Maximum temperature during normal use - non-skin contact

[jddad19]

I'm developing a single use, sterile ME device that will be used in surgery. There is a portion that will be handled by the surgeon (who will always be wearing a sterile glove when handling it) and a portion that will contact the patient.

I understand that the portion in contact with the patient should meet the temperature limits per Table 24 (Clause 11.1.1) and do not have any problem with that. However, the portion that is handled by the surgeon houses electronics that generate heat, and I'm trying to determine what the appropriate temperature limits for these surfaces should be. Table 23 does not seem appropriate since those limits are "applicable for touching the healthy skin of adults." Since these portion of the device will not touch skin but rather will only touch a gloved hand, how do I identify what the temperature limits should be?

Thanks in advance!

 

[Peter Selvey]

In principle they should be the same limits, and in practice even lower given that sustained temperatures at the levels in Table 23 would be unconformable for the operator.

More likely the true issue is about the measurement method.

A device tested in air and relies only on unforced convection and radiation for cooling, and using a base temperature of 40°C (typical IFU spec for conditions of use) can easily get over 48°C with small amounts of power, since it is only +8K temp rise and the heat has nowhere to go. However, the surgeon's hand is a large heatsink and blood flow can also move small amounts of heat away to the much larger body. Plus the surface of the body such as the hand is typically 35°C which should be used as the base, not room 40°C (ref: IEC 60601-2-37). This means a device that gets to say 50°C (+10K) in open air bench test, failing IEC 60601-1, might only be 37°C (+2K) if tested in simulated actual use conditions.

Simulating a hand is of course difficult, so if the power in the electronics is low (a few watts) it's possible to choose a substitute material that is obviously less conductivity/heat capacity than the human and but still clearly shows the temperatures are safe. Often just using wood can be enough to pull the temperatures down. But if there is say a high power motor or more serious heat, then it might require closer simulation of the human body.

 

[CharlieUK]

Given the nature of the product and where it will be used, can you stipulate a lower maximum ambient temperature?

 

[jddad19]

In principle they should be the same limits, and in practice even lower given that sustained temperatures at the levels in Table 23 would be unconformable for the operator.

More likely the true issue is about the measurement method.

A device tested in air and relies only on unforced convection and radiation for cooling, and using a base temperature of 40°C (typical IFU spec for conditions of use) can easily get over 48°C with small amounts of power, since it is only +8K temp rise and the heat has nowhere to go. However, the surgeon's hand is a large heatsink and blood flow can also move small amounts of heat away to the much larger body. Plus the surface of the body such as the hand is typically 35°C which should be used as the base, not room 40°C (ref: IEC 60601-2-37). This means a device that gets to say 50°C (+10K) in open air bench test, failing IEC 60601-1, might only be 37°C (+2K) if tested in simulated actual use conditions.

Simulating a hand is of course difficult, so if the power in the electronics is low (a few watts) it's possible to choose a substitute material that is obviously less conductivity/heat capacity than the human and but still clearly shows the temperatures are safe. Often just using wood can be enough to pull the temperatures down. But if there is say a high power motor or more serious heat, then it might require closer simulation of the human body.

Thanks for the detailed response and insight, this is very helpful. It certainly makes sense to me, however given that the standard specifies a method for how measurements are taken (11.1.3), would a test lab accept an alternate test method as you suggest? Any recommendations on how to make this argument?

 

[jddad19]

Given the nature of the product and where it will be used, can you stipulate a lower maximum ambient temperature?

I'm sorry I'm not sure I understand the question, do you mean specifying a maximum ambient temperature for our operating conditions that is lower than a typical ambient temperature? And if so how does that relate to the maximum surface temperatures of the device?

 

[CharlieUK]

At what ambient temperature did you determine that the hand contact temperature for the surgeon exceeded those in table 23 ?

 

[jddad19]

At what ambient temperature did you determine that the hand contact temperature for the surgeon exceeded those in table 23 ?

We've just been testing at room temperature, about 23C. We're still developing the design but when tested in air at this ambient temperature we're measuring around 53C max. We've historically assumed an upper operating condition range in the OR to be up to 30C (which I realize sounds a bit high but was derived from an ASHRAE standard).

 

[Peter Selvey]

53°C in a 23°C environment sounds like a lot, 30K rise. Note that with a significant heat source, trying to measure the "temperature" itself is highly error prone, as it requires careful attention to the attachment of sensing element (e.g. thermocouple) and also environment such as air flow from air conditioning, as well as selection of the point to be measured, and also needs careful measurement of the "ambient" reference. Which means that a measurement of 30K by the manufacturer's rough test could be 40K or more in a third party lab that takes care to get the maximum temp possible. But anyway, 30K already "feels" fairly high for an open air test, it indicates the handpiece has a fair bit of power.

As background, cells start to die above 41C (which is 6K margin above the typical body's surface temp). It's a non-linear relation and at 43C the cell death rate is minor and needs about 2-4hrs to have a noticeable effect, like a visible scar that remains for months (I know this from testing on myself). Above this I don't know the real time/temp numbers but time to burn drops of quickly and again is non-linear. For me, the limits in Table 23 are a bit high (48C is well into the burn zone), but they probably take into account that the measurement is made in air, and the actual skin and tissue temperature is likely to be lower due to the thermal loading effect, plus an operator in contact with warm parts will take action before any significant damage. Plus, tests in 11.1.1 are done at worst case conditions (room temp, test duration, supply voltage etc etc) which means real world temperatures are likely to be lower.

In this case, the operator has to touch the part to perform the work, so 48C seems too high. But again there could be other factors at play, for example the actual procedure is only short and the handpiece itself takes a long time to get up to warmer temperatures.

The key point here is that Table 23 is highly simplified and the actual situation is complicated. It's possible that Table 23 is too strict but it's also possible that Table 23 is too relaxed. So for this case, I'd ignore Table 23 in the first instance and just confirm if the temperatures are within an 8K rise (35C + 8K = 43C) taking into account expected real world factors such as thermal loading and usage time. If it's more than 8K, you should be looking at modifying the design as first priority. If it is less than 8K in "real world", but could fail Table 23, then look at how to deal with a test lab.

If it's more than 8K and the design cannot be changed, then depending on the rise, increasing design effort should be done to make sure that the measurements are accurate, the effects are known, regardless of Table 23.

 

[jddad19]

53°C in a 23°C environment sounds like a lot, 30K rise. Note that with a significant heat source, trying to measure the "temperature" itself is highly error prone, as it requires careful attention to the attachment of sensing element (e.g. thermocouple) and also environment such as air flow from air conditioning, as well as selection of the point to be measured, and also needs careful measurement of the "ambient" reference. Which means that a measurement of 30K by the manufacturer's rough test could be 40K or more in a third party lab that takes care to get the maximum temp possible. But anyway, 30K already "feels" fairly high for an open air test, it indicates the handpiece has a fair bit of power.

As background, cells start to die above 41C (which is 6K margin above the typical body's surface temp). It's a non-linear relation and at 43C the cell death rate is minor and needs about 2-4hrs to have a noticeable effect, like a visible scar that remains for months (I know this from testing on myself). Above this I don't know the real time/temp numbers but time to burn drops of quickly and again is non-linear. For me, the limits in Table 23 are a bit high (48C is well into the burn zone), but they probably take into account that the measurement is made in air, and the actual skin and tissue temperature is likely to be lower due to the thermal loading effect, plus an operator in contact with warm parts will take action before any significant damage. Plus, tests in 11.1.1 are done at worst case conditions (room temp, test duration, supply voltage etc etc) which means real world temperatures are likely to be lower.

In this case, the operator has to touch the part to perform the work, so 48C seems too high. But again there could be other factors at play, for example the actual procedure is only short and the handpiece itself takes a long time to get up to warmer temperatures.

The key point here is that Table 23 is highly simplified and the actual situation is complicated. It's possible that Table 23 is too strict but it's also possible that Table 23 is too relaxed. So for this case, I'd ignore Table 23 in the first instance and just confirm if the temperatures are within an 8K rise (35C + 8K = 43C) taking into account expected real world factors such as thermal loading and usage time. If it's more than 8K, you should be looking at modifying the design as first priority. If it is less than 8K in "real world", but could fail Table 23, then look at how to deal with a test lab.

If it's more than 8K and the design cannot be changed, then depending on the rise, increasing design effort should be done to make sure that the measurements are accurate, the effects are known, regardless of Table 23.

Thank you again for your detailed response. I understand and can appreciate your points overall. We are indeed still doing everything we can in the design to reduce the temperature, but I will say the device does not "feel" hot when it is being used for extended periods of time. This is where I'm most intrigued by your comment about testing with a simulated hand.

When you say

If it is less than 8K in "real world", but could fail Table 23, then look at how to deal with a test lab.

do you have any experience or recommendation for how to approach this with a test lab? My experience has been that they test according to the standard, so I'm skeptical I could convince the lab the use an alternate test method. One thing I could think of would be to perform a study to equate the temperature when held by a human (and presumably measuring below 48C) with the temperature when tested in air, and use the air temperature as the safety limit and document this in the risk management file, with the justification that the device is not in contact with bare skin so the limits of Table 23 don't apply.

 

[Peter Selvey]

I think the use of the gloves to avoid Table 23 is itself not relevant, since presumedly the gloves are fairly thin and wouldn't reduce the temperature significantly. But it could be worth a shot (to get around Table 23) as long as if the actual temperatures are anyhow safe.

In my opinion (of course, it might not count!) the use of an object to simulate the thermal loading of hand is not outside of the standard, as it's incorporated in the first line of Clause 11.1.1, compliance is tested "When ME EQUIPMENT is operated in worst-case NORMAL USE ... ". While Clause 11.1.3 a) 3) does refer to putting a hand held device in still air, this clause (title: positioning) should itself not create an abnormal test condition. But, there is the side issue that 48C seems high for a part that must be held by the operator, especially if it is more than 10min, but it gets complicated as the wrong ambient is used. So the 48C might be itself a massaged messed up limit based on open air test.

As a side note: the standard IEC 60601-2-37 for ultrasounds is a good reference, as it has good methods and limits for the ultrasound probes which can get warm and need to contact the patient. I just check and found for skin contact it uses 33C (not 35C) and has a limit of 43C, which means an effective 10K rise for skin contact, when tested with material mimicking tissue for thermal properties. It also has a separate test in still air which allows up to 27K rise (50C in 23C ambient).

Anyway, step one for dealing with a test lab is to have your own data or plausible rational to show that it is safe with good design margins. It could be cooling by the body, it could be short usage time, it could be that the gloves are somehow really reducing the temperature. Then show this to the lab in advance of the test. If necessary document it in risk management. A good lab should be able to appreciate the effort and work with it.