Biohazard Hood to Prevent Infection During Microbiological Procedures

Transcription

1 APPUED MICROMOLOGY, Dec. 1968, p Copyright 1968 American Society for Microbiology Vol. 16, No. 1 Printed in U.S.A. Biohazard Hood to Prevent Infection During Microbiological Procedures LEWIS L. CORIELL AND GERARD J. McGARRITY Institute for Medical Research, Camden, New Jersey 81 Received for publication 6 September 1968 A microbiological hood was designed to reduce the danger of airborne infection of laboratory workers. The hood uses absolute filters to deliver sterile air in a laminar flow to the work area. An air curtain across the hood opening permits easy access but separates the worker from aerosols produced in the hood, and protects material inside the hood from contamination by room air. Tests with bacterial and viral aerosols showed that the air curtain is at least 99.96% effective in preventing airborne particles from entering the work area. A practical hood was developed to prevent airborne infection among laboratory workers. Pike et al. (9) reported 61 laboratory-acquired infections caused by bacterial, viral, rickettsial, fungal, and parasitic human pathogens with 1 deaths over a period of 1 years, and a committee of the American Public Health Association has records of over,7 cases of laboratory-acquired infections with 17 fatalities (6, 7). Wedum observed that the infective dose most often arises in an aerosol at the laboratory bench during routine laboratory procedures (1). It is believed that the present hood will adequately confine aerosols generated during normal laboratory manipulations. Another application of the hood is to reduce contamination of microbial and cell cultures. A recent survey of long-term cell cultures used for vaccine production showed between 5 and 9% to be contaminated with mycoplasma (Barile, personal communication). Bacterial, fungal, and intra- and interspecies contamination with L or HeLa cells occurs frequently (, 5). MATERIALS AND METHODS Biohazard hood. A biohazard hood measuring ft by.5 ft by 8 ft high (11.9 by 76. by.8 cm) was made to our specifications (manufactured by The Baker Co., Biddeford, Me.). The objective of the hood is to reduce or eliminate airborne infections of cultures and laboratory personnel. Airflow patterns in the hood are shown in Fig. 1. Air is passed through a high efficiency particulate air (HEPA) filter that meets government standards by removing 99.97% of dioctylphthate particles having an average diameter of. pm (). The filter extends across the entire roof of the work cabinet and air moves downward in the hood with minimal turbulence at an average velocity of 1 ft (.8 m) per min (FPM). Air leaves the work area of the cabinet through two perforated exhaust grills located at the bottom of the rear wall and at the front of the work surface. Ninety per cent of the air is recirculated through the main HEPA filter and 1% is exhausted through a second HEPA filter at the top of the hood. The exhausted air is replaced by a similar volume of room air drawn through the 1-inch (5. cm) opening of the hood. This room air does not enter the work area of the hood but passes down through the grill in front of the work area. The air curtain across the front opening moves at a speed of 17 to FPM and creates a barrier to passage of particles into or out of the hood. Generation and enumeration of bacterial and viral aerosols. Aerosol clouds of a -hr culture of Serratia marcescens were generated with a DeVilbis no. nebulizer. Approximately 18 bacteria were nebulized per test, and samples were collected on Trypticase Soy Agar in Andersen samplers or on settling plates with the filter unit turned off. The filter unit was then turned on, 11 bacteria were nebulized, and samples were taken at the same locations. Andersen samplers were calibrated to sample 1 ft of air per min. Plates were incubated for hr at 5 C, and results were expressed as colony-forming units (CFU) per cubic foot of air for Andersen samples or as CFU per settling plate. Approximately 19 plaque-forming units (PFU) of T coliphage were nebulized in a similar manner, and samples were collected in Andersen samplers or in 5- ml plastic syringes containing 1 ml of tryptose phosphate broth. Andersen samplers containedpetri dishes filled with 7 ml of gelatin, and 1 ft of air was sampled per min. After the test, the gelatin was liquified at 7 C and assayed for bacteriophage by plating on a culture of E. coli strain B. The plastic syringe collected a -ml sample of air, and the Andersen sampler inthese tests collected 8, ml of air; but in 175 nebulizations of T bacteriophage, the sensitivity of the two samplers was about equal when testing dense viral aerosols. 1895

2 1896 CORIELL AND MCGARRITY APPL. MICROBIOL. FiG. 1. Air is pushed through the main HEPA filter, flows through the work chamber at the rate of 1 ft (.8 m) per min, and exits through two grills located in front ofand at the back of the work space. Ninety per cent of the air is recirculated and 1% is discarded through a second HEPA filter. The discarded air is replaced as the blower draws an equal volume ofroom air through the hood opening to create an air curtain across the opening of the hood; this air curtain prevents airborne particles from entering or leaving the hood. RESULTS AND DIscussIoN Preliminary tests showed the hood to be effective in the removal of viable and nonviable airbome particulates. The efficiency of the air curtain was tested by controlled release of smoke and bacterial and viral aerosols at various locations inside of and outside of the hood. Smoke tests were recorded on film to indicate the most likely spots in which to look for contamination when testing with bacteria and viruses. With the filter tumed on, the number of airbome particles inside of the hood larger than.5,um were in the range of to 7 per fts of air (detected by a dust particle counter, model M-11, Dynac Corp., Portland, Me.), as compared to 5, to 1, in unfiltered room air. Tests for bacteria in the hood were usually sterile, and to obtain reproducible characterization of the performance of the hood it was necessary to generate artificial aerosols of microorganisms. In a representative test, S. marcescens was nebulized inches (5.8 cm) outside of the hood opening with the filter unit turned on; no organisms were detected by Andersen samplers and 7 settling plates in the work area inside of the hood. In a repeat experiment with the filter unit turned off, Andersen samplers collected an average of 67 colonies per ft of air and 7 settling plates collected an average of 85 colonies per plate. Colonies detected by Andersen samplers and settling plates with the unit off are illustrated in Fig.. In this experiment, 6% of the S. marcescens recovered had an average diameter of. tsm or less, as determined by the Andersen sampler. In a representative test with T phage, the virus was nebulized inside of the hood and samples were collected simultaneously inside and outside. A total of. X 17 PFU/ft were obtained in the Andersen sampler outside of the hood when the filters were not in operation, as compared to 9 PFU/ft obtained at the same site when the filter unit was turned on. At this rate, 1 PFU out of X 1' penetrated the air curtain and was detected by the sampler, for a.% penetration. The syringe sampler outside of the hood collected.7 X 1 PFU/ft with the filter off and no PFU with the filter on. The average mean diameter of the T phage particle nebulized from a DeVilbis no. has not been accurately determined, although our results with Andersen samplers showed that the pre- i ? QL18 118a 1 79 CFU/ CU. FT AIR (ANDERSEN) 1 = CFU/SETTING PLATE FiG.. Number of colonies of S. marcescens collected by three Andersen samplers and 7 settling plates inside of the biohazard hood with the filter unit off. Striped area represents exhuast grill located in front of the work area, and the nebulizer at the left represents the release site ofs. marcescens from DeVilbis no. nebulizer.

3 VOL. 16, 1968 BIOHAZARD HOOD TO PREVENT INFECTION 1897 dominant particle size of the samples collected is 1. Am or less. According to Williams and Fraser (11), the head of the T phage is 7 nm wide and the dimensions of the tail are 15 by 1 nm. It is not known how closely the aerosolized T particles approximate these dimensions. The work of Barach, cited by Dautrebande, showed that the average mean diameter of particles generated with the DeVilbis nebulizer was 6. pum, with a range between 1.6 and 1.8 pum (). To produce an aerosol of Ti coliphage with an average diameter of.1 pam, Harstad et al. used a Dautrebande DI aerosol generator (8). In their study, the penetration of Ti phage through four different HEPA ifiters averaged.%. It should be pointed out that the Dautrebunde generator is more effective in producing monodispersed submicron particles than are most nebulizers commonly employed in the generation of microbial aerosols. The particle size of aerosols produced from liquid suspensions is not determined by the size of bacteria or viruses but by the shearing force applied, the nozzle aperture, and the type and concentration of solutes in the spray suspension. The observed efficiency of HEPA filtration in removal of nebulized bacteria and viruses from air, when an effort has been made to produce an aerosol of very small particle size, makes us confident that these filters will be equally or more effective in removal of particles generated by normal laboratory procedures, such as pipetting, pouring solutions, streaking plates, etc. These operations are likely to produce only larger particles. Table 1 lists the results of 1 experiments in which we tested the penetration of the air curtain by S. marcescens. When bacteria were nebulized inside of the hood and samples were taken outside of the hood, the average number of CFU/ftt collected in the Andersen samplers with the unit on was.%o of the average CFU/ft collected at the same site with the unit off. In these tests, Andersen samplers detected airborne Serratia outside of the hood with the unit in operation in one test of five. In the positive test, one sampler outside of the hood collected two colonies of Serratia and two other samplers outside of the hood remained sterile, whereas an average of 57 CFU/ft were obtained at the same sites when the filters were not in operation. When the bacteria were nebulized inside of the hood and collected outside of the hood on settling plates, the results of eight tests showed a penetration of.6%. Bacteria nebulized outside of the hood were released inches (7.6 cm) in front of and 9 inches (.86 cm) above the work surface at the left hand side of the hood opening. The bacterial spray was directed toward the right, parallel to the opening. Andersen samplers were used to detect the nebulized organisms in six tests. Three Andersen samplers were inside of the hood in each test. As shown in Table 1, the Andersen samplers inside of the hood did not detect organisms in any test when the filter unit was running, compared to an average value of 551 CFU/ ft monitored by the three samplers when the filter unit was not running. When settling plates were used to sample the aerosol in 1 tests, an average of.1 CFU were collected per plate with the filter unit in operation as compared to 5 CFU/plate with the filter turned off, or.8% penetration. The results of spraying T phage inside of the hood are listed in Table. Phage was not detected outside of the hood with the filter unit turned on in eight tests and it was detected in two tests. In one of the positive tests, 9 PFU/ft were monitored with the filter unit running and. X 17 were monitored with the unit off, a difference of more than 5 logs. On the other occasion,.5 X 1 PFU/ft were observed with the unit on and. X 18 PFU/ft with the unit off, a difference of logs. The average penetration of phage through the air curtain in 1 tests was.8%. In another test, twenty 5-ml Erlenmeyer flasks containing tryptose phosphate broth were left open in the laboratory and such flasks were left in the biohazard hood with the unit turned on. All flasks in the hood were sterile after 5 days, whereas 17 of flasks in the laboratory were contaminated. Tests to show the absence of turbulence within the hood were conducted as follows. The work surface of the hood measures 5 by 16 inches (11. by.6 cm) and was covered with standard petri dishes in 1 rows of each. With the petri dishes exposed and the filter unit turned on, S. marcescens was nebulized 18 inches (5.7 cm) above the work surface in the middle of the hood, with the spray directed toward the right. Bacteria were recovered from dishes located to 18 inches ( to 5.7 cm) to the right of the release point, whereas dishes beyond this point and those located to the left of the nebulizer remained sterile. When this test was performed with the filter unit turned off, every dish in the hood was heavily contaminated as were dishes exposed on a table in front of the hood. The force of the spray jet carries organisms 18 inches (5.7 cm) to the right because the jet emerges at a speed of 1, FPM and slows to less than the speed of the air stream at 18 inches (5.7 cm). When the nebulizer was placed 6 inches (15.

4 1898 CORIELL AND MCGARRITY APPL. MICROBIOL. TABLE 1. Penetration of air curtain of biohazard hood by nebulized S. marcescens Expt no. lb 5 1c ld S 6 1'e No. of bacteria detected (CFU) with HEPA filtera On Off >5 >5 > Penetration < a CFU refers to colony-forming units or the number of viable bacteria recovered based on colony count. bin this group of five experiments, bacteria were nebulized inside of the hood and collected outside of the hood opening with Andersen samplers. The Andersen sampler measures CFU per cubic foot of air; each experiment is the average of six to eight Andersen samplers, each containing six petri dishes. c In this group of eight experiments, bacteria were nebulized inside of the hood and collected outside of the hood opening with settling plates. Setting plates measure CFU per plate; each experiment is the average of 8 settling plates. d In this group of six experiments, bacteria TABLE. Penetration of the air curtain of a biohazard hood by nebulized T bacteriophagea Expt no. 1 S No. of phage (PFU/ft)b detected with HEPA filter On.5 X 1 9. X 11.5 X 1 Off 1.7 X 17.7 X 17.8 X 1S 5.5 X X 18. X 18 X 18.7 X 1W. X 18. X 17.9 X 18 Penetration.1..8 a Bacteriophage nebulized inside of the hood and collected outside of the hood with Andersen samplers. bpfu refers to plaque-forming units or the number of viable bacteriophage recovered per cubic foot of air sampled. cm) above the work surface and directed toward the back of the hood, only 1 to rows of petri dishes were contaminated (Fig. ). In another test with the filter unit in operation, Serratia was nebulized 6 inches (15. cm) above the work surface in the middle of the hood with the spray directed toward the right. Two Andersen samplers were positioned inches (7.6 cm) to the front and rear of the release site, and one Andersen sampler was placed 8 inches (.) to the left of the nebulizer. The orifice of the Andersen sampler was approximately 8 inches (. cm) above the work surface and inches (5.8 cm) above the nebulizer. No viable organisms could be detected by any of the samplers when the hood was in operation. With the unit turned off, the samplers in front of and behind the nebulizer collected 575 and 1,177 CFU/ft, respectively, whereas the sampler located 8 inches (.) to the left of the release site collected 769 CFU/ft. These tests suggest that with the filter in operation an infectious aerosol produced in one portion of the hood during a microbiological procedure is not apt to contaminate an open container in a distant area within the hood, bewere nebulized outside of the hood opening and collected inside of the hood with Andersen samplers. e In this group of 1 experiments, bacteria were nebulized outside of the hood opening and collected inside with settling plates.

5 VOL. 16, 1968 BIOHAZARD HOOD TO PREVENT INFECTION 1899 FIG.. With the hood in operation and the work surface covered with open petri dishes, S. marcescens was nebulized 6 inches (15. cm) above the work surface at the spot indicated by the arrow. After incubation of the plates, heavy growth of bacterial colonies was observed on one row of plates and a few colonies were observed on one adjacent row. All of the other plates remained sterile. cause the aerosol does not spread vertically or laterally from its site of origin. The air velocity in the hood produces approximately 1,6 air changes per hr, which flushes airborne contaminants away from their site of origin and out of the chamber in a fraction of a second. This is important because we believe that most of the contamination of cell cultures occurs via the aerosol route and that most static or turbulent flow hoods in common use increase, rather than decrease, this possibility by confining the airborne particles within the air over the work area. All of the tests reported above indicate that the air curtain of the biohazard hood has an efficiency of 99.96% or greater under the conditions of the tests. In actual use, the hands and arms of the operator penetrate this air curtain and may lower its efficiency or mechanical equipment used to generate high-speed air currents may breach the air current. To test this, we carried out 1 experiments inside of the hood with a Waring Blendor containing S. marcescens or T coliphage. This procedure did not produce sufficient aerosol to challenge the hood, because in only one test were a few organisms recovered outside of the hood opening. To measure contamination of room air caused by rapid removal of hands and contaminated objects from the hood, the blendor was operated for 1 min and then the blendor lid was removed and kept in the hood while samples were collected outside of the hood opening. This was repeated, but the blendor lid was immediately removed from the hood and held over Andersen and syringe samplers just outside of the hood opening. A few organisms were recovered in two tests, and all samples were sterile in nine tests when the hands and blendor were in the hood. When the hands and Waring Blendor lid were immediately removed from the hood, all samples were sterile in 5 of 11 tests. Three Andersen samplers were sterile in two tests and contaminated on one test. Syringe samplers detected organisms in five of eight tests under these adverse conditions. It is obvious that good microbiological techniques should be employed while working in the hood in order to permit the airstream to remove aerosols and to prevent contamination of the hands. This can be supplemented by hand washing or by use of surgical gloves as an added preventive measure. The capability of the present hood to remove generated aerosols can be mathematically expressed by the logarithmic equation used by Bourdillon and Colebrook (1). The rate of biological decay of a substance at any moment is proportional to the concentration present at that moment and to the rate of removal. In equation form this may be expressed: 18 t T = 1 log1on1 - logion where T = equivalent air changes per hour, and t = the time in minutes between the moments at which the number of bacteria-carrying particles are n and n (1). Each successive air change reduces airborne contaminants in the ratio l/e; i.e., 1/.7. Three air changes decrease it to 1l/e or approximately 1/, and six air changes reduce it to approximately 1/. In the present system,

6 19 CORIELL AND MCGARRITY APPL. MICROBIOL. 1,6 air changes are employed per hr, and the airborne concentration of contaminants is reduced to l/el6l per hr or 1 /e17 per min. The theoretical reduction is approximately 1 per 5 X 111 or 5 X 1-11 per min. These figures do not consider the natural death rate of aerosolized microorganisms. If desired, a glove panel can be attached to the front opening of the hood to convert it to a class III hood. Use of this panel is recommended when extremely hazardous manipulations are to be performed. With the glove panel in place, the hood is airtight and may be sterilized with ethylene oxide. An outside cock is provided for introducing ethylene oxide and a pressure gauge indicates when the gas pressure reaches 1 inch (.5 cm) of water. Many tests of the air discarded through the small HEPA filter on top of the hood while bacteria and viruses were being nebulized in the hood were always negative for viable organisms. The exhaust can, therefore, be safely discharged into the laboratory, or, if desired, it can be connected to any discharge duct system. The work surface is a stainless-steel pan depressed.5 inch (1.7 cm) to collect liquids accidentally spilled and to permit chemical decontamination of the work surface. The average useful life expectancy of HEPA filters in continuous use in this laboratory is to years; therefore, maintenance is not a significant problem. ACKNOWLEDGMENTS This investigation was supported by grants from The John A. Hartford Foundation and by Public Health Service grant CA 95-9 from the National Cancer Institute. Suggestions for the design and performance objectives of the Biohazard Hood were contributed by Max Hesselgesser and by Arnold G. Wedum, and their encouragement and assistance are gratefully acknowledged. We thank Arlene Valloreo for technical assistance. LITERATURE CITED 1. Bourdillon, R. B., and L. Colebrook Air hygiene in dressing rooms for burns or major wounds. Lancet 1: Coriell, L. L Detection and elimination of contaminating organisms. Natl. Cancer Inst. Monograph 7, p Dautrebande, L Microaerosols. Academic Press, Inc., New York.. Decker, H. M., L. M. Buchanan, L. B. Hall, and and K. R. Goddard Airfiltrationof microbial particles. Am. J. Public Health 5: Gartler, S. M Genetic markers as tracers in cell culture. Natl. Cancer Inst. Monograph 6, p Hammon, W. M Human infection acquired in the laboratory. J. Am. Med. Assoc. : Hanson, R. P., S. E. Sulkin, E. L. Buescher, W. McD. Hammon, R. W. McKinney and T. H. Work Arbovirus infections of laboratory workers. Science 158: Harstad, J. B., H. M. Decker, L. M. Buchanan, and M. E. Filler Air ifitration of submicron virus aerosols. Am. J. Public Health 57: Pike, R. M., S. E. Sulkin, and M. L. Schulze Continuing importance of laboratory acquired infections. Am. J. Public Health 55: Wedum, A. G Airborne infection in the laboratory. Am. J. Public Health 5: Williams, R. C., and D. Fraser Morphology of the seven T-bacteriophages. J. Bacteriol. 66: 58-6.