Kevin O'Donnell04.01.09

It's one of cold chain's great mysteries: Why do healthcare companies that collect temperature data from essentially the same distribution environments, use the same modes of transportation, the same routes, and the same service providers, end up with very different ambient temperature profiles from one another? Shouldn't all the profiles look more-or-less the same?

The answer is not as elementary as it might appear; there are multiple explanations for this phenomenon.

First of all, there is no universally accepted method for collecting temperature data, performing statistical analysis, distilling the data down and converting it into a meaningful and representative profile that accurately reflects a particular distribution process. Companies, based on their quality systems and acceptable levels of risk and confidence intervals, each reach the ends by different means - some more sound than others. Regardless of how this is achieved, it must represent an optimal balance of testing validity and actual observed conditions applicable to a specific distribution process and the process must be sound enough and logically presented, in the event of a regulatory audit.

Variability in the logistics chain can also be a significant contributor and could include differences in handling practices, seasonal variation, delays in delivery, modes of transport, and time-of-day pack out procedures. USP General Information Chapter <1079>

All this potential variability only underscores the need for shippers of time- and temperature-sensitive medicinal products to thoroughly understand the nature and extent of the hazards within a given distribution environment and know the limits of environmental exposure necessary in designing, developing, and implementing effective and efficient packaging. Such knowledge will not only maximize package performance and minimize cost, but maintain the quality of the drug product.

Among the critical factors necessary to consider when determining a representative ambient temperature profile are:

During the package design process, it is important to determine the amount of heat exposure the packaged product is likely to be exposed to during the distribution process. Such a calculation is helpful in determining the amount of insulation required, the amount of refrigerant necessary, and the size of the package, all necessary elements for maintaining the product within a specified temperature range. A simple process for determining a relative number representing the total amount of heat is by calculating the area under the heat curve. This can be achieved by multiplying the length times the width of the curve.

Where A = area, L = temperature and W = (time) and where all widths are broken down to an equal interval

In the simple example below of an IATA 7D profile (Fig. 1), the length, L = (temperature) at each interval is multiplied by the width, W = (time) and where all widths are equal.

The resulting area in this example A = 7465. This relative number represents all the heat under the curve and can be used to compare other profiles (whose area is determined by the same method), for determining which are more "severe." This can be critical information to have when determining package design parameters.

For more complex applications such as the saw-tooth profile in Fig. 2, the intervals of time are much more frequent than in Fig. 1. This shorter frequency allows for greater resolution necessary to capture the slope. In Fig. 1, the increments are measured in hours. In Fig. 2, they are measured in seconds and calculated using an Excel spreadsheet.

The resulting number, using the same calculation as in Fig. 1, is A = 5364

Comparing the results from Fig 1 and Fig. 2, there is 40% more heat exposure using Fig. 1 profile than using the profile in Fig. 2. This is visually obvious by overlaying the graphs.

But in more complex profiles, the difference is not so obvious. At times, in fact, it is quite deceiving. Take, for example the two profiles designated by the red and blue lines in fig. 3. Which has more heat?

By analyzing the area under the curves (total heat), the two profiles are essentially equivalent, A = 7392.

Another way to compare the two profiles is to look at the Mean Kinetic Temperature (MKT) of the ambient profiles. Although MKT is generally associated with longer storage durations of pharmaceutical products, it still provides a relative comparison of the temperature of the environment and how it could affect a temperature-sensitive payload. MKT is another way to compare temperature effects of a profile compared to a standard average because it takes into account a weighted average of the effect of temperature spikes on a payload. The profile marked in red has an MKT of 34.1°C compared to the blue profile ISTA 7D summer profile, which has an MKT of 30.1°C. Although the area under the curve is essentially the same, using the MKT analysis method reveals that the red profile is 10% more extreme.

But use of MKT alone as a means of determining severity and total heat does not necessarily tell the whole story of what can - and often does - happen to a product in distribution. Short-term spikes in temperature are a frequent occurrence in distribution and can have a great impact on the quality of the drug product. Most vaccines, for example, cannot be allowed to freeze. Exposure to freezing temperatures, even for a brief period of time, can render many vaccines ineffective. The same is true of heat exposure for many biological products.

Fig. 4 represents a profile to which a package was qualified in an Operational Qualification (blue line) and actual internal package temperatures in a distribution environment (dashed line) over a 72-hour period. Although the test profile is representative of what the package actually experiences in distribution, and the MKT for the product was well within parameters, neither the MKT for the product or the test profile captures the fact that the product might have frozen between hours 24 and 28.

While the use of MKT can be applied as a useful tool, these illustrations point out the importance of understanding the temperature hazards within a specific distribution environment based on routes, duration, and modes of transport, and why two profiles utilizing the same transportation services may not necessarily look the same. The more data collected and the more closely distribution processes are monitored, the higher the resolution and greater the accuracy. Equally important is the method of analysis for determining the accuracy and severity of a temperature profile and the hidden potential dangers of relying solely on one method.

The answer is not as elementary as it might appear; there are multiple explanations for this phenomenon.

First of all, there is no universally accepted method for collecting temperature data, performing statistical analysis, distilling the data down and converting it into a meaningful and representative profile that accurately reflects a particular distribution process. Companies, based on their quality systems and acceptable levels of risk and confidence intervals, each reach the ends by different means - some more sound than others. Regardless of how this is achieved, it must represent an optimal balance of testing validity and actual observed conditions applicable to a specific distribution process and the process must be sound enough and logically presented, in the event of a regulatory audit.

Variability in the logistics chain can also be a significant contributor and could include differences in handling practices, seasonal variation, delays in delivery, modes of transport, and time-of-day pack out procedures. USP General Information Chapter <1079>

*Good Storage and Shipping Practices*recognizes this criticality and states, "unlike shock, vibration, and other physical hazards, thermal hazards tend to be unique to a given system."All this potential variability only underscores the need for shippers of time- and temperature-sensitive medicinal products to thoroughly understand the nature and extent of the hazards within a given distribution environment and know the limits of environmental exposure necessary in designing, developing, and implementing effective and efficient packaging. Such knowledge will not only maximize package performance and minimize cost, but maintain the quality of the drug product.

Among the critical factors necessary to consider when determining a representative ambient temperature profile are:

- the cumulative amount of heat exposure during the distribution process (the area under the curve), and
- the assignment of temperature spikes at the appropriate place along the timeline, (the elapsed time during the distribution process capturing day/night exposure and critical touch points).

#### Simple Calculation for Determining the Heat Under the Curve

During the package design process, it is important to determine the amount of heat exposure the packaged product is likely to be exposed to during the distribution process. Such a calculation is helpful in determining the amount of insulation required, the amount of refrigerant necessary, and the size of the package, all necessary elements for maintaining the product within a specified temperature range. A simple process for determining a relative number representing the total amount of heat is by calculating the area under the heat curve. This can be achieved by multiplying the length times the width of the curve.

A = L x W

Where A = area, L = temperature and W = (time) and where all widths are broken down to an equal interval

In the simple example below of an IATA 7D profile (Fig. 1), the length, L = (temperature) at each interval is multiplied by the width, W = (time) and where all widths are equal.

A = L1 x W + L2 x W + L3 x W + L4 x W

The resulting area in this example A = 7465. This relative number represents all the heat under the curve and can be used to compare other profiles (whose area is determined by the same method), for determining which are more "severe." This can be critical information to have when determining package design parameters.

Fig. 1 |

For more complex applications such as the saw-tooth profile in Fig. 2, the intervals of time are much more frequent than in Fig. 1. This shorter frequency allows for greater resolution necessary to capture the slope. In Fig. 1, the increments are measured in hours. In Fig. 2, they are measured in seconds and calculated using an Excel spreadsheet.

The resulting number, using the same calculation as in Fig. 1, is A = 5364

Comparing the results from Fig 1 and Fig. 2, there is 40% more heat exposure using Fig. 1 profile than using the profile in Fig. 2. This is visually obvious by overlaying the graphs.

But in more complex profiles, the difference is not so obvious. At times, in fact, it is quite deceiving. Take, for example the two profiles designated by the red and blue lines in fig. 3. Which has more heat?

Fig. 2 |

#### More Than One Way To Average

Another way to compare the two profiles is to look at the Mean Kinetic Temperature (MKT) of the ambient profiles. Although MKT is generally associated with longer storage durations of pharmaceutical products, it still provides a relative comparison of the temperature of the environment and how it could affect a temperature-sensitive payload. MKT is another way to compare temperature effects of a profile compared to a standard average because it takes into account a weighted average of the effect of temperature spikes on a payload. The profile marked in red has an MKT of 34.1°C compared to the blue profile ISTA 7D summer profile, which has an MKT of 30.1°C. Although the area under the curve is essentially the same, using the MKT analysis method reveals that the red profile is 10% more extreme.

Fig. 3 |

But use of MKT alone as a means of determining severity and total heat does not necessarily tell the whole story of what can - and often does - happen to a product in distribution. Short-term spikes in temperature are a frequent occurrence in distribution and can have a great impact on the quality of the drug product. Most vaccines, for example, cannot be allowed to freeze. Exposure to freezing temperatures, even for a brief period of time, can render many vaccines ineffective. The same is true of heat exposure for many biological products.

Fig. 4 represents a profile to which a package was qualified in an Operational Qualification (blue line) and actual internal package temperatures in a distribution environment (dashed line) over a 72-hour period. Although the test profile is representative of what the package actually experiences in distribution, and the MKT for the product was well within parameters, neither the MKT for the product or the test profile captures the fact that the product might have frozen between hours 24 and 28.

Fig. 4 |

**Profile MKT:**12.7°C**Product MKT:**0.8°C**Low Temperature Excursion:**-4.7°C (@ hour 28)While the use of MKT can be applied as a useful tool, these illustrations point out the importance of understanding the temperature hazards within a specific distribution environment based on routes, duration, and modes of transport, and why two profiles utilizing the same transportation services may not necessarily look the same. The more data collected and the more closely distribution processes are monitored, the higher the resolution and greater the accuracy. Equally important is the method of analysis for determining the accuracy and severity of a temperature profile and the hidden potential dangers of relying solely on one method.