Trend analysis of microbiological contamination within a pharmaceutical compounding unit : towards a risk-based revision of the environmental monitoring sampling plan

Introduction
The entry into force of the French Good Preparation Practices guidelines (Bonnes Pratiques de Préparation, BPP 2023) formalized the requirement to define a documented microbiological sampling plan (SP) for cleanrooms, incorporating monitoring points, frequencies, and action limits. Aseptic preparations are compounded in a class C cleanroom fitted with an airlock and four isolators (class A within a class C environment). Implementation of this SP represents a substantial burden in terms of sampling time and consumables. To assess its relevance and initiate a revision, a trend analysis was conducted over a 12-month period.

Materials and Methods
The SP complies with the sampling frequencies, sample types, and limits recommended by the BPP 2023, as well as the number of monitoring points defined according to standard NF EN ISO 14644. A trend analysis was performed for the year 2025 using data extracted from Lyra® software, covering active air and surface sampling in the class C cleanroom and airlock, as well as active air, settle plates, surfaces, and glove sampling within the isolators. All sampling was performed under operational conditions (‘in operation’).

Results
A total of 2,921 samples were analyzed. In the class C cleanroom and airlock (n=73; active air n=33, surfaces n=40), no sample was out of specification (OOS). Mean active-air contamination was 22 CFU/m³, corresponding to 44% of the BPP action limit (50 CFU/m³). Within the isolators (n=2,848; active air n=48, settle plates n=904, gloves n=1,457, surfaces n=437), 21 samples were OOS (0.74%): settle plates (n=7; 0.77%), gloves (n=13; 0.89%), surfaces (n=1; 0.23%), active air (n=0). The overall compliance rate was 99.3%.

Discussion
These data will form the basis of a risk analysis conducted in accordance with ICH Q9 principles. In the class C cleanroom, the absence of OOS results together with contamination at only 44% of the action limit may support a reduction in sampling frequency: active air from monthly to quarterly, and surfaces from weekly to monthly. Within the isolators, the 21 OOS results show no clinical significance and no unfavorable trend; the predominance of glove-related OOS results (62% of all OOS) points toward reinforcing decontamination procedures rather than intensifying monitoring. Class A monitoring points, which carry the highest criticality, should be maintained; only the monthly active-air sampling, consistently negative, could reasonably be extended to a quarterly frequency.

Conclusion
With a 99.3% compliance rate across 2,921 samples and class C contamination levels well below the applicable thresholds, these findings confirm effective control of environmental bioburden. The ICH Q9-based trend analysis will support a documented and proportionate revision of the sampling plan which, contingent on the absence of drift and subject to annual review, will reduce the monitoring burden without compromising the microbiological safety of the preparations.

See the poster