Biosafety Levels BSL-1 to BSL-4: Containment Requirements Compared

Biosafety levels are frequently presented as four tiers of dangerous organisms. That framing is wrong, and it is the single most consequential error a lab makes when planning containment. The CDC/NIH Biosafety in Microbiological and Biomedical Laboratories (BMBL), 6th edition, is explicit: a biosafety level is a combination of practices, safety equipment and facility design selected for a specific protocol, not a label attached to an agent. The same organism can be handled at two different levels in two different rooms and both assignments can be correct, because the procedure, the volume, the concentration and the strain differ.

This page compares BSL-1, BSL-2, BSL-3 and BSL-4 side by side across the four elements BMBL actually uses to define them, then sets out the decision logic that assigns a level. It deliberately does not restate the build-out detail for any one level. If you already know your level and need facility and equipment specifications, go straight to the BSL-2 facility and equipment requirements guide or the BSL-3 laboratory requirements guide.

The four elements every biosafety level is built from

BMBL Section IV states the structure plainly: “The essential elements of the Biosafety Levels 1–4 are standard microbiological practices, special practices, safety equipment, and laboratory facilities.” Every comparison below is organised on those four axes, because that is how the source document itself is organised — and because a comparison built on any other axis (agent lists, popular examples, room photographs) will not map onto anything a biosafety professional or an inspector asks you for.

• Standard microbiological practices — the baseline behaviours (training, hand hygiene, no mouth pipetting, sharps discipline, decontamination of work surfaces, a biosafety manual). These are present at all four levels and are the only element BSL-1 relies on.
• Special practices — the additional, level-specific procedural controls layered on top. BMBL lists none required for BSL-1; the count and stringency rise sharply at BSL-3 and BSL-4.
• Safety equipment (primary barriers) and PPE — containment placed directly at the hazard. BMBL defines a primary barrier as a “physical containment measure(s) placed directly at the level of the hazard”: biological safety cabinets, sealed rotors, centrifuge safety cups, and the PPE worn when a cabinet cannot be used.
• Laboratory facilities (secondary barriers) — the built environment: doors, sinks, surface finishes, ventilation, effluent treatment, change rooms and showers.

Two structural points follow, and both are routinely lost in simplified explainers. First, the levels are cumulative: BMBL states that “BSL-2 builds upon BSL-1,” and that secondary barriers at BSL-3 and BSL-4 “include those previously mentioned” for the levels below. A higher level does not replace the lower level’s controls; it adds to them. Second, an element can shift roles: BMBL notes that where traditional primary barriers cannot be used — large-animal work is its example — “the facility becomes the primary barrier,” and personnel must then rely on administrative controls and PPE, often with added engineering controls such as HEPA filtration on exhaust air.

BSL-1 to BSL-4 compared

All rows below are drawn from BMBL 6th edition Sections III and IV. Where BMBL uses conditional language (“considered,” “based on a risk assessment”), that conditionality is preserved rather than flattened into a yes/no, because that is precisely where most secondary summaries introduce error.

Dimension	BSL-1	BSL-2	BSL-3	BSL-4
Suitable for	Well-characterised agents not known to consistently cause disease in immunocompetent adult humans; minimal potential hazard	Agents associated with human disease of varying severity, posing moderate hazard; generally safe on an open bench where splash/aerosol potential is low	Indigenous or exotic agents that may cause serious or potentially lethal disease through the inhalation route	Dangerous and exotic agents posing high individual risk of aerosol-transmitted, frequently fatal disease, for which no vaccine or treatment exists
BMBL’s own examples	Bacillus subtilis, Naegleria gruberi, infectious canine hepatitis virus, agents exempt under the NIH Guidelines	Hepatitis B virus, HIV, Salmonella, Toxoplasma; also unknown-status human, animal or plant specimens and primary cell lines	Mycobacterium tuberculosis, St. Louis encephalitis virus, Coxiella burnetii	Marburg virus, Crimean-Congo haemorrhagic fever virus; plus any agent with a close or identical antigenic relationship to a BSL-4 agent, until data support re-designation
Special practices	None required	Access controlled while work is in progress; demonstrated proficiency; medical surveillance and immunisations offered as appropriate; documented waste decontamination method; incident evaluation and records	All BSL-2 practices plus specific entry/exit requirements; near-miss and accident reporting system; double-containment of materials leaving the lab; verified decontamination processes; whole-laboratory decontamination considered after gross contamination or major change of use	All lower-level practices plus a change of clothing on entry, a personal (and, in suit labs, chemical) shower on exit, and a fully controlled materials-transfer regime through the containment barrier
Where open manipulation happens	Open benchtop; special containment equipment not generally required	Open bench permitted; a BSC or other containment device used for procedures with aerosol or splash potential, and for high concentrations or large volumes	Inside a BSC or other primary containment device. BMBL: “No work with open vessels is conducted on the bench”	Inside a Class III BSC (cabinet laboratory model) or at an open bench by personnel in a positive-pressure supplied-air suit (suit laboratory model)
Primary barrier	None generally required	Class I or Class II BSC; sealed rotors and centrifuge safety cups	BSC mandatory for agent manipulation; where a procedure cannot be done in a BSC, a risk-assessed combination of PPE, sealed rotors and safety cups	Gas-tight Class III BSC with double-door pass-through autoclave, dunk tank or fumigation chamber, one supply HEPA and two exhaust HEPA filters in series — or the suit itself
PPE	Lab coat, gown or uniform; eye protection for splash-generating work	Designated lab clothing not taken home; eye/face protection for splash or spray work; respiratory protection considered by risk assessment	Solid-front protective clothing (tie-back or wrap-around gown, scrub suit, coverall), not worn outside the lab; two pairs of gloves where appropriate; respiratory protection and shoe covers considered	Complete clothing change; in cabinet labs, disposable gloves worn under the cabinet gloves; in suit labs, a full-body positive-pressure HEPA-supplied-air suit
Access control	Doors for access control; lab not necessarily separated from general building traffic	Self-closing, lockable doors; access restricted while work is being conducted	Separated from unrestricted traffic; access through two consecutive self-closing doors, optionally with a change room or anteroom between them	Separate building or clearly demarcated, isolated zone; sequential exit through inner (dirty) change area, personal shower, and outer (clean) change room
Ventilation	No specific requirement	No specific requirement; BMBL says new-build planning considers mechanical ventilation giving inward airflow without recirculation outside the lab	Ducted mechanical system required, giving sustained directional inward airflow, designed so airflow does not reverse at the containment barrier under failure; visual airflow monitor at entry; exhaust not recirculated to other building areas	Dedicated, specialised supply and exhaust systems with HEPA filtration and emergency power for exhaust, alarms, lighting, entry/exit controls, BSCs and door gaskets
Waste and effluent decontamination	Covered by standard practices	“A method for decontaminating all laboratory waste is available” — autoclave, chemical disinfection, incineration or another validated method	A decontamination method available in the facility, preferably within the laboratory. Note the wording: BMBL does not mandate an in-room autoclave at BSL-3	Double-door pass-through autoclave at the barrier; liquid effluents from sinks, floor drains and autoclave chambers decontaminated by a proven method, preferably heat, with the process validated physically and biologically at least annually
Surfaces and envelope	Cleanable design; impervious benchtops; non-porous chair coverings; no carpet; exterior windows screened	As BSL-1, plus vacuum lines protected with disinfectant traps and in-line HEPA filters; opening exterior windows not recommended	Seams, floors, walls and ceilings sealed; seamless or integral-cove flooring; spaces around doors and vents capable of being sealed for space decontamination; all windows sealed	As BSL-3, plus a containment envelope engineered for gas or vapour decontamination, with gas-tight dampers and decontamination ports on HEPA housings
Verification cadence	None specified	BSCs certified at least annually	BSCs and HEPA filter housings certified at least annually; the facility’s design, operational parameters and procedures verified before operation and re-tested annually or after significant modification	Class III BSC certified at least annually; effluent decontamination biologically validated at least annually; cabinet gloves replaced annually at recertification

For the cabinet class question specifically — which of Class I, II or III belongs at which level, and why a Class II A2 is not interchangeable with a Class II B2 — see the operational comparison of Class I, II and III biosafety cabinets. For how to work in one correctly once installed, see biosafety cabinet airflow, sash height and technique.

How a biosafety level is actually assigned

BMBL’s 6th edition describes a six-step risk management process, and it is the operative content of the document — the level criteria in Section IV are the menu, not the decision. The steps, in the source’s own order:

1. Identify the hazardous characteristics of the agent and assess inherent risk — risk in the absence of mitigating factors: capability to infect and cause disease in a susceptible host, disease severity, availability of preventive measures and effective treatments, likely laboratory routes of transmission, infectious dose, environmental stability, host range, whether the agent is indigenous or exotic, and its genetic characteristics.
2. Identify laboratory procedure hazards — BMBL names four: agent concentration, suspension volume, equipment and procedures that generate small-particle aerosols and larger droplets, and use of sharps. Animal procedures add bites, scratches, zoonotic exposure and experimentally generated aerosols.
3. Determine the appropriate biosafety level and select any additional precautions indicated by the assessment.
4. Review the assessment and the selected safeguards with a biosafety professional, a subject-matter expert, and the Institutional Biosafety Committee or equivalent, before implementing controls. BMBL calls IBC review of potentially high-risk protocols something that “should become standard practice.”
5. Evaluate staff proficiency and the integrity of safety equipment on an ongoing basis.
6. Revisit regularly, verify the strategies and adjust as procedures, equipment or regulations change.

Steps 1 and 2 are the pair that most explainers collapse into one. They are separate because the agent and the procedure contribute independently to risk: a low-hazard agent handled in a way that generates respirable aerosols at high concentration can require more containment than a higher-hazard agent handled in closed vessels at microlitre scale. BMBL lists pipettes, blenders, centrifuges, sonicators, vortex mixers, cell sorters and MALDI-TOF mass spectrometers as recognised aerosol sources.

Risk group is not biosafety level

This is the distinction that most often produces a wrong assignment, and BMBL states it in unusually direct language: the four Risk Groups “correlate with, but do not equate to, Biosafety Levels,” and “just because a biological agent is listed as a Risk Group 3 agent, it does not mean the activities conducted with that biological agent must occur in a BSL-3 laboratory.”

The two classifications answer different questions:

• A Risk Group (RG1–RG4) is a property of the agent — its intrinsic ability to cause disease in healthy human adults and to spread in the community. The authoritative US list is Appendix B of the NIH Guidelines for Research Involving Recombinant or Synthetic Nucleic Acid Molecules. ABSA International maintains a cross-national compendium of risk group assignments at my.absa.org/Riskgroups.
• A biosafety level (BSL-1–BSL-4) is a property of the protocol — the containment combination selected for a specific set of manipulations in a specific facility by specific staff.

BMBL treats a risk group assignment as an initial estimate that “must be modified appropriately based on the unique risks faced by each laboratory for the specific work being done.” The recommended levels in BMBL’s own agent summary statements (Section VIII) carry the same caveat: they represent suggested practices for work with that agent using standard protocols. Depart from the standard protocol and the recommendation stops applying on its own terms.

Where the level moves without the agent changing

Four situations, all drawn from BMBL, in which the same organism lands at a different level:

• Volume and concentration. BMBL states that handling large volumes or high concentrations “may require additional practices outlined in Sections IV and V.” Section V covers large-scale work; the NIH Guidelines separately treat culture volumes greater than 10 litres as large-scale research with its own containment regime.
• Aerosol-generating procedures. At BSL-2 the BSC requirement is triggered by the procedure, not the agent: pipetting, centrifuging, grinding, blending, shaking, mixing, sonicating, opening containers of infectious material, intranasal inoculation of animals, and harvesting infected tissue from animals or eggs are BMBL’s enumerated triggers.
• Genetic modification. BMBL requires that a modified agent be assessed on the same factors as the wild type, with specific attention to whether the modification could alter pathogenicity or susceptibility to antibiotics or other treatments — and warns that unanticipated enhanced virulence has been reported with engineered agents. It also cautions that “vaccines should not be considered non-pathogenic simply because they are vaccine strains.”
• Unknown specimens. Clinical and diagnostic laboratories rarely know the agent; BMBL notes most public and animal health clinical laboratories operate at BSL-2 facility, engineering and practice standards, and directs readers to its Appendix N. Where human blood or other potentially infectious material is involved, the OSHA Bloodborne Pathogens Standard (29 CFR 1910.1030) applies independently of and in addition to BMBL.

What “BSL-3 enhanced” and “BSL-2+” actually refer to

These terms are used constantly and defined almost nowhere. There is no fifth or intermediate level in BMBL. What exists is BMBL Section IV, BSL-3 facility criterion D.13, which states that enhanced environmental and personal protection may be necessary based on risk assessment and applicable local, state or federal regulation, and enumerates the candidate enhancements: an anteroom for clean storage with dress-in/shower-out capability; gas-tight dampers to allow laboratory isolation; final HEPA filtration of laboratory exhaust air; laboratory effluent decontamination; containment of other piped services; and advanced access control such as biometrics.

So an “enhanced BSL-3” is a BSL-3 laboratory carrying one or more of those six named features — a defined and auditable claim, not a vague upgrade. “BSL-2+” has no equivalent anchor in BMBL at all; it is institutional shorthand, typically for BSL-2 facilities operating some BSL-3 practices, and should always be written out as the specific practices adopted rather than asserted as a level.

What layers on top of the biosafety level

A biosafety level assignment is necessary but rarely sufficient. Several separate regimes attach independently, and none of them substitutes for another:

• The NIH Guidelines and IBC review. Institutions conducting covered recombinant or synthetic nucleic acid research must maintain a standing IBC, and IBC approval precedes the start of covered work. BMBL calls the NIH Guidelines “the key reference in assessing risk and establishing an appropriate Biosafety Level for work involving recombinant DNA molecules.” Where BSL-3 or BSL-4 containment, large-scale work, or gene-drive modified organisms are involved, the institution must also appoint a Biosafety Officer.
• The Federal Select Agent Program. BMBL notes that the biosafety plan required under 9 CFR Part 121 and 42 CFR Part 73 must rest on an assessment addressing the risk of the select agent or toxin given its intended use. Registration, security and personnel-suitability requirements apply regardless of what level the work sits at.
• DURC and pandemic-pathogen oversight. Dual-use review is complementary to, not a substitute for, biosafety review — see DURC, select agents and biosafety oversight.
• Animal work. BMBL defines four parallel Animal Biosafety Levels (ABSL-1 to ABSL-4), plus ABSL-3Ag for USDA APHIS High-Consequence Foreign Animal Diseases and Pests in large or loose-housed animals, where the building itself serves as the primary barrier. Appendix D also covers ABSL-2Ag and ABSL-4Ag. Animal protocols carry IACUC review alongside IBC review.
• Respiratory protection. Where the risk assessment calls for respirators — explicitly “considered” at BSL-2 and BSL-3 — staff must be enrolled in a properly constituted respiratory protection programme, which brings fit testing obligations with it.

National frameworks are not interchangeable

BSL-1 to BSL-4 is the US nomenclature, from BMBL. Other jurisdictions use containment levels that are similar in intent but differently drafted and differently enforced — Canada’s Containment Levels 1–4 under the Canadian Biosafety Standard, and the WHO Laboratory Biosafety Manual, which BMBL itself acknowledges as an alternative risk assessment methodology. A US BSL-3 specification is not a drop-in substitute for a Canadian CL3 specification; see Containment Level 3 vs CL2 under the Canadian Biosafety Standard for where the two diverge. If you are writing a multi-site protocol, state which framework each site is being assessed against.

Source access note

BMBL 6th edition (revised June 2020) remains the current edition; no seventh edition has been published as of this writing. The CDC’s own hosting of the document at cdc.gov/labs/bmbl returns HTTP 403 to automated retrieval, including with a standard browser user agent, so the quotations and criteria above were taken from the full 6th edition PDF as republished by a university biosafety office. Before designing, commissioning or certifying containment space, read the criteria in the CDC-hosted document directly rather than relying on this or any other summary — BMBL is the authority, and Section IV’s full criteria run to far more detail than any comparison table can carry. Note also that BMBL is explicitly advisory: it “is not intended to be a regulatory document,” though it is incorporated by reference into frameworks that are binding, including the Federal Select Agent Program regulations and the NIH Guidelines.

Frequently asked questions

What is the difference between BSL-2 and BSL-3?

Three differences carry most of the weight. First, route: BSL-3 is defined by the potential for serious or lethal disease via inhalation, whereas BSL-2 agents are chiefly a percutaneous, mucosal and ingestion risk. Second, where work happens: at BSL-2 the bench is permitted and a cabinet is required for aerosol- or splash-generating steps; at BSL-3 no work with open vessels happens on the bench at all. Third, engineering: BSL-2 has no ventilation requirement, while BSL-3 requires a ducted system delivering sustained directional inward airflow, two consecutive self-closing doors, sealed surfaces and sealed windows, with the facility verified before operation and re-tested annually.

How are biosafety levels determined?

By documented risk assessment, following BMBL’s six-step process — agent hazards first, procedure hazards second, then the level determination, then review with a biosafety professional and the IBC before controls go in, then ongoing proficiency and equipment checks, then periodic reassessment. The agent’s risk group and BMBL’s agent summary statement are inputs to that assessment, not the answer to it.

Is there a BSL-5?

No. BSL-4 is the highest level defined in BMBL, and no other major national or international biosafety framework defines a fifth tier. See the BSL-5 entry for why the term persists in circulation.

Does a BSL-3 laboratory need its own autoclave in the room?

Not per BMBL’s text. The criterion is that a method for decontaminating all laboratory waste is available in the facility, preferably within the laboratory. “Preferably” is doing real work in that sentence: an in-room autoclave is the preferred configuration, not a stated requirement. At BSL-4, by contrast, a double-door pass-through autoclave at the containment barrier is specified outright. Institutional policy, state rules or a Select Agent registration may impose a stricter requirement than BMBL does.

Can the same organism be handled at two different biosafety levels?

Yes, and routinely is. The level attaches to the protocol. Diagnostic handling of a specimen, propagation of the same organism in culture, and aerosol-challenge work with it can sit at three different levels in the same institution. This is why BMBL’s agent summary statements are framed as recommendations for standard protocols, and why a protocol-specific assessment reviewed by the IBC is the operative control rather than a lookup table.

What happens if the risk is genuinely unclear?

BMBL gives a direction of travel: “where there is insufficient information to make a clear determination of risk, consider the need for additional safeguards until more data are available.” It pairs this with the opposite warning — that safeguards more rigorous than needed impose expense and burden “with little enhancement of laboratory safety.” Over-containment is a real cost, not a free default; the resolution is a documented assessment and a biosafety professional’s judgement, not reflexive escalation.

Related CASRAI resources

• BSL-2 requirements: facility and equipment — the build-out and procurement detail for the level most research labs actually operate at.
• BSL-3 laboratory requirements — facility, equipment, certification and cost detail, including what a BSL-2 to BSL-3 upgrade involves.
• Class I vs II vs III biosafety cabinets — choosing the primary barrier the table above assumes.
• BMBL, biosafety level, BSL-1 and BSL-4 — controlled-vocabulary definitions.
• Laboratory design for compliance and safety — how containment requirements translate into a floor plan.
• Lab compliance — the full cluster hub for accreditation, containment, chemical safety and facility controls.