCIA

IL)O

.0

m

C.

C.)N

Co0

L.0)

C.)

CI)a.O

-

0

-o

-

00

0)

:to

.0

0

0

0

C)

.4-

r.

as

Q

in u, Lo _4 0 Lo

-4c

lc

q -4 t o OL4OLIN I LO

C0O

.- cli -4 14IRr. O -4

O-4

U- 00no ul m C1 Lone O

nt- e0M L L 0L

LO 00 "cs

LO CDsu : :e-t

os -n N - ttt~~~~~~~~~~C

tel oo £ ee eco) C'icu C; 4o~ing Lo, iLo Lo Looo

C-4~~~

42- __~~~~~~C- A~

Cs)

LOC C CrS o Lo Q o iUDLo o o __\CCC I" 1 A0~~

C' C- A-

CJC. A - > fiA"> AcqCD -CDcq U U CD

-4 ~ ~ AA A

CD CNNn e . D N N LOLf C toul L- to 0n C:>4: C1 _ 0

QE Lo AoL6L.Oo0 _

CO 't _iCSi o0 oL oa

A0"t1 oC o""0 Lo "A

S! CS L 0CI C"C C D0

LO A A 'QbOON

A AAA

Co~~~~~~~~~~~~~~~~"

.s C, CC'tO -O X

A

. CS C'n_ O C D-:_

C'ICD~~~~~~~'! t

'cq AScrcJ_ t _es N~ tt t X A

*~~~~~~~~~- Lo000 tb

~C

M L

o S A

O O0CD i u LO C4 0

..0* - rE- Ci C'S O O O u-: O C CS C'i 4 Lo O O

,.c

C')

C0r.C.

cec' t to

VOL. 6, 1974

4;

r._

C-Q:Yr.

O6C4X~.54)

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

382 TRUST AND CHIPMAN

species, whereas product 8 required more than30 gg of penicillin per ml. In all cases, thesebacteriostatic concentrations were markedlyhigher than the recommended concentrations,ranging from 8 to 83 times that recommended.The minimal inhibitory concentrations of the

standard reference antibacterial compounds areshown in Table 3. Inhibition of all test specieswas obtained with erythromycin at 500 Ag/ml,neomycin at 250 gg/ml, streptomycin at 500ug/ml, or tetracycline at 105 yg/ml. With theassay technique used, 500 Mg each of sulfamer-zine, sulfamethazine, or sulfathiazole per ml,singly or the three combined, failed to inhibitbacterial growth. Comparison of the titers ob-tained with the test formulations (Table 2) andthe control antibacterial compounds (Table 3)indicates that the concentrations of antibacte-rial compounds reputed to be present in prod-ucts 2, 6, and 8 were accurately stated by themanufacturer. In most cases identical results, orresults differing by only a single dilution tube,were obtained. It is apparent that addition ofCuSO, to the neomycin in product 1 enhancesthe antibacterial activity of the formulation,whereas merbromin enhances the antibacterialactivity of streptomycin in product 3. Acrifla-vine appears to enhance the antibacterial activ-ity of sulfathiazole in product 4, and the combi-nation of three sulfa compounds with neomycinin product 5 has markedly greater antibacterialactivity than that displayed by the individualcompounds or by triple sulfa alone.The numbers of viable bacteria in the water

of aquaria containing goldfish were reduced byproduct 1 from 107 to 103 per ml within 6 h(Table 4). However, when used at the concen-tration recommended by the manufacturer, fiveof the assay fish were dead within 90 min of theaddition of the formulation to the aquarium. Anadditional four dead fish were removed after 6h, and the, remaining fish was dead at 12 h. Atthe higher concentrations tested, all the fishwere dead 6 h after addition of the formulation.Fish mortalities were also obtained when theexperiment was repeated with a separate batchof fish. In this second experiment, three assayfish were dead after 24 h in the aquariumcontaining the concentration recommended bythe manufacturer. Another two fish were deadafter 48 h. At twice this concentration, sevenfish died in the first 6 h and the remainder weredead at 12 h. At four times the recommendedconcentration, all 10 fish were dead within 4 h.When product 2 was used at four, eight, andsixteen times the recommended concentration,the viable bacterial numbers were only lowered

co

Sz.o

0

C)

0C.C3

L.Ck)Ci

CiS

-ok

C)

-0

CS

L.CI

Ci0.

Hi

E

130

0bD._

._

DS0.0

0

C)

0._

0

D

._

as0

o

C)o0Q.

Q.rCD

Cq-

* 4D-9 U- Q-at)C-J C-I

A

A

A

Ci c: : :CSN_1C)9

A

0C' o ot0A

00) at~~~)4

Q Q00_ . 1

')eA,eX> CcqE0 o

A

es A

0*Lat) A

to~~~~~~~~~t

ciC4 C13 C

A A

onXa A LO

oEAOo o

E U| °~ *n Log

.2C.-OOONat

A

*-~Ca iC)eat)

A

5S,. -

E.a0 8-L

*-- A a4t)!Ee e~~o.wt X 8 E 8 8~~at

0

C-I~

_4 _-4 Att;O~~

......

ANTIMICROB. AG. CHEMOTHER.

0

0

._

.e r

> co

_ ca

0

Xc^0.bC) 0

_0s

._0

~CD)*@ ._

O~ *

aq 0

0._NO

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

AQUARIUM ANTIBIOTIC FORMULATIONS 383

TABLE 4. Effect of antimicrobial formulations on viable bacterial numbers in water containing goldfish

Concn of active Estimated no. of bacteria in aquarium water (x 105/ml) after

Antimicrobial agent ingredients addition of test formulationtested (Ag/ml) Oh [0.5h lh 3h 6h 12h 24h | 48h |72h

Product 1 (neomycin, CuSOj)a

Product 2 (tetracycline)

Product 3 (streptomycin, mer-bromin)

Product 4 (sulfathiazole, acri-flavine)

Product 5 (sulfamerazine, sulfa-methazine, sulfathiazole,neomycin)

Product 6 (erythromycin)

Product 7 (NFP)c

Product 8 (penicillin G)

0,00.3, 0.80.6, 1.61.2, 3.2

0234692

0,00.3, 0.50.5, 1.01.1, 4.0

0,03.2, 1.16.8, 2.212.8, 4.4

0,0,0,08.1, 8.1, 8.1,0.316.2, 16.2, 16.2,

0.632.4, 32.4, 32.4,

1.2

010.621.242.4

00.40.81.7

00.61.22.4

136165270140

49534261

84160320200

15010010871

1163643

41

5992190108

3152

35310

137221

48423838

1019216086

150631339

893925

24

667614088

30.40.40.2

370.99

a Active ingredients.-, Not tested.

c NFP, 6-Hydroxymethyl-2- [2-(5-nitro-2-furyl)vinyl]pyridine.

1370.30.20.5

37161924

1188113023

139611320

952628

25

675513078

30.60.60.04

S50.622

1.500.080.040.1

46191719

1308119034

11848814

843126

19

50729673

4630.5

4182

36

1970.020.010.02

42101212

1327412031

150379

32

1203239

33

636912068

21128

6116849

930.040.0010.005

21752

19013511021

45401824

424220

50

35407845

316

11

S149896

85110.004

6562

56195175165

65364250

1120150

170

23394356

92555

244462

312

23300

0.30.03

537

26

1362120270

5132726

65

34

40

16162756

261

1118

266033176

48

120130

1163297

0.623

5

2445384

26

2252

41

70320

to 105 viable cells per ml at 24 h. This was alsothe case when products 3, 4, 5, 6, 7, and 8 wereadded to the aquaria at four times the recom-mended concentration. Moreover, four dead fishwere removed from the aquarium containing42.4 ,ug of erythromycin per ml at the end of the72-h assay period. It should be noted thatalthough the total count was generally onlyreduced to 105 viable cells per ml in the aquaria,there was an effect on the bacterial speciespresent. Generally, one or two colony types werefound to predominate in aquaria to whichformulations had been added, whereas numer-ous colony types were present in the control

aquaria. It should also be noted that in severalassays, the viable bacterial numbers in thewater of the control aquaria decreased duringthe assay period. This probably reflects theattachment of many of the organisms to thesurfaces of the aquaria.Table 5 shows the effect of the standard

antibacterial compounds on the viable bacterialnumbers in test aquarium water. Even whenadded to the water at concentrations six timeshigher than that recommended in the case oferythromycin and tetracycline and 167 timeshigher in the case of streptomycin, the antibac-terial compounds failed to reduce the viable

VOL. 6, 1974

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

TABLE 5. Effect of standard antibacterial compounds on viable bacterial numbers in water containing goldfish

Concn Estimated no. of bacteria in aquarium water (x 105/ml)Antibacterial compound tested, after addition of test compound

(mg/ml) Oh 0.5 h 1 h 3 h 6 h 12 h 24h 48h 72 h

Erythromycin 0 15 15 23 17 14 105 4 3 215 16 11 11 13 8 6 9 12 430 22 11 8 10 8 3 8 11 860 87 30 21 21 27 13 33 20 21

Neomycin 0 17 16 19 58 18 10 24 14 4830 80 10 10 45 8 14 55 12 22460 24 0.06 0.04 0.02 0.02 2 50 12 480120 88 1 0.8 0.7 0.4 5 34 10 136

Penicillin G 0 6 7 7 9 9 14 40 26 60.6 5 7 5 4 3 5 3 5 71.2 4 6 5 4 5 6 8 16 292.4 6 5 5 4 5 9 10 31 74

Sulfathiazole 0 10 10 10 10 10 10 18 10 3300 10 10 10 11 1 7 9 5 1

Sulfamerazine, sulfamethazine, 0 3 3 3 4 2 3 9 26 2sulfathiazole 300 3 0.4 0.3 1 7 11 16 21 27

Streptomycin 0 26 37 37 41 29 28 77 24 3212.5 81 21 8 4 4 0.5 39 62 4725 79 4 5 1 4 0.5 41 25 4050 100 14 18 10 14 0.7 40 38 45

Tetracycline 0 29 4 25 34 22 17 8 3 430 12 0.7 6 3 5 2 1 3 460 16 0.8 6 5 5 2 4 6 3130 39 2 16 16 11 2 24 10 8

a Expressed in terms of active antibacterial component(s).

bacterial numbers to much less than 105 per ml.In addition, neomycin was toxic to the assayfish. In the aquarium containing 30 ,ug/ml, onefish was dead at 24 h, another was dead at 48 h,and an additional four were dead at 72 h. In theaquarium with 60 ug of neomycin per ml, onefish was dead at 48 h and another six were deadat 72 h. In the aquarium containing 120 ug ofneomycin per ml, one fish was dead at 24 h,another two fish were dead at 48 h, and anotherfour were dead after 72 h.

DISCUSSIONThe resistance of bacteria to antibiotics and

other synthetic chemotherapeutic agents hasbeen recognized for many years (9). In the lastdecade, strains of bacteria resistant to manyantibacterial agents have been isolated, and theincidence of these antibacterial-resistant orga-nisms in the environment appears to be on theincrease (15). Not unexpectedly, possession of

the property of antibiotic resistance by bacteriacausing disease in man has become a matter ofconsiderable concern.There seems little doubt that the incidence of

such antibiotic-resistant strains reflects thewidespread use of antibiotics (15). The resist-ance can arise by a number of mechanisms (9,15), but whatever the route whereby resistanceappears in a hitherto susceptible cell, selectionpressure most operate if that cell is to become amajor component of the bacterial population.There is no doubt that once a resistant cell hasemerged, selection with appropriate antibioticswhen they are present will lead to the displace-ment of the susceptible population by a resist-ant population (15). Conversely, withdrawal ofthe antibiotics commonly leads to a decrease inthe number of resistant organisms encountered(15).As a result of the increased awareness that

the development of antibiotic-resistant bacteriaimpairs the effectiveness of antibiotics in treat-

384 TRUST AND CHIPMAN ANTIMICROB. AG. CHEMOTHER.

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

AQUARIUM ANTIBIOTIC FORMULATIONS 385

ing human and animal disease, controls havebeen imposed on the sales of antibacterialagents for the therapy and prophylaxis ofhuman and animal disease (7, 14, 17). Suchrestrictions do not apply to the sale of antibiot-ics for the therapy and prophylaxis of ornamen-tal fishes. People can purchase products con-taining erythromycin, neomycin, nitrofuran,penicillin, streptomycin, sulfa compounds, ortetracycline from a large number of retail out-lets in North America for use in their householdaquaria. Even if these formulations were effec-tive in controlling fish disease, the lack ofrestrictions would not be desirable and obvi-ously deserves consideration by the appropriateauthorities.The method most commonly used to treat

diseases of ornamental fishes is to bathe the fishin the water-soluble antibacterial compound(8). Our results show that the prescribed levelsare likely to be subtherapeutic since they fail toinhibit the growth of bacterial species known tocause infections in aquarium fishes and speciescommonly found in aquarium water. The re-sults also show that markedly higher levels ofthe formulations often fail to significantly de-crease the numbers of viable bacteria in theaquarium water. Moreover, one of the moreeffective antibacterial formulations was toxic tothe fish. It is worth noting that control ofbacterial growth in household aquaria mayfurther be complicated by the presence of sand,plants, and food, since these will reduce theefficiency of the antibacterial compound bydirect inactivation or by mechanically protect-ing the bacteria from attack. In view of theapparent lack of effect on the bacteria in theaquarium water, the formulations would proba-bly be unable to reduce tha viable bacterialnumbers of surface infections of the fish. Inaddition, although few studies have directlycompared drug levels attained by differentroutes of administration, experience has taughtthat therapeutic tissue levels can rarely beattained by bathing fish in chemotherapeutics(8). For example, less than 3% of the neomycinreaching the alimentary tract is absorbedthrough the gut wall.

Various antibiotics and synthetic chemo-therapeutics have been used for cultured fish asfeed additives in various contries to prevent andtreat various infections. They have sometimesbeen administered directly into fish pond waterto control fish infections. Fish pathogens carry-ing transferable drug resistance factors havebeen reported. (1). It is likely that bacteriacarrying transferable drug resistance factorswill also emerge either in household aquaria

containing ornamental fish or in the aquaria ofthe distributors of the ornamental fishes ifantibiotics are used to treat the fish in thesesituations. This is undesirable, especially sincewater containing ornamental fish has beenfound to contain bacterial species that arepotential pathogens of man (19). The potentialpublic health risk of water containing drug-resistant bacteria can readily be seen sinceaquaria containing ornamental fish are found inschools, eating establishments, shopping cen-ters, retails stores, hospital wards, homes for theelderly, and doctors' and dentists' offices.

In addition to the potential for rapid emerg-ence of strains of multidrug-resistant bacteriaby virtue of their acquisition of transferabledrug resistance factors, a rapid emergence ofresistance to streptomycin in vitro can alsooccur by the so-called "one-step" resistancepattern (9). Gram-negative species and Myco-bacterium in particular can be made to yieldmutants with a high resistance to streptomycin(9). In fact, aquarium-borne mycobacteriosesthat are refractory to chemotherapy have al-ready been reported (2, 10). Therefore, it seemsundesirable to continue to allow the unre-stricted sale of streptomycin and other antibiot-ics for aquarium use.The philosophy of the use of bacteriostatic

antibiotics to bathe fish also needs to be reeval-uated. The in vivo success of chemotherapywith bacteriostatic compounds often dependson the ability of these compounds to controlbacterial growth until the immune response cancope with the invaders. This is not likely to beapplicable when poikilothermic species arebathed in such bacteriostatic agents. Not only isit unlikely that sufficient tissue levels of anti-bacterial agents will be attained, but the im-mune system of the poikilothermic species isnot as efficient as that of the homeotherms.Antibody titers of poikilotherms are generallylow and attained slowly (16).

In view of the questionable value of theseformulations in the therapy and prophylaxis ofinfections of ornamental fishes, and since itseems wise that all efforts should be directedtowards reducing the selection pressures favor-ing antibiotic-resistant bacteria (15), it wouldseem desirable to withdraw from the marketproducts containing antibiotics used in humanchemotherapy. This would contribute to thereduction in the reservoir of resistant bacteria inthe environment.

ACKNOWLEDGMENTSThis work was supported in part by the University of

Victoria.We thank David Burdge for assistance in the laboratory.

VOL. 6, 1974

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from

386 TRUST AND CHIPMAN

LITERATURE CITED

1. Aoki, T., S. Egusa, Y. Ogata, and T. Watanabe. 1971.Detection of resistance factors in fish pathogen Aero-monas liquefaciens. J. Gen. Microbiol. 65:343-349.

2. Adams, R. M., J. S. Remington, J. Steinberg, and J.Seibert. 1970. Tropical fish aquariums. A source ofMycobacterium marinum infections resembling sporo-trichosis. J. Amer. Med. Ass. 211:457-461.

3. American Public Health Association. 1965. Standardmethods for the examination of water and wastewater,12th ed., p. 545-563. American Public Health Associa-tion Inc., New York.

4. Association of Official Analytical Chemists. 1970. Officialmethods of analysis of the A.O.A.C., 11th ed. Wash-ington, D.C.

5. Axelrod, H. A. 1973. Tropical aquarium fishes as labora-tory animals. Trans. Amer. Fish. Soc. 102:249-250.

6. Bullock, G. L., D. A. Conroy, and S. F. Snieszko. 1971.Diseases of fishes, p. 7-9. In S. F. Snieszko and H. A.Axelrod (ed.), Bacterial diseases of fishes. THF Publi-cations, Jersey City, N.J.

7. Health and Welfare Canada. 1973. The use of antimi-crobials in animal feeds. Information letter 387. Healthand Welfare Canada, Ottawa.

8. Herman, R. L. 1972. The principles of therapy in fishdiseases, p. 141-152. In L. E. Mawdesley-Thomas (ed.),Diseases of fish. Symposia of the Zoological Society ofLondon, no. 30. Academic Press Inc., London.

9. Hughes, W. H., and H. C. Stewart. 1970. Conciseantibiotic treatment, p. 83-85. Butterworth and Co.,Ltd., London.

10. Janssen, W. A. 1970. Fish as potential vectors of humanbacterial diseases, p. 284-290. In S. F. Snieszko (ed.), Asymposium on diseases of fishes and shellfishes. Amer.Fish. Soc. Spec. Publ. no. 5. American Fisheries Society,

Washington, D.C.11. Kavanagh, F. 1963. Dilution methods of antibiotic as-

says, p. 125-140. In F. Kavanagh (ed.), Analyticalmicrobiology, vol. 1. Academic Press Inc., New York.

12. Kavanagh, F. 1973. Analytical microbiology, vol. 11.Academic Press Inc., New York.

13. Miles, A. A., and S. S. Misra. 1938. The estimation of thebactericidal power of the blood. J. Hyg. 38:732-749.

14. New Scientist. 1972. U.S. restricts antibiotics in feeds.New Scientist 53:413.

15. Richmond, M. H. 1972. Some environmental conse-quences of the use of antibiotics: or "what goes up mustcome down." J. Appl. Bacteriol. 35:155-176.

16. Saunders, G. C. 1970. Development of the immuneresponse, p. 93-135. In P. Abramoff and M. F. La Via(ed.), Biology of the immune response. McGraw-HillBook Co., New York.

17. Swann Committee. 1969. Report of the joint committeeon the use of antibiotics in animal husbandry andveterinary medicine. Her Majesty's Stationery Office,London.

18. Trust, T. J. 1972. Inadequacy of aquarium antibacterialformulations for the inhibition of potential pathogensof freshwater fish. J. Fish. Res. Bd. Can. 29:1425-1430.

19. Trust, T. J., and K. H. Bartlett. 1974. Occurrence ofpotential pathogens in water containing ornamentalfishes. Appl. Microbiol. 28:35-40.

20. Trust, T. J., and V. G. Money. 1972. Bacterial populationof diets for aquarium fishes. J. Fish. Res. Bd. Can.29:429-433.

21. Von Graevenitz, A., and J. Weinstein. 1971. Pathogenicsignificance of Pseudomonas fluorescens and Pseudo-monas putida. Yale J. Biol. Med. 44:265-273.

22. Wilson, G. S., and A. A. Miles. 1964. Topley and Wilson'sprinciples of bacteriology and immunity, 5th ed., vol 1.Edward Arnold Ltd., London.

ANTIMICRoEL AG. CHEMOTHER.

on May 29, 2018 by guest

http://aac.asm.org/

Dow

nloaded from