Federation of American Scientists Working Group on BW Verification

Estimate of

The Number of Declared Facilities¹

Revised September, 1997

In order to provide a more explicit basis for determining the feasibility of the BWC compliance measures under consideration by the Ad Hoc Group, we have made a systematic, detailed survey of the number of US sites containing/working with agents listed in the Rolling Text, the number of US sites with microbial production capability greater than 50 liters, and the number of US BW defense facilities, including contractors. The sources of the information are included. Because of time limitations we have not attempted to calculate exact figures for each sub-category, but have obtained close approximations and have erred on the high side in determining the approximate maximum number of each type of facility. These maximum estimates include a fair amount of overlap. The summary table below presents estimates of the probable numbers of US facilities (after discounting these factors). The degree of uncertainty in these numbers is approximately 20%.

To arrive at estimates for the probable numbers of declarable facilities in all States Parties, it is assumed that the total numbers of facilities with listed agents or microbial production capabilities are three times the US total, and the number of BW defense facilities and contractors is twice the US total (see footnotes for the conditions and assumptions in the table). These are the weakest assumptions made. (The order of magnitude of the total, however, is not highly sensitive to these factors--eg, if the world totals were all four times the US figures, the net world total would be 39% greater).

THE EXPECTED NUMBER OF DECLARED FACILITIES WORLDWIDE IS OF THE ORDER OF 2500

Conclusions

1. The five facility declaration triggers employed here, chosen to capture the most relevant dual-use facilities, would not capture an unmanageable number of facilities. Properly designed declarations based on these five facility triggers could be broken down by computer analysis to yield additional categories, such as vaccine facilities and high containment facilities.

2. Identification of the approximately 900 US facilities requiring declaration would not be difficult. About 300 of these are commercial.

3. The total number of declared facilities, worldwide, under a BWC protocol with these triggers, will be substantially smaller than the number of declared chemical facilities under the CWC. It would be appropriate to adopt a provision for assistance to States Parties that request help in compiling their lists of declared facilities.¹⁸

4. With 2500 declared facilities, a small number of declaration visits (eg, 50/yr) would be effective to ensure accurate declarations and deter misuse of declared facilities. Declaration visits could be determined through a random, weighted process with quota limits.

_______________________________________________________________________________________________

¹ Proposed definition of "facility:" A largely independent unit consisting of all operations and services (whether or not a service is shared with other units), that are relevant to one program (e.g., a line of production), and located at a single physical site. If more than one unit at a site is declarable, then all declarable units should be considered as a single "facility." If a unit's program is carried out at more than one site, then the unit occupies more than one "facility," and each will declare separately. Institutions should have some flexibility in defining their facilities, since all are different.

² See attached Estimated Number of US Sites With Agents Listed in the Rolling Text, pages 4-11.

³ Because the US has a much larger number of pharmaceutical and biotechnology companies and research activities than any other country, it is assumed here that the numbers of related facilities, worldwide, are less than three times the US numbers.

⁴ Priority agents are those listed in the draft Rolling Text of June, 1997 (see appendix to "Estimated Number of US Sites Containing/Working With Agents Listed in the Rolling Text," pages 10 and11).

⁵ Purely diagnostic or therapeutic facilities are excluded here. They should not have to declare unless another facility at the same site is declarable, provided that isolates of listed agents are destroyed immediately following diagnosis and no live samples are retained.

⁶ Small quantities of listed toxins, some of which are widely used in research, are excluded here as triggers. A low threshold would eliminate almost all users. Toxins themselves are not self-reproducing, and small quantities below the threshold would not be useful as biological weapons. However, the presence of toxin-producing microorganisms, in any quantity, should trigger declaration.

⁷ See attached Estimate of US Sites with Microbial Production Capacity Greater than 50 Liters, pp 12-14.

⁸ See footnote 3.

⁹ Active, bonafide human food or drink production facilities are excluded; their production could not be interrupted safely for pathogen production on a routine basis. Although they could conceivably be designed for future con-version, their inclusion would not be practical. In any event, such facilities would be subject to challenge inspection.

¹⁰ Microbial production capacity of 50 l is generally considered adequate for significant BW production. There are also many facilities with lesser capacity that could be scaled up quite readily for BW production.

¹¹ Trigger 2 captures vaccine facilities, which can readily be compiled electronically from declaration forms without the need for a separate trigger, which some believe might damage their public image.

¹² Estimate of overlap is based on the data in the attached Agent and Production estimates.

¹³ For the approximate number of facilities, including contractors, in the US Biological Defense Program, see the attached "Estimated Number of US Sites Containing/Working With Agents Listed in the Rolling Text," items 1 and 2: a total of 83 + 79 + ? sites, estimated as approximately 200.

¹⁴ For the following reasons, it is assumed that the number of biological defense facilities worldwide is roughly double that of the US, and that overlap with other triggers is also roughly double. In BWC CBMs, the US has reported 24 biological defense facilities and all the other Parties, 27 (I. Hunger, p. 81 in "Strengthening the BWC: Key Points for the 4th Review Conference," ed. G. Pearson, M. Dando, Quaker UN Office, Geneva, 1996. The US Defense budget is approximately equal to that of the rest of the world. Based on historical evidence (eg, the former Soviet Union), in some countries the number of biological defense facilities may be disproportionate to the overall defense budget, but there are also many Parties that have no biological defense program at all.

¹⁵ For the overlap between BW defense facilities/contractors and trigger 1, see the items cited in footnote 13, but select the number of sites using listed agents: 55 + <20. For overlap with trigger 2, it is estimated that there are at most several microbial production facilities > 50 l under the BDP. Thus, total overlap is probably about 70.

¹⁶ Definition of the aerosol trigger needs futher consideration. It is assumed here that it will exclude small scale industry testing and routine agricultural spraying. There may be substantial overlap with trigger 3.

¹⁷ Transfers of listed priority agents and listed priority equipment for microbial production and aerosol work should be declared. The equipment list should be such that, except for manufacturers, the sources of the transfers would already be declared under other triggers, and the recipients would become declarable upon receipt (if they were not already). Thus, this trigger will add only producers of listed equipment for export to the total number of declared facilities. The number will depend on types of equipment listed and export activity, and is unpredictable at present.

¹⁸ A provision for aid in compiling declarable facilities could be adopted similar to that in the CWC Verification Annex, Part IX-A, paragraph 7.

Number of Declarable US Facilities

1. With listed agents:

< 60 Commercial Biologics Producers (for human/animal use, licensed or not)

< 50 Other Commercial Sites

<200 Research Sites (human agents); much overlap

<200 Research Sites (animal agents); much overlap (126 of these use Vesicular stomatitis virus (widely used as a model in research)).

<100 Research Sites (plant agents); much overlap (~45 of these use Xanthomonas campestris)

< 75 Biological Defense Program sites, government and contractors

____

< 685 Total

Probable number of US sites with listed agents: ~ 550

(Could be reduced to about 379 by removing two cited agents from list)

2. With Microbial Production Capability >50 Liters:

~95 Commercial Biologics Producers

<90 Other Commercial Sites (pharmaceuticals/chemicals)

<107 Research and Government Sites

______

< 292 Total

Probable number of US sites with microbial production capacity >50 liters: ~ 250 (About 30 of these also possess listed agents)

3. BW Defense Facilities and Contractors

200 (70 of these fall into the two categories above)

_____________________________________

Grand Total ~1000, less 100 overlap:

~ 900 Declarable US Facilities

(excluding aerosol facilities)

Listed Priority Pathogens as Triggers for Declarations:

ESTIMATED NUMBER OF US SITES CONTAINING/WORKING WITH AGENTS LISTED IN THE ROLLING TEXT

(Purely diagnostic or therapeutic facilities excluded)

Federation of American Scientists Working Group on BW Verification

August 1997

The US government controls nearly all the listed pathogens, and their locations are on record under:

1. US Department of Agriculture, Animal and Plant Health Inspection Service:

a) All the listed animal pathogens are restricted to a small number of sites, which must obtain permits.

b) Permits are required for all the listed plant pathogens.

c) Veterinary biologics producers must be licensed for interstate commerce.

2. US Food and Drug Administration (FDA): Producers of biologics for human use must be licensed, for interstate commerce.

3. US Centers for Disease Control (CDC): Requires registration of all sites transferring or receiving listed human agents.

4. US Department of Defense, Chemical and Biological Defense Program: all its primary sites and its many contractors, and their projects, are accessible for compilation by government.

There are a number of other readily accessible sources for information on commercial and research activities with listed pathogens. These include Federal research funding agencies such as NIH, NSF, USDA, etc.

1. Participants in the (Medical) Biological Defense Research Program

(Source: Environmental Impact Statement, BDRP, April 1989*)

3 Primary (government) sites; those using agents: 3

80 contractors; those US sites using agents: 52

(9 government, 5 commercial, 38 non-profit)

*Later detailed information is not publicly available. However, the number of sites has probably not increased significantly, if at all, since the year reported (1988), for which total obligations were $ 61,760,000, compared to the total medical BDRP budget for 1995: $ 50,213,000.

2. Participants in the Non-Medical BDRP

(Sources: US Annual BWC CBM report, April, 1995; DARPA)

Total: 79 sites, plus secondary contractors; those using listed agents: <15

(Numbers of sites not given); those using agents: est. <5

3. Non-BDRP US Government Sites

DOD, FDA, CDC, NIH, USDA, etc. (some have several sites): Total with >50 l, est. 20

4. Licensed Biologics Producers

(Sources: FDA: "Establishments and Products Licensed Under Section 351 of the Public Health Service Act, 1 March 1991;" Center for Veterinary Biologics, APHIS, USDA: "Veterinary Biological Products, Licensees and Permittees, February 1997;" K. Johnston, Bayer Agriculture Division)

Licensed US producers using listed agents for human use: <10

(Agents used: anthrax, botulinum, cholera, diptheria, plague, rabies, tetanus, yellow fever. Some probably use avirulent or attenuated strains or cloned antigens, which are exempt)

There are < 43 licensed US producers using listed agents for veterinary biologics; some of these no doubt use only attenuated or avirulent organisms. Twenty companies produce products for further manufacture; many of the others probably buy and re- formulate, using only killed, attenuated or avirulent organisms. Those handling live, virulent listed agents, <40

There are also some very small unlicensed intrastate companies, but few are likely to handle listed agents. Those that do, est. <10

5. Other Commercial Sites

(Sources: Institute for Biotechnology Information (IBI): "Pharmaceutical Industry Guide, 1997," "Biotechnology Industry Guide, 1996," and telephone consultation; Stanford Research Institute: "SRI Directory of US Chemical Producers, 1989;" expert consultation)

IBI lists <15 pharmaceutical companies that could possibly use listed agents;

IBI lists 1330 biotech companies (excluding pharmaceuticals and including many very small start-up companies, often without marketed products); 318 of these work in areas that might involve pathogens of any kind; est/ <<50 use listed agents;

There are fewer than 100 companies working on antimicrobials, of which <15 may use listed agents for testing products;

SRI and experts: 200-250 companies use fermentation for producing chemicals of any kind (including pharmaceutical and biotech companies but not biologicals producers); of these, <<50 are estimated to use listed agents.

Other (non-biologicals) commercial sites using listed agents: Total, est. <50

(Subtotal: sites (excluding non-BDRP research sites) with listed agents: <205)

6. Non-BDRP Research Sites

ANALYSIS BASED ON INDIVIDUAL AGENTS

These are mainly academic research institutions. There is some overlap with BDRP contractors; a few government laboratories are also included. Thus, there is some double-counting in adding these data to the above categories.

Human pathogens (43 listed; see page 10)

(Sources: expert opinions and Medline literature search)

a) Viruses (17 listed)

The following agents require BL4 containment:

1. Crimean-Congo hemorrhagic fever virus

4. Ebola virus

5. Hantavirus (Hantaan)

7. Junin virus

8. Lassa fever virus

9. Machupo virus

10. Marburg virus

12. Tick-borne encephalitis (Russian spring-summer) virus

17. Kyasanur Forest fever virus

There are 9 BL4 labs in the US, plus 3 or 4 that may work with BL4 agents in enhanced BL3 facilities.

Therefore, about 13 sites might use these agents. Those using listed human viruses requiring BL4 containment, 13

No. 11, Rift Valley Fever, is restricted by USDA to 2 or 3 sites: 3

No. 13, Smallpox, is restricted by WHO to 1 site: 1

Nos. 14 (Venezuelan encephalitis virus) and 16 (Yellow fever) are usually studied in attenuated form or as incomplete genomes; Those using whole virulent agents, based on 3-yr. literature search: no. 14: 6 no. 16: 5

Nos. 2 (Chikungunya) and 6 (Japan. encephalitis) require BL3 or enhanced; Those using no. 2, expert est. 6 Those using no. 6, expert est. 6

Nos. 3 (Eastern encephalitis virus) and 15 (Western encephalitis virus) are indigenous to the US. Based on 3-yr literature search, Those using no. 3, 10 Those using no. 15, 6

b) Bacteria (8 listed, all endemic in parts of the US)

No. 1, Bacillus anthracis: exp. est. <<50 sites; est. based on 3-yr lit. search, 9

No. 2, Brucella: exp. est. <<50 sites; est. based on literature search, 13

No. 3, Chlamydia psittaci: expert est. <10

No. 4, Clostridium botulinum: expert est. <20

No. 5, Francisella tularensis: expert est. <10

No. 6, Pseudomonas mallei: expert est. <10

No. 7, Pseudomonas pseudomallei: expert est. <10

No. 8, Yersinia pestis: exp. est. <20 sites; based on 3-yr literature search, 9

c) Rickettsiae (3 listed)

No. 1, Coxiella burnetti: expert est. <<10 sites

No. 2, Rickettsia prowazekii: expert est. <<10 sites

No. 3, Rickettsia rickettsii: expert est. <<10 sites T

otal number of sites for all three (considerable overlap expected): <10

d) Fungi (1 listed)

Histoplasma capsulatum: exp. est. <60 sites; based on literature search, about 20

e) Toxins (14 listed)

Only toxin-producing organisms and significant toxin stocks considered here. (Source: Dr. Jack Melling, Director, Salk Institute Biologicals Development Center)

Toxins are widely used in research, but in very small quantities, which are almost always bought from a few suppliers. There are very few companies in the toxin- producing business, including those making products for medical use. Botulin and ricin are the most commonly used. Botulin: only two or three companies make botulin (of which it is estimated that several tens of thousands of vials, containing of the order of 10-20 ng, are sold each year in the US). Thus, even a low threshold would eliminate almost all users. Ricin: for medical purposes, only the A chain (incomplete toxin) is used; for production, the A chain is probably cloned in an otherwise harmless bacterium.

Total number of sites with organisms that can produce listed toxins, and/or with significant toxin stocks: est. <20

(Subtotal: non-BDRP research sites with listed human agents: <197 )

(There is tremendous overlap here. Many laboratories study a variety of pathogens.)

Animal pathogens (18 listed: see page 11)

(Sources: Animal and Plant Health Inspection Service (APHIS), US Dept. of Agriculture; Dr. L. King, Dean, Michigan State U. School of Veterinary Medicine, former Director of APHIS)

No. 9, Herpes B (monkey) is a primate virus requiring BL4 for propagation. Number of sites, est. <5

All other listed animal pathogens are restricted by APHIS (USDA) to a few sites:

No. 3, Bluetongue virus 30

No. 11, Newcastle disease virus 24

No. 18, Vesicular stomatitis virus 126

All others, total 15

(Subtotal: Non-BDRP research sites with listed animal agents: <200)

(Considerable overlap here)

Plant Pathogens (19 listed; see page11)

(Sources: Plant Protection Quarantine, APHIS, USDA: Web site and telephone consultation)

For more than 20 years, USDA has required permits for all the agents on the list. The permits are valid for two years and must be renewed unless the agent is de- stroyed. The permits specify no further distribution. In 1996 there were 147 permits issued for plant pathogens that are exotic or not widely distributed in the US; about 51 of these were for listed pathogens. Thus, about 102 such permits are issued in two years, and this is equivalent to the approximate number of currently valid permits for listed agents. Many of these permits were issued to the same person or institution; thus, there is considerable overlap.

Total number of sites with listed plant pathogens (corrected for overlap) about 75

(Subtotal: Non-BDRP research sites with listed plant agents: 75)

_________________________________________

Grand Total:

BDRP, government and commercial sites <205

Non-BDRP research sites with listed human agents <197

Non-BDRP research sites with listed animal agents <200

Non-BDRP research sites with listed plant agents 75

Approximate Maximum Number of US Sites with Listed Agents, TOTAL <677 (This number is the sum of high individual estimates with significant overlap)

PROBABLE TOTAL No. OF US SITES WITH LISTED AGENTS, about 550

Appendix

LIST OF AGENTS AND TOXINS

from Rolling Text, 1 August 1997

Human Pathogens

Microbial Production Capability as a Trigger for Declarations:

Estimate of the Number of

US SITES WITH MICROBIAL PRODUCTION CAPACITY GREATER THAN 50 LITERS

(Active producers of human food and drink excluded)*

Federation of American Scientists Working Group on BW Verification - August 1997

1. Biologics Producers

a) Licensed Producers of Biologics for Human Use

(Source: FDA, "Establishments and Products Licensed Under Section 351 of the Public Health Service Act, 1 Mar. 1991)

About 22 companies make licensed products requiring microbial production, some probably with less than 50 l. capacity; there may also be a few unlicensed producers.

Total with >50 l fermentation capacity:                                                                                                                                                                    about 25

b) Producers of Veterinary Biologics and Feed

(Sources: Institute for Biotechnology Information (IBI): "Biotech Industry Guide, 1996"; Center for Veterinary Biologics, APHIS, USDA: "Veterinary Biological Products, Licensees and Permittees, Feb. 1997; K. Johnston, Bayer Agriculture Division)

IBI lists 26 companies that make products "using living cells" for "animal agriculture," some of which may carry out microbial production (eg, to produce feed supplements such as amino acids) (Note: all feed ingredient manufacturers are licensed by the States, from which exact information can be obtained.)

Total number of microbial feed, etc. producers with >50 l capacity:                                                                                                                    est. <20

There are 124 veterinary biologics companies licensed by USDA for interstate commerce; some of these produce serum or other non-microbial products not involving fermentation; some buy and reformulate pre-made materials--about 41 companies produce microbials for further manufacture. Only about 20 companies produce 99% of the marketed products. There are also a number of very small unlicensed intrastate companies. Few other than the largest companies have >50 l capacity.

Total no. of vet. biologics producers with >50 l microbial production capacity:                                                                                                 est. 30

c) Producers of Agricultural Biologics (biopesticides, plant innocula, etc.)

(Sources: IBI (ibid); EPA, Biopesticides and Pollution Division, Office of Pesticide Programs; Farm Chemicals Handbook '96)

IBI lists 55 biotech companies in the field of "plant agriculture"; not all are likely to use fermentation, and fewer still with >50 l capacity.

EPA registers all manufacturers of microbial products for pesticide or plant innoculum use; there are 146, many of which only buy and reformulate (eg, 116 companies have products containing B. thuringiensis, but only 11 companies worldwide produce it). Some producers have less than 50 l capacity.

The Farm Chemicals Handbook lists about 21 US companies producing microbial biocontrols.

Total no. ag. biologics producers with >50 l microbial production capacity:                                                                                                   about 20

2. Other Commercial Sites (pharmaceutical and biotech companies and producers of industrial, agricultural, etc. chemicals of any kind using fermentation)

(Sources: IBI: "Pharmaceutical Industry Guide 1997" and "Biotech Industry Guide, 1996," and telephone consultation; Dr. Lynn Klotz, biotechnology industry consultant and company founder; SRI International: "Directory of Chemical Producers, USA," 1988 and 1989; and Dr. Leo Zeftel, DuPont (retired) and CWC expert)

IBI lists 1330 US companies "using living cells for commercial purposes" of any kind (excluding companies that are primarily pharmaceutical), of which about 900 have >15 employees. A great many do not yet have marketed products.

About 16 of the listed companies cite fermentation (contract work) as their primary focus; some of these may have >50 l capacity:                                                                                                                                                                                      total <16

IBI also lists 108 pharmaceutical companies.

According to IBI, few pharmaceutical or biotechnology companies use fermentation for production; many use living cells for small-scale research purposes only.

According to Klotz, few biotech companies use fermentation for production; small companies that do are mostly bench-scale (<50 l).

According to SRI and Dr. Zeftel, about 200-250 producers of chemicals of any kind use fermentation for production. Most of those that do are small companies producing specialty chemicals on a scale <50 l. The number of large companies using fermentation is small. For example, there are 33 sites producing antibiotics in commercial quantities, but some use chemical methods rather than fermentation.

Sites that produce any chemical by fermentation and have >50 l capacity                                                                                                                                                                                                                                           est. <75

3. Research Institutions and Accredited Universities and Technical Institutes

(Sources: World of Learning, 1982-3; Lovejoy's College Guide, 1995; American Assn. of Engineering Societies; US Institute of Engineering; Dr. George Georgiou, Prof. of Chemical and Bioengineering, Univ. Texas, Austin)

None of these institutions is likely to have more than 1-2 fermenters, if any; many of the fermenters in the range 50-300 l capacity are old and unused. Today, because of the efficiency of production by cloned organisms, capacities of 20-50 l are preferred. (The efficiency of pathogen production, however, remains unchanged.) Moreover, a large variety of cell products that once had to be isolated by researchers are now commercially available. Thus, if required to be declared, fermenters >50 l intended for production of research materials in institutions of this type would probably be discarded.

There are 29 biological and 38 agricultural engineering schools, which are likely to have fermentation pilot plants, but not all >50 l; those with >50 l capacity,                                                                                                                                                                                                                    <67

There are 145 chemical engineering schools, some of which have fermentation pilot plants; those with >50 l capacity,                                                                                                                                                                                                              est.<30

4. Government Sites

USAMRIID and probably a few others                                                                                                                                                                                                                                          est. 10

_______________________________________

Approximate Maximum No. US Sites with Microbial Production >50 L (based on high estimates for each category)                                                                                                                                                                                                                                       Total <293

PROBABLE TOTAL NUMBER OF US SITES WITH MICROBIAL PRODUCTION CAPACITY >50 L                                                                                                                                                                                                                                       about 250

_______________________________________________________________________________________________

* Active, bonafide human food or drink production could not be interrupted safely for pathogen production on a routine basis. Such facilities could, however, be designed for rapid conversion to BW production if called upon to do so in the future.