Outbreak of Mycobacterium chelonae
Infection Associated With Use
of Jet Injectors
Jay D. Wenger, MD; John S. Spika, MD; Ronald W. Smithwick, MS; Vickie Pryor, RN, MPH; David W. Dodson, MD;
G. Alexander Carden, MD; Karl C. Klontz, MD, MPH
Between January 1 and May 15, 1988, foot infections due to Mycobacterium February 22 and April 30, 1988. Within
chelonae subspecies abscessus were diagnosed in eight persons who had 2 weeks of podiatric procedures per¬
formed in office A (one of two offices in
undergone invasive procedures at a podiatry office. A cohort study was per- which a podiatrist treated patients), pa¬
formed to evaluate risk factors for disease. Persons who underwent procedures
tients complained of pain and redness
before 10:30 AM were more likely to have developed infection than those with
(often with a blister), followed by drain¬
procedures after that time (relative risk, 5.6). In addition, procedures involving age near the procedure site. All patients
any of the second through fourth toes were more likely to have resulted in were afebrile, but evaluation of the af¬
infection than procedures involving only the first and/or fifth toes (relative risk, fected extremity revealed tender, ery-
4.4). Persons with 0, 1, or 2 risk factors had attack rates of 5%, 14%, and 60%, thematous, and swollen areas close or
respectively. Mycobacterium chelonae subspecies abscessus organisms of the slightly distal to the procedure or injec¬
same antimicrobial resistance pattern as the patients' strains were cultured from tion site. Incision and drainage of the
distilled water in a reusable, nonsterilized container. A jet injector used to lesions yielded scanty amounts of sero-
administer lidocaine was held between procedures in a mixture of the distilled sanguineous fluid but little purulent ma¬
water and a disinfectant as recommended by the manufacturer. Inoculation of terial. Three patients with osteomyeli¬
tis had prolonged intravenous antibiotic
patients with mycobacteria by the jet injector may have only occurred early in the
day due to slow killing of the bacteria by the disinfectant. The outbreak empha- therapy and underwent bony débride-
ment. The index case required amputa¬
sizes the pathogenicity of this water-associated organism and the need for high\x=req-\ tion of the distal
level disinfection of jet injectors. phalanx of the fourth
toe. The lesions resolved over a period
(JAMA. 1990;264:373-376) of weeks to months during treatment
with a variety of antimicrobial agents.
investigation was begun after the podia¬
MYCOBACTERIUM chelonae is a pathogen and most commonly causes trist discontinued procedures in office A
rapidly growing species found occasion¬ soft-tissue infections, although osteo¬ following identification of the outbreak.
ally in soil, water, and sewage.12 The myelitis, keratitis, and disseminated A procedure was defined as anything
organism is an opportunistic human disease have been reported.** Mycobac¬ done by the podiatrist resulting in
terium chelonae has recently been de¬ breaking the skin barrier (ie, an injec¬
From the Division of Bacterial Diseases, Center for scribed as a cause of epidemic disease tion or incision). A definite case of
Infectious Diseases, Centers for Disease Control, Atlan-
ta, Ga (Drs Wenger and Spika and Mr Smithwick);
associated with specific medical or sur¬ M chelonae soft-tissue infection was de¬
Florida Department of Health and Rehabilitative Ser- gical invasive procedures.712 We de¬ fined as a foot seen by the podiatrist in
vices, Palm Beach County Public Health Unit, West scribe an outbreak of M chelonae soft- either office between January 1 and
Palm Beach (Ms Pryor); and Preventive Health Ser-
vices, Florida Department of Health and Rehabilitative tissue infections related to use of jet May 31, 1988, in which symptoms of
Services, Tallahassee (Dr Klontz). Drs Dodson and Car- injectors, in which an environmental pain, redness, and drainage at or near
den are in private practice in West Palm Beach, Fla. source of the pathogen was identified. the site of a procedure were accompa¬
Use of trade names is for identification only and does
not imply endorsement by the US Public Health Service
Patients and Methods
nied by culture of acid-fast bacteria
or by the US Department of Health and Human from the wound. Procedures on differ¬
Services. Six patients hospitalized with M che¬
Reprint requests to Meningitis and Special Patho-
ent days performed on a single patient
gens Branch, Division of Bacterial Diseases, Centers lonae infection of the foot were reported were considered separate procedures.
for Disease Control, Atlanta, GA 30333 (Dr Wenger). to one county public health unit between A probable case was defined as a foot
Downloaded from www.jama.com by guest on August 11, 2010
from which no cultures were obtained
but that otherwise fit the case definition
and had symptoms that persisted for
more than 2 weeks after the procedure.
The podiatrist and all nurses working
at either office were interviewed about
office practices during the study period.
All patient visits to office A between
January 1 and April 30,1988, were eval¬
uated by review of office visit logs and
patient charts. At office B, charts of all
patients who underwent operative pro¬
cedures were reviewed for results of
Samples of all injectable materials,
disinfectant and cleansing solutions, tap
water, and bandaging material as well
as swabs of environmental surfaces and
operative instruments were collected
from both offices. Operating procedures
were observed at office B. Production
facilities of the distilled-water manufac¬
turer supplying office A were observed,
and water samples were obtained.
Laboratory Investigation.—The 71
specimens submitted for mycobacterial
culture included individual samples or
mixed combinations of dexamethasone, Fig 1 .—Number of patients who underwent surgical procedures each week in office A during the outbreak of
bupivacaine, lidocaine, sterile water, Mycobacterium chelonae infection. The number of patients who became infected is shown in white; the
number of patients who did not become infected is shown In gray. The arrow denotes the date procedures
isopropyl alcohol, Mada Sonic disinfec¬ were discontinued in office A.
tant (a mixture of quaternary ammoni¬
um halide compounds plus the chelating
agent ethy lenediaminetetraacetate, solution before inoculation to media. injections of bupivacaine hydrochlo-
Mada Medical Products Inc, Carlstadt, Swabs from surfaces thought to be con¬ ride-dexamethasone mixture only and a
NJ), carboys of distilled water from a taminated with other bacteria were variety of operations (such as bunionec-
local supplier, tap water from various treated as above, except that a 1% sodi¬ tomies, exostectomies, and osteoto¬
sites in the clinics, jugs of distilled water um hydroxide solution replaced the mies), all followed by an injection of bu-
for daily needs that were filled from the sterile water. pivacaine-dexamethasone mixture. All
larger carboys, autoclave reservoir wa¬ Incubation and Identification.— persons who underwent a procedure
ter, and swabs of various surfaces in the All inoculated media were incubated at had their skin cleansed with povidone-
two clinics. 28°C. Isolates were identified by stan¬ iodine or alcohol and had skin wheals of
Up to 50 mL of each liquid specimen dard biochemical methods.u Minimal in¬ 2% lidocaine hydrochloride solution
was placed in a sterile, 60-mL plastic hibitory concentrations of antimicrobial raised with a hand-held jet injector. For
disposable syringe and filtered under drugs against all M chelonae subspecies persons receiving bupivacaine-dexa-
plunger pressure through a presteri- abscessus isolates14 were determined. methasone injections without additional
lized 25-mm (0.4-p.m pore size) polycar¬ Survival of M chelonae in Disinfec¬ operative procedures, the injection
bonate membrane filter (Nucleopore tant Solution.—A working solution of mixture was given through a 21-gauge
Corp, Pleasanton, Calif). Organisms Mada Sonic disinfectant was prepared needle into the deep and subcutaneous
were freed from the filter by shaking in according to the manufacturer's in¬ tissue. A sterile bandage was placed
phosphate buffer and inoculated to structions, using water from the gallon over the site. Persons receiving an inci¬
Lowenstein-Jensen slants. The suspen¬ jug known to contain M chelonae organ¬ sion had additional lidocaine injec¬
sion was inoculated to Dubos-Middle- isms. At prescribed times (0,0.5,1,2,4, tions. The operative procedure was
brook oleic acid albumin agar and to 24, and 48 hours), a 10-mL sample of the then performed with new sterile scalpel
4 mL of modified 7H-9 broth. Swabs of mixture was removed and handled as blades and autoclaved instruments. Af¬
environmental sites were each placed in described above for liquid specimens. ter the procedure, deep injections
a 16 x 100-mm screwcap tube containing of bupivacaine-dexamethasone solution
10 mL of sterile distilled water, and the were given in and around the trauma¬
tube was shaken by hand for 10 to Results tized area. The site was bandaged, and
20 seconds. The swabs were then The podiatrist routinely spent 3 days the patients were given a short course of
pressed against the side of the tube to per week at office A and 2 days per week an oral cephalosporin therapy. Al¬
express free fluid, and the suspension at office B. Two nurses also rotated though the amount of bupivacaine-
was filtered through a polycarbonate from office to office, while a third nurse dexamethasone solution prepared for
membrane filter and handled as des¬ worked 3 days per week at office A only. each type of operation was similar (2 to
cribed for the liquid specimens. Speci¬ Twenty to 40 patients per day were seen 3 mL), the precise amount used in each
mens thought to be contaminated with at office A, with 0 to 8 procedures per¬ operation was not known. However,
bacteria other than mycobacterial spe¬ formed each day. Slightly fewer pa¬ more solution was deposited in the easi¬
cies were treated for 10 minutes in 1% tients were seen at office B. ly distensible tissue on either side of the
(final concentration) sodium hydroxide Procedures.—Procedures included metatarsal bones than in the thin tissue
Downloaded from www.jama.com by guest on August 11, 2010
0 12 3 4
Fig 3.—Survival of the epidemic strain of Mycobac¬
terium chelonae organisms in disinfectant solution.
Distilled water from the gallon jug from which M
chelonae organisms were isolated was used to di¬
lute Mada Sonic disinfectant according to the manu¬
facturer's instructions. Asterisks indicate the log of
the number of colony-forming units recovered after
the disinfectant was mixed.
Association of Risk Factors* With Disease
No. of Risk Factors
Fig 2.—Attack rate (number of cases per number of separate procedures) of infection following surgical
procedures performed during each hour of the working day. The total number of procedures during each
period is indicated by n. Since no procedures were performed between noon and 1 pm, the 11:30-1:29 Interval Outcome, No. of patients
1 -hour time III
was considered a period. Well 20 18 4
Attack rate, % 5 14 60
The risk factors were (1) procedure performed by
10:30am and (2) procedure involving any of toes 2
on the medial side of the first toe and the between January 1 and April 15, 1988. through 4.
lateral side of the fifth toe. The mean time of follow-up was 2
Procedures and supplies were similar months for patients in both offices.
in most respects for both offices. Dis¬ Risk Factors.—To identify risk fac¬
tilled water was purchased from sepa¬ tors for disease, we studied records for factors(procedure before 10:30 am and
rate suppliers and poured into gallon all patients who underwent invasive procedure including any of toes 2
jugs for daily use. Nurses routinely pre¬ procedures in office A between January through 4) further increased a person's
pared the disinfectant solution—in 22, 1988 (the date of the first invasive likelihood of infection (Table).
which the jet injector tip was immersed procedure associated with infection), Laboratory Results. —Eleven of 71
before and after use on each patient— and April 27, 1988 (when procedures samples submitted for culture were pos¬
using distilled water from the gallon were discontinued in office A). Fifty-six itive for mycobacteria. Mycobacterium
jug. The solution was prepared fresh procedures were performed during this chelonae subspecies abscessus organ¬
each morning at approximately 8 AM period. Age, sex, preoperative hemo¬ isms were isolated only from water in
and discarded that evening. globin level, preoperative serum glu¬ the gallon jug of distilled water (two
A total of 726 patient visits to office A cose concentration, and the type and samples). Mycobacterium gordonae or¬
were recorded between January 1 and duration of operation were not associ¬ ganisms were also isolated from the gal¬
April 30,1988. By the end of the investi¬ ated with infection. However, when all lon jug. Mycobacterium gordonae or¬
gation, 9 definite and 1 probable case of procedures for which site of procedure ganisms grew in seven other positive
disease were detected among 8 pa¬ was known were compared, procedures samples, and an orange-pigmented,
tients. In the following analysis, defi¬ involving an inner toe (any of toes 2 slowly growing acid-fast organism that
nite and probable cases were combined. through 4) were more likely to result in was not further identified grew in two
Results are similar if the probable case infection than procedures involving other samples. No cultures of speci¬
is deleted. The first procedure associ¬ only an outer toe (toe 1 or 5,7 of 18 vs 3 of mens from office B were positive for
ated with an infection was performed on 34 procedures; relative risk, 4.4; 95% M chelonae, including those from the
January 22,1988 (Fig 1), and the last on confidence interval, 1.3 to 15). gallon jug used to store distilled water.
April 15, 1988. In 2 cases, infection fol¬ The time of the procedure was also a Cultures of all samples (including dis¬
lowed a bupivacaine-dexamethasone in¬ risk factor (Fig 2). Persons who under¬ tilled water from unopened containers)
jection without additional invasive pro¬ went a procedure before 10:30 am were from the distilled-water manufacturer
cedures. The incubation period ranged more likely to develop infection than were negative for mycobacteria.
from 14 to 42 days, with a mean of 21 those with procedures after that time The results of the investigation to de¬
days. No cases occurred in 46 patients (relative risk, 5.7; 95% confidence inter¬ termine the length of survival of M che¬
who underwent procedures at office B val, 1.3 to 24). The presence of both risk lonae subspecies abscessus organisms
Downloaded from www.jama.com by guest on August 11, 2010
in disinfectant are shown in Fig 3. The were isolated from any sample of water 3. Wolinsky E. Nontuberculous mycobacteria and
number of viable M chelonae subspecies from the distilled-water producer. Al¬ associated diseases. Am Rev Respir Dis. 1979;
abscessus colony-forming units was re¬ though M chelonae organisms were not 4. Good RC. Opportunistic pathogens in the genus
duced by 98% after 4 hours. No viable isolated from tap water in this investi¬ Mycobacterium. Annu Rev Microbiol. 1985;39:
mycobacteria could be isolated after 24 gation, they can be recovered from mu¬ 347-369.
5. Wallace RJ, Swenson JM, Silvox VA, Bullen
or 48 hours. nicipal and well-water samples.11219 MG. Treatment of nonpulmonary infections due to
Results of standard biochemical Thus, although nursing personnel de¬ Mycobacterium fortuitum and Mycobacterium
tests13 were compatible with descrip¬ nied rinsing the gallon jug with tap wa¬ chelonei on the basis of in vitro susceptibilities. J
tions of M chelonae subspecies absces¬ ter, the nearby sink (under which the Infect Dis. 1985;152:500-514.
6. Woods GL, Washington JA II. Mycobacteria
sus for all isolates. The minimum inhibi¬ jug was stored) was a likely source for other than Mycobacterium tuberculosis: review of
tory concentrations for both environ¬ initial contamination. microbiologic and clinical aspects. Rev Infect Dis.
mental and patient isolates were within To our knowledge, this is the second 1987;9:275-294.
a twofold dilution of one another. The outbreak of an infectious disease linked 7. Band JD, Ward JI, Fraser DW, et al. Peritonitis
due to a Mycobacterium chelonei-like organism as-
minimum inhibitory concentration for to jet injectors. Hepatitis B virus was sociated with intermittent chronic peritoneal dialy-
cefoxitin was 16 to 32 mg/L; for amika- transferred from person to person sis. J Infect Dis. 1982;145:9-17.
cin sulfate, 16 mg/L; for ciprofloxacin through subcutaneous jet injections in a 8. Mycobacterium chelonei infections following
hydrochloride, 8 to 16 mg/L; for rifam- weight-loss clinic.20 Experimental stud¬ eye surgery: Texas. MMWR. 1983;32:591-598.
ies in mice have demonstrated mouse- 9. Kuritsky JN, Bullen MG, Broome CV, Silcox
pin, 8 to 16 mg/L; and for doxycycline NA, Good RC, Wallace RJ. Sternal wound infec-
hyclate, greater than 32 mg/L. to-mouse transmission of lactate dehy- tions and endocarditis due to organisms of the My-
drogenase virus during subcutaneous cobacterium fortuitum complex. Ann Intern Med.
inoculation using jet injectors.21 Many 1983;98:938-939.
10. Bolan G, Reingold AL, Carson LA, et al. Infec-
Although several outbreaks of injec¬ jet injections require touching the skin tions with Mycobacterium chelonei in patients re-
tion-related M chelonae disease have of each recipient with the tip of the in¬ ceiving dialysis and using processed hemodia-
been reported, a source of the organism strument. Use of these instruments for lyzers. J Infect Dis. 1985;152:1013-1019.
identified.15"18 We identified subcutaneous or intramuscular injec¬ 11. Safranik TJ, Jarvis WR, Carson LA, et al.
tions in persons infected with hepatitis Mycobacterium chelonae wound infections after
two risk factors for disease and isolated plastic surgery employing contaminated gentian
the organism from distilled water in a B virus or human immunodeficiency vi¬ violet skin-marking solution. N Engl J Med.
reusable gallon jug. A jet injector disin¬ rus can result in detectable viral mark¬ 1987;317:197-201.
fectant solution diluted with water from ers on the skin surface after injection.22 12. Lowry PW, Beck-Sague CM, Bland LA, et al.
The World Health Organization has Mycobacterium chelonae infection among patients
this jug was the probable source of the receiving high-flux dialysis in a hemodialysis clinic
organism, which was transferred to the suggested that use of jet injectors be in California. J Infect Dis. 1990;161:85-90.
patients by the jet injector. restricted to special circumstances until 13. Kent PT, Kubica GP. Public Health Mycobac-
The reason for the increased risk of the risk of transmission of infectious dis¬ teriology: A Guide for the Level III Laboratory.
eases can be further clarified.23 Atlanta, Ga: Centers for Disease Control; 1985.
disease associated with operations on 14. Swenson JM, Thornsberry C, Silcox VA. Rap-
the inner portion of the foot is not clear, This outbreak of M chelonae infection idly growing mycobacteria: testing of susceptibility
but it may be due to the different demonstrates a different risk of jet in¬ to 34 antimicrobial agents by broth microdilution.
jectors. Use of slowly acting mycobac- Antimicrob Agents Chemother. 1982;22:186-192.
amounts of bupivacaine-dexametha¬ 15. Inman PM, Beck A, Brown AE, Stanford JL.
sone solution injected into intermeta- tericidal disinfectant solutions led to the
Outbreak of injection abscesses due to Mycobacte-
tarsal areas vs outer-toe tissues. De¬ transmission of M chelonae organisms. rium abscessus. Arch Dermatol. 1969;100:141-147.
creased local inflammatory responses in Since such bacteria are often resistant 16. Borghans JGA, Stanford JL. Mycobacterium
to many standard disinfectant agents, chelonei in abscesses after injection of diphtheria-
the intermetatarsal soft tissues may pertussis-tetanus-polio vaccine. Am Rev Respir
have resulted from higher concentra¬ use of a rapidly tuberculocidal disinfec¬
tions of the long-acting steroid in those tant, eg, 2% alkaline glutaraldehyde, is 17. Pettini B, Hellstrand P, Ericksson M. Infection
areas. The microbiologie basis for the appropriate. Such instruments should with Mycobacterium chelonei following injections.
association of disease with the time of be bathed in effective disinfectant solu¬ Scand J Infect Dis. 1980;12:237-238.
tions for at least 30 minutes between 18. Jackson PG, Keen H. Nobel CT, Simmons NA.
the procedure was demonstrated by the Injection abscesses due to Mycobacterium chelonei
laboratory studies. Patients were ex¬ use. Storing distilled water in single- occurring in a diabetic patient. Tubercle. 1981;62:
posed to the contaminated distilled wa¬ use containers will also reduce the dan¬ 277-279.
ger of contamination. 19. Carson LA, Petersen NJ, Favero MS, Aguero
ter through the jet injector disinfectant SM. Growth characteristics of atypical mycobacte-
bath. Mycobacteria present in the dis¬ We thank Robert C. Good, PhD, George P. Ku- ria in water and their comparative resistance to
tilled water-disinfectant solution were bica, PhD, and Claire V. Broome, MD, for con¬ disinfectants. Appl Environ Microbiol. 1978;36:
killed over a period of several hours. structive criticism of the manuscript; James 0. Kil- 839-846.
burn, PhD, Vella A. Silcox, MS, Ray Butler, MS, 20. Hepatitis B associated in jet gun injection: Cali-
Thus, patients injected early in the day and Charles L. Woodley, PhD, for assistance in the fornia. MMWR. 1986;35:373-376.
may have been exposed to a large inocu¬ laboratory evaluation of the mycobacterial isolates; 21. Brink PRG, Van Loon AM, Tromellen JEM,
lum, while patients injected later were and Janis Sexton for secretarial assistance. Gribnau FWJ, Smale-Novakova IRO. Virus trans-
mission by subcutaneous jet injection. J Med Mi-
exposed to fewer or no mycobacteria. A References crobiol. 1985;20:393-397.
combination of both factors—jet injec¬ 22. Zachoval R, Deinhardt F, Gurtler L, Eisen-
tion (with a high mycobacterial inocu¬ 1. Won Jin B, Saito H, Yoshi Z. Environmental burg J, Korger G. Risk of virus transmission by jet
lum) in the morningplus larger amounts Mycobacteria in Korea, I: distribution of the organ- injection. Lancet. 1988;9:189.
of steroid injection—resulted in the isms. Microbiol Immunol. 1984;28:667-677. 23. Global Advisory Group. Expanded programme
2. Lowry PW, Jarvis WR, Oberle AD, et al. Myco- on immunization. Weekly Epidemiol Rec. January
greatest risk of developing disease. bacterium chelonae causing otitis media in an ear- 16,1987;62:5-9.
It is unclear how the gallon jug be¬ nose-and-throat practice. N Engl J Med. 1988;
came contaminated. No mycobacteria 319:978-982.
Downloaded from www.jama.com by guest on August 11, 2010