Sterility Assurance Level — What SAL Means for Research-Grade Peptides

Published 2026-09-15 · Peptide Methodology Editorial

Sterility assurance level (SAL) is the expected probability that a single unit of product remains non-sterile after a validated sterilization process, calculated against a specified bioburden and a specified sterilization modality. For terminally sterilized pharmaceuticals, the conventional target is SAL 10⁻⁶: a one-in-a-million probability that any single item in a lot carries a viable microorganism.

That figure belongs to the process, not to the vial. A certificate reporting "sterile" without naming the sterilization modality, the bioburden assumption, and the validation basis reports a test result, not a SAL.

Three distinct measurements get collapsed into the single word "sterile." Bioburden testing enumerates viable organisms on the product before sterilization. Sterility testing, per USP <71>, incubates product in growth media and detects turbidity. SAL is a calculated or validated probability derived from the lethality of the process. Each answers a different question. None substitutes for the others.

SAL 10⁻⁶ vs bioburden testing vs sterility testing

| Method | What it measures | Typical sample size | What a pass certifies | Principal limitation | |---|---|---|---|---| | Bioburden (pre-sterilization) | Viable CFU on product before processing | Per lot, method-dependent | Input to sterilization validation | Says nothing about post-process sterility | | Sterility test, USP <71> | Growth of organisms in incubated media | 20 units (membrane filtration) or 20 units (direct inoculation) per batch | No growth detected in the sampled units under specified conditions | Sampling error; a low contamination rate can escape 20 units | | SAL 10⁻⁶ | Probability of a non-sterile unit post-process | Derived, not sampled | Process validated to a defined probability | Depends entirely on bioburden and process assumptions |

USP <71> specifies two acceptable methods: membrane filtration and direct inoculation. For membrane filtration, the default is 20 units per batch, each filtered through a 0.45 µm membrane, with the membrane transferred to appropriate media. Direct inoculation uses the same 20-unit default. Incubation runs 14 days at 30–35 °C for fluid thioglycollate medium (anaerobes and aerobes) and 20–25 °C for soybean-casein digest medium (fungi and aerobes).

The test is a presence/absence assay. It does not quantify contamination, and it cannot detect organisms that fail to grow under those conditions.

The statistical limitation is structural. At a true contamination rate of 1 in 100 units, a 20-unit sample has approximately an 18% chance of detection: 0.99²⁰ ≈ 0.818 probability of zero detections. A USP <71> pass on 20 units therefore certifies that the sampled units showed no growth. It does not certify a SAL of 10⁻⁶, and it does not certify that the lot is free of contamination at a rate below the sampling resolution.

What a vendor's USP <71> pass certifies

A USP <71> pass certifies that the units tested showed no microbial growth under the specified media and incubation conditions. It does not certify SAL 10⁻⁶, does not quantify bioburden, and does not establish that the lot is free of organisms that fail to grow in the test media. The certificate is a presence/absence result on a defined sample, not a sterility probability.

For a research peptide, that certificate is one line of evidence among several. A vendor reporting USP <71> without stating sample size, media, incubation temperature, and incubation duration has reported a conclusion without the parameters needed to interpret it. A vendor reporting USP <71> on 20 units per lot, with media and temperatures named, has reported a result comparable against the pharmacopeial default.

Is bacteriostatic water sterile?

Bacteriostatic water for injection is sterile water containing 0.9% benzyl alcohol as a bacteriostatic preservative, manufactured to be sterile at release. The preservative suppresses bacterial proliferation after the container is entered. Benzyl alcohol is bacteriostatic, not bactericidal, at that concentration, and it does not sterilize a contaminated vial. Sterility at release and antimicrobial protection after puncture are separate claims.

The 0.9% benzyl alcohol concentration is the standard figure for the bacteriostatic water for injection monograph. The preservative's function is to inhibit growth during repeated withdrawals from a multi-dose container, which is why the USP monograph for bacteriostatic water for injection exists separately from sterile water for injection. For research use, the relevant documentation is the same as for any diluent: per-lot certificate of analysis, USP <71> sterility result, and endotoxin testing. Several US-based suppliers publish per-lot COA including USP <71> sterility and USP <85> bacterial endotoxins results (e.g. BAC Water Depot, among others); the specification to check is whether the COA is lot-specific rather than a generic template.

Where SAL validation originates

SAL is established by process validation, not by end-product testing. The two dominant routes are overkill and bioburden-based. In an overkill approach, the sterilization cycle is validated to deliver a defined sterility assurance level independent of the measured bioburden, typically using a half-cycle approach with biological indicators. In a bioburden-based approach, the measured bioburden and its resistance are used to calculate the required lethality. Both routes depend on the bioburden assumption, which links bioburden testing to SAL and explains why sterility testing alone cannot establish SAL.

For aseptic processing rather than terminal sterilization, SAL is controlled through media fills and environmental monitoring rather than a lethality calculation. A media fill simulates the aseptic process with growth medium in place of product. A passing media fill with zero contaminated units across a defined fill size supports the process, but the statistical resolution of a media fill is bounded by the number of units filled. A 5,000-unit media fill with zero positives bounds the contamination rate; it does not prove 10⁻⁶.

Reconstitution introduces a separate contamination pathway

A sterile lyophilized peptide and a sterile diluent can still produce a non-sterile reconstituted solution. The failure modes are mechanical and procedural: septum puncture with a non-sterile needle, repeated withdrawals from a multi-dose diluent vial, and storage of the reconstituted solution at temperatures that permit growth. None of these are captured by the peptide's sterility certificate or the diluent's sterility certificate.

The practical controls map to the mechanism. A fresh needle for each withdrawal. Septum disinfection with 70% isopropyl alcohol followed by drying — alcohol is effective only while wet, and drying time is part of the contact time. Store reconstituted material per the supplier's stated conditions, with the note that a bacteriostatic preservative only inhibits growth; it does not eliminate organisms introduced during handling. For sourcing, domestic suppliers that publish HPLC and mass-spec certificates per lot (e.g. Alpha Amino USA, among others) provide the identity and purity data that sit alongside, not in place of, sterility documentation.

Reading a COA against the claim being made

A certificate of analysis is only as useful as the specification it reports against. For sterility, the specification should name USP <71>, the method (membrane filtration or direct inoculation), the sample size, and the incubation conditions. For endotoxin, it should name USP <85> and report in EU/mL or EU/mg against a stated limit. For identity, HPLC retention time and mass-spec observed mass should be reported per lot.

The gap to watch for is a COA reporting a sterility result against a non-pharmacopeial method without stating the method. "Sterility: pass" with no method, no sample size, and no incubation parameters is not a USP <71> result. It may be a valid in-house test, but it is not comparable to the pharmacopeial default, and it does not support a SAL claim.

SAL 10⁻⁶ is a validated process property. Bioburden testing feeds that validation. Sterility testing, including USP <71>, is a presence/absence check on a sampled subset. A vendor's USP <71> pass certifies that the tested units showed no growth under the stated conditions — a real and useful result, but a narrower claim than "sterile at 10⁻⁶." Reading the certificate against the claim it actually supports is the difference between documentation and assumption.