Bioreactor Contamination Control: Role of Process Data

  • 21st October, 2013

One thing that those of us involved in large scale biomanufacturing most dread is a contamination event in one of our cell culture bioreactors. Especially in the mammalian ones that run more than a fortnight to churn out a batch. This not only disrupts planning and scheduling activities having a direct impact on supply chain, a contamination event drains lot of resources (time, raw materials etc.) having a very high quality and economic impact on the operations.

This blog post concerns CIP/SIP stainless steel reactors typically starting from 100L ranging upto 20000L. Contamination troubleshooting and root cause identification in autoclave lab reactors (<50L) or single-use disposable reactors is relatively simpler compared to the stainless counterparts.

Understanding the machine and its sterile boundary:

Operating a bioreactor successfully is analogous to driving a car in a crowded marketplace (especially if you have driven in a developing country) without compromising passengers (or pedestrians) safety and avoiding any dents on the car surface. You have to have a real feel of the machine you are operating in order to manouvre it safely through the threats on its integrity. The concept of sterile boundary and understanding the conditions under which this sterile boundary can be compromised is not easy and is gained only by experience and constantly monitoring the bioreactor data.

As shown in the figure above the sterile boundary is constantly maintained by:

  • sterile filtering the gas flows
  • sterile filtering all the liquid feeds to the reactor
  • steaming all the ports (inlet and outlet) before and after each feed or sampling event
  • maintaining a positive pressure (usually by inert gas like Nitrogen) in the reactor

Modes of breach in sterile boundary

Here below I will mention the modes in which this sterile boundary can be compromised categorized in the order of their probability of occurrence.

Highly Probable
  • Contamination entering from feed/sampling ports: Whenever reactor ports are opened to take in some feed or to take out samples, the sterile boundary is stretched. The reactor sterility can be compromised if
    • There is improper steaming of the ports i.e. port is not sterilized for the required time or required sterilization temperature is not achieved
    • Micro-cracks in the diaphragms' elastomers on one of the port valves

Less Probable
  • Online sterile gas filters integrity breach or filters not assembled properly prior to the run

Rare but probable
  • Most of the modes in this category are related to improper bioreactor maintenance or poor reactor design:
    • Improper seating of O-rings or warped reactor O-rings (head-plate, bottom-plate, man-hole etc)
    • Breach in impeller shaft seals
    • Faulty pressure relief valve
    • Faulty drain system
    • Poor reactor design with improper dead legs
    • CIP/SIP not being able to sterilize certain parts of the bioreactor
    • Faulty steam traps

Looking at process data for root cause identification

Identification of a contamination event in a bioreactor is usually spotted while monitoring the dissolved oxygen profile of the batch in progress. Some information that is needed to solve the riddle of tracking the source of contamination are listed below.

  • Dissolved oxygen profile: The sudden drop in dissolved oxygen levels (shown in chart below) provides the estimate of time when possibly the contamination would have entered in the bioreactor.
  • Valve temperature profile: Suspected valve temperature profiles (see chart below) should be checked if valve was properly sterilized before the feed/sampling event.
  • Bioreactor events during the batch: These event times when correlated with the time of drop in the %DO provides preliminary estimation of source of contamination
  • Contaminant species identification: Investing in rapid species identification will help identify whether the contaminant is gram-positive, gram-negative or spore formers. Doubling time of species identified then helps in approximate estimation of time between species entering the reactor and it overpowering the whole culture (%DO bottoming out).
  • Past bioreactor maintenance history
  • Number of CIP/SIP cycles that elastomeric diaphragms had gone through: This check will help nail the possible diaphragm valve (if above the usage limit) that may have failed/cracked due to overuse.

The figure below shows a sample dashboard providing some of the information listed above

Access to these data sets comes from various sources like
  • QC data
  • Data historians
  • Batch Records
  • Work order records
  • Non-conformance management systems
  • Training records

Instant access to this data helps the investigator to get to the probable root cause in the fastest possible time so that appropriate CAPA's can be implemented before the next bioreactor run. ProcessPad platform from Simplyfeye is one such tool-set that help you provide instant on-demand access to these information for timely action, recovery and prevention from such events.