When validating the steam sterilization instructions for a reusable medical device, first-time pass rates are important to manufacturers. Failures in validation testing require resolution, adding to the total validation time, which can result in a delay to a regulatory submission, getting a device to market, or both.

Nelson Labs analyst placing a BI in a reusable device. (Credit: Nelson Labs)

In the United States, the validation of the sterility assurance level delivered to a medical device by a sterilization process requires microbiological testing. The lethality the process delivers to the test device is measured by inactivation of resistant spore-forming bacteria. With reusable devices, microbiological methods that use knowledge of device bioburden levels cannot be used for the validation of a sterilization process. Instead, biological indicators (BIs), populated with Geobacillus stearothermophilus, are used as surrogates for device bioburden. The concept relies on the high resistance of G. stearothermophilus to the steam process in combination with validated cleaning processes for reusable medical devices to provide a challenge that will exceed the sterilization resistance presented by microorganisms left on the device after cleaning. If the BIs chosen for the validation testing present a microbiological challenge that is overly robust, the odds for having a validation that passes in the first set of test cycles decreases.

As a testing laboratory, it is important to provide guidance to meet industry requirements and guidance without designing validation studies that over-challenge test devices. In an effort to better understand validation failures, all steam sterilization validation test results over a period of time were trended for first-time pass rates. This article helps to demonstrate that designing validation studies that use higher microbiological challenges than required by industry standards and guidance increases the chances of experiencing a validation failure.


In the reusable medical device industry, the microbiological validation of a sterility assurance level (SAL) for a medical device is commonly performed using the overkill validation method outlined in ANSI/AAMI/ISO 17665-1: 2006/(R)2013, Annex D.1 This methodology is well known in the industry and is defined in many other industry standards and guidance documents.

In practice, this methodology consists of using a BI, placed into the most difficult-to-sterilize locations in the medical device, to act as a surrogate for any bioburden that may be found on the medical device after validated cleaning. The population of the BIs placed into the device is commonly chosen to be greater than or equal to 1.0 × 106 CFU/BI. By demonstrating complete inactivation of the BIs placed into the medical device in a half cycle, a log reduction of ≥12 can be demonstrated for the full processing cycle.

One of the issues with this approach is that the resistance of the BI is often not taken into consideration. In ANSI/AAMI/ISO 17665-1:2006/(R)2013, an overkill process is described as providing a minimum 12 log reduction of microorganisms that have a D value of 1.0 minute at 121 °C. However, commercial BIs for steam sterilization commonly have D values at 121 °C that are greater than or equal to 1.5 minutes. When using a BI with a minimum population of 1.0 × 106 CFU/BI, this higher resistance can result in a testing scenario that exceeds the description of an overkill process for steam described by ANSI/AAMI/ISO TIR 17665-2:2009/(R)2016.2

When determining the appropriate BI challenge for a reusable medical device steam sterilization validation, the Fbio value is helpful to ensure that a BI is selected for validation testing that provides an appropriate, but not overly robust, challenge to the process. The calculation for the BI Fbio value can be found in AAMI TIR 12:2010 and is calculated as outlined in the following equation:3

  • Fbio value = log10(BI population) × D121°C value

For a validation using the partial cycle approach (ANSI/AAMI/ISO 17765-1: 2006/(R)2013, Annex D), a BI providing a Fbio value of 6 or greater is required for successful demonstration of process validation. By inactivating all BIs placed into the product, a log reduction of greater than or equal to 6 can be demonstrated in the half cycle for an organism that has a D value of 1.0 minutes at 121 °C:

  • Fbio value = log10(1.0 × 106) × 1.0 minutes = 6 minutes

As stated previously, commercial BIs used in many reusable medical device steam sterilization validations have D values that are greater than or equal to 1.5 minutes at 121 °C. As demonstrated in the following calculation, a population of 4 logs will allow for successful performance of a partial cycle validation, assuming all BIs placed into the product are inactivated in the half cycle:

  • Fbio value = log10(1.0 × 104) × 1.5 minutes = 6 minutes

Since the D value is higher, testing can show the same amount of lethality as that measured with 6 logs of the less resistant BI with a lower population of the more resistant BI.

A lesser known overkill validation approach is the full cycle approach (ANSI/AAMI/ISO 17665-1:2006/(R)2013 Annex D). When this approach is selected, the BI used for validation testing must provide a much more robust challenge than a BI selected for the partial cycle approach. The following steps are used, as described in ISO 17665-1 Annex D, to determine a BI that provides an appropriate challenge for this validation approach:

  1. Calculate the BI population required to demonstrate an Fbio value of 12 minutes:

    1. Fbio = [log10(pre-exposure population) ‒ log10(ending population)] × D121°C value

  2. Add 0.5 log10 to the population determined for an Fbio value of 12 minutes

    1. BI test population = 10[log10(N0) + 0.5)

With this approach, the ending population of the BI is assumed to be 1 CFU. By demonstrating complete inactivation of the BI in a full cycle, the actual surviving microorganisms will have a population of <1 CFU.

To illustrate how to select an appropriate BI challenge for the full cycle approach, a BI with a D value at 121 °C of 2.0 minutes is assumed:

  1. 12 minutes = [log10(pre-exposure population) – log10(1)] × 2.0 minutes

    1. [log10(pre-exposure population) ‒ 0] × 2.0 minutes = 12 minutes

    2. log10(pre-exposure population) = 12 minutes/2.0 minutes = 6

    3. Pre-exposure population = 1.0 × 106 CFU/BI

  2. BI test population = 10(6 + 0.5) = 10(6.5) = 3.2 × 106 CFU/BI

  3. With a D121°C value of 2.0 minutes, our BI must have a population of 3.2 × 106 CFU/BI or greater

As described in ANSI/AAMI/ISO 17665-1:2006/(R)2013, this approach adds 0.5 log10 to the BI population to account for, “...variations in microbiological manipulations and changes in D value for the test microorganism, which can be caused by contact with the product or a contaminating material.”

Experience with reusable medical device steam sterilization validation testing has been that many medical device manufacturers choose BI challenges that are overly robust. This occurs when a partial cycle validation approach is chosen and BIs with D121°C values of 1.5 minutes or greater are selected with accompanying population of greater than or equal to 1.0 × 106 CFU/BI.

While this level of challenge allows for the device manufacturer to demonstrate lethality levels that are higher than required, it also leads to unnecessary validation failures in common U.S. healthcare cycles for devices that present greater steam penetration resistance that more simple device types. Even when attempting to adjust the challenge presented by the BIs to the sterilization validation cycles, manufacturers are reluctant to use BI populations that are lower than 1.0 × 105 CFU/BI, as this is the minimum population outlined in ISO 11138-3. Assuming a D value of 1.5 minutes and a population of 1.0 × 105 CFU/BI, the Fbio value of the BI is calculated to be 7.5 minutes:

  • Fbio value = log10(1.0 × 105) × 1.5 minutes = 7.5 minutes

To better understand the effect of using an overly resistant BI in validation testing, pass/fail results for validation studies were tracked over a period of time. The results are outlined in the following section.

Validation Testing Pass Rates

All SAL validation testing conducted for individual devices and device sets in an 18-month period were analyzed for pass/fail trends. Over this period, 255 studies were reviewed (110 individual device studies and 145 device set studies). The first-time pass rate for individual devices was 85/110 or 77 percent. The first-time pass rate for device sets was 99/145 or 71 percent. Combined, the original first-time pass rate for all steam SAL validation testing was 72 percent.

The most frequently used failure resolution was to decrease the testing population to a minimum of 1.0 × 105 colony forming units/biological indicator (CFU/BI) from higher test populations (e.g., ≥1.0 × 106 CFU/BI). If all the studies that passed using a population of 105 CFU/BI with no other modifications are totaled into the first-time pass rate, the first-time pass rate increases to 76 percent. Of all the 48 studies that used 105 CFU/BI, the pass rate was 90 percent.

Per ANSI/AAMI/ISO standards, a half cycle needs to show complete inactivation of a population of 1.0 × 106 CFI/BI or its equivalent. The equivalent is defined as demonstrating an Fbio value of 6 minutes. The average Fbio value for 105 carriers used during testing was 11 minutes, almost twice the required value. Reducing the population to 105 alone increases the first-time pass rate to ~88 percent despite being well above the requisite Fbio value.

Steam sterilizers at Nelson Labs. (Credit: Nelson Labs)

A second consideration for SAL optimization would be the full-cycle approach, which through linear extrapolation is specified to demonstrate an Fbio value of at least 12. Full-cycle testing uses inoculated carriers with a minimum Fbio value of >12 minutes while conforming to the guidance outlined in ISO 17665-1, Annex D. Since ISO 17665-1 requires the addition of 0.5 log10 to the population required to provide an Fbio value of 12 minutes, actual test BIs have Fbio values approaching or exceeding 13 minutes. While 105 testing results in an Fbio value that is 5 minutes greater than what is needed and has ~88 percent pass rate, the full-cycle approach uses a BI that has an Fbio value that is only 1–2 minutes greater than what is required. By inference, the full-cycle approach provides a more optimized BI Fbio value considering commonly available commercial BI attributes (i.e., resistance and population).

First-time pass rate tracking for the full-cycle approach is currently ongoing. While first-time pass rates appear to be high with this validation approach, at the current time the sample set is small. First-time pass rates will continue to be tracked with this validation approach to obtain a more statistically significant sample set.


Lowering BI population to a minimum of 1.0 × 105 CFU/BI decreases the chance of over-challenging the process for partial cycle validation testing, while still meeting industry overkill validation requirements.

The full-cycle validation approach allows for the optimal BI challenge to be selected due to the BI attributes that are commercially available to device manufacturers for validation testing and is currently showing promising first-time pass rates.

This article was written by Robert Mueller, Chemistry Project Coordinator; Blake Northrup, Study Director; and Jason Pope, Senior Scientist for Nelson Labs, Salt Lake City, UT. For more information, visit here .