10-12-2020, 12:10 PM
I'm sure everyone has heard that CSIRO (Australia) has a new study out showing much longer persistence than a March US study. The difference is being generally attributed to keeping the samples in the dark.
I will remind you that different studies use different starting- and end-points, which affects the news-reported number of days or hours. I suggest thinking of half-lives, and then figuring out your own criteria for ending. The US study was often quoted by when it last detected the virus (subject to its testing methodology) rather than when it first failed to detect the virus. My impression was that people then took that duration and decided that was when it was safe to handle.
Multiply the half-life by 10 to compute how long it takes to reduce the virus by 1000 (often used to report survivability of virus), or by 20 to reduce by 1,000,000 (French standard for a virus to be considered killed). The D-value of the recent study is for a 90% reduction. The NIH study used their test sensitivity vs original concentration as the measure for reporting duration, so it was inconsistent in how much the virus had been reduced. (roughly 10x for aerosols, 100x for copper, 1000x for others).
The key results from the CSIRO study are found in the parentheses in this table (which links back to the study paper).
I would critique the study at least for failing to consider cooler temperatures. They only reported on 20, 30, and 40 degrees C. As long as they where doing that, I would have suggested 0, 5, 10, and 15 degrees C to cover freezer, refrigerator, and temperate-climate unmanaged temperatures.
Both studies reported half-life of the virus on stainless steel
CSIRO: 20C 50%RH half-life 43 hours
NIH: 21-23C 65%RH ("rotating drum") half-life 5.5 hours
There may have been other differing conditions (potentially a wind across the surface in the rotating drum).
At this time, I'd consider it a hypothesis that the presence or absence of light (particularly UV light) is the cause of the difference. The NIH study did not indicate whether the samples were exposed to any light during the tests. If the difference turns out to be UV light, then a good question is how much UV light is typically present indoors.
I removed some cash from my wallet last night and thought about how rarely that cash would have been exposed to any light. However, it probably is at closer to 30C for 12 hours a day and below 20C for 8 hours or so.
I find the difference in half life for paper notes (cash) and cotton curious and maybe troublesome.
With only 3 temperature measurements, I find their supposition of a Z-value (temperature change to change 90% reduction time by a factor of 10) to be questionable. I expect it to be a curve and that the 3 points only gives you an estimate of the slope on that portion of the curve. It may or may not be accurate at 50C or 10C. (Hey, but they report it to 4 digits of precision, so it must be accurate, right?)
From their graphs, it appears their half-life values at 40C for several of the materials may be computed from a single measurement (at 1 hour) of each replicate, whose value was right at their limit of detection. I wonder about the accuracy of those half-lives.
In multiple studies I've seen on SARS persistence, there has been evidence of a breakdown of the assumption of exponential decay, when the level of virus got low. In some measurements, the low level persisted longer than expected. In this study, "For paper notes, infectious virus was detected for 21 days, although there was less than 1 log of virus recovered for both 14 day and 21 day time points.". Whether this has any practical effect may be questioned.
I suspect New Zealand is reviewing their policies in light of this study.
I will remind you that different studies use different starting- and end-points, which affects the news-reported number of days or hours. I suggest thinking of half-lives, and then figuring out your own criteria for ending. The US study was often quoted by when it last detected the virus (subject to its testing methodology) rather than when it first failed to detect the virus. My impression was that people then took that duration and decided that was when it was safe to handle.
Multiply the half-life by 10 to compute how long it takes to reduce the virus by 1000 (often used to report survivability of virus), or by 20 to reduce by 1,000,000 (French standard for a virus to be considered killed). The D-value of the recent study is for a 90% reduction. The NIH study used their test sensitivity vs original concentration as the measure for reporting duration, so it was inconsistent in how much the virus had been reduced. (roughly 10x for aerosols, 100x for copper, 1000x for others).
The key results from the CSIRO study are found in the parentheses in this table (which links back to the study paper).
I would critique the study at least for failing to consider cooler temperatures. They only reported on 20, 30, and 40 degrees C. As long as they where doing that, I would have suggested 0, 5, 10, and 15 degrees C to cover freezer, refrigerator, and temperate-climate unmanaged temperatures.
Both studies reported half-life of the virus on stainless steel
CSIRO: 20C 50%RH half-life 43 hours
NIH: 21-23C 65%RH ("rotating drum") half-life 5.5 hours
There may have been other differing conditions (potentially a wind across the surface in the rotating drum).
At this time, I'd consider it a hypothesis that the presence or absence of light (particularly UV light) is the cause of the difference. The NIH study did not indicate whether the samples were exposed to any light during the tests. If the difference turns out to be UV light, then a good question is how much UV light is typically present indoors.
I removed some cash from my wallet last night and thought about how rarely that cash would have been exposed to any light. However, it probably is at closer to 30C for 12 hours a day and below 20C for 8 hours or so.
I find the difference in half life for paper notes (cash) and cotton curious and maybe troublesome.
With only 3 temperature measurements, I find their supposition of a Z-value (temperature change to change 90% reduction time by a factor of 10) to be questionable. I expect it to be a curve and that the 3 points only gives you an estimate of the slope on that portion of the curve. It may or may not be accurate at 50C or 10C. (Hey, but they report it to 4 digits of precision, so it must be accurate, right?)
From their graphs, it appears their half-life values at 40C for several of the materials may be computed from a single measurement (at 1 hour) of each replicate, whose value was right at their limit of detection. I wonder about the accuracy of those half-lives.
In multiple studies I've seen on SARS persistence, there has been evidence of a breakdown of the assumption of exponential decay, when the level of virus got low. In some measurements, the low level persisted longer than expected. In this study, "For paper notes, infectious virus was detected for 21 days, although there was less than 1 log of virus recovered for both 14 day and 21 day time points.". Whether this has any practical effect may be questioned.
I suspect New Zealand is reviewing their policies in light of this study.