Three Bacillus strains (velezensis, pumilus, licheniformis NRRL B-67647–67649) as a gut flora stabiliser for poultry for fattening (4b1898): counts on TSA to EN 15784, PFGE — native Bacillus in control feed up to 9.8 × 10⁷ CFU/kg against a dose of 3 × 10⁷, some batches below specification
Commission Implementing Regulation (EU) 2025/2500, published on 12 December 2025, authorises until 1 January 2036 a preparation of Bacillus velezensis NRRL B-67647, B. pumilus NRRL B-67648 and B. licheniformis NRRL B-67649 (1:1:1, Microsaf, holder: S.I.Lesaffre) for poultry for fattening and ornamental birds, at min. 3 × 10⁷ CFU/kg feed. EFSA (opinion of 6 May 2025) identified the strains by ANI from genome sequences and found no acquired resistance genes and no toxicity to Vero cells; the additive is not irritant to skin or eyes provided calcium carbonate is the only carrier, but it is a skin and respiratory sensitiser — also because of nickel.
- Methods in the annex: identification — DNA sequencing or PFGE (CEN/TS 17697); enumeration in the additive, premixtures and compound feed — spread plate on tryptone soya agar (EN 15784). The annex requires min. 2 × 10⁸ CFU/g, although there are two forms: diluted GT 3B (2 × 10⁸) and concentrated GT 3B-C (5 × 10⁹ CFU/g). Five batches: GT 3B on average 1.9 × 10⁸ CFU/g (1.4–2.2 × 10⁸), GT 3B-C 1.2 × 10¹⁰ (4.0 × 10⁹–3.4 × 10¹⁰); EFSA notes that three batches of GT 3B and four of GT 3B-C had total counts below the specification, as did the batches used in stability studies.
- Impurities (three batches of each form): lead 4.4–5.2 and 3.7–5.5 mg/kg, nickel 0.2–0.3 mg/kg, cadmium < 0.4, arsenic < 0.2, mercury < 0.02 mg/kg; PCDD/Fs + PCBs up to 0.247 ng WHO-TEQ/kg; mycotoxins below the LOQ; Salmonella absent in 25 g, Enterobacteriaceae, E. coli, C. perfringens, sulphite-reducing anaerobes, yeasts and moulds and B. cereus < 10 CFU/g. Dusting potential by the Stauber-Heubach method 3002–4080 and 4940–6070 mg/m³. No losses (< 0.5 log) after pelleting at 95 °C and 24 months at 4–25 °C; in premixture at 40 °C after 6 months a loss of > 1 log; homogeneity CV 2 %.
- Efficacy: three 42-day trials in chickens at 3 × 10⁷ CFU/kg — higher final body weight (2618 versus 2346 g, 2988 versus 2950 g, 2860 versus 2755 g), and in two a better feed-to-gain ratio (1.91 versus 1.95 and 1.99 versus 2.07). Native Bacillus were measured in the control feed: 4 × 10⁶–9 × 10⁶, < 5 × 10⁶–1.1 × 10⁷ and 2.5 × 10⁶–9.8 × 10⁷ CFU/kg; the presence of the additive strains in control groups was excluded by WGS analysis. Analysed in the supplemented feed: from 1.9 × 10⁷ to 1.6 × 10⁸ CFU/kg.
- Compatibility with coccidiostats: in vitro the MICs of monensin, narasin, salinomycin, maduramicin, lasalocid, robenidine and narasin/nicarbazin were below four times the dose, so an in vivo study was needed (counts in ileum and caecum with and without heat treatment, criterion a difference ≤ 1 log). Lasalocid reduced vegetative cells of B. velezensis in the ileum and halofuginone those of B. pumilus in the caecum by more than 1 log, yet EFSA accepted compatibility with all of them. The act lists diclazuril, narasin, maduramicin ammonium, salinomycin sodium, monensin sodium, robenidine hydrochloride, narasin/nicarbazin, lasalocid and halofuginone; decoquinate, which EFSA found compatible in the text because it has no antibacterial action, appears neither in the opinion’s conclusions nor in the act.