【CDCレポート】「ロングフライトにおけるSARS-CoV2感染伝播」(9/18) 航空機内でのマスク装着は必須 (2020.9.19作成)

Joseph Allen @j_g_allen

5/ I made this point in op-ed from May: "Billions of people travel by plane every year, yet there have only been a handful of documented disease outbreaks attributable to airplanes in the past 40 years." washingtonpost.com/opinions/2020/…

Joseph Allen @j_g_allen

6/ Can you get this on an airplane? Yes, sure, you can get this virus anywhere. Time on airplane is just lower risk than people think. People tend to point to the airplane first, ignoring the other travel-associated activity Now, to the report...

Joseph Allen @j_g_allen

7/ They report 15 "flight-associated" cases, and rule out that these travelers had pre- or post-flight exposure. First thing that caught my eye showing that this couldn't be correct: "median symptom onset was 8.8 days (interquartile range 5.8–13.5) after arrival" hmmm...

Joseph Allen @j_g_allen

8/ CDC reports median incubation period of 4-5 days, with 97.5% of symptoms showing within 11.5 days. A sign we should dig deeper in this airplane study...

Joseph Allen @j_g_allen

9/ And when you dig, you uncover. Things like this: "We also note that cases 3 and 14 experienced symptom onset 17 days after flight VN54." 17 days after the flight! Yet, the authors say these are 'flight-associated'. (there's more...) pic.twitter.com/m8Jg8ijeMd

拡大

Joseph Allen @j_g_allen

10/ five of the 15 "airplane-associated' cases had symptoms beyond the time when 97.5% of people have symptoms (11.5 days) Unlikely these five people caught this on flight.

Joseph Allen @j_g_allen

11/ So, the '15 cases of airplane-associated cases' is already down to 10 with just some basic review of symptom data. What else can we learn about those other 10...

Joseph Allen @j_g_allen

12/ The authors took pains to explain away post-flight transmission: "We consider local transmission after arrival in Vietnam unlikely." But when you peek at the appendix, you find more info that suggests post-flight exposure occurred...

Joseph Allen @j_g_allen

11/ in addition to the symptom data supporting that post-flight exposure happened, so does the time-activity info There sure seems like *a lot* of important post-flight exposure activity Like this: 3 of the 'flight-associated' cases went on *CRUISE SHIPS* after flight (blue) pic.twitter.com/AehOIDHKf3

拡大

Joseph Allen @j_g_allen

12/ Really, look at that figure hard Combine it with the earlier statement that median incubation was 8.8 days after flight, as opposed to 4-5 days from CDC Walk back 4-5 days from when their symptoms appeared. --> cruise ships and hotels is what you see

Joseph Allen @j_g_allen

13/ For example: Case 2 Symptom onset March 10 activity 5 days before? cruise ship (same with Case 3)

Joseph Allen @j_g_allen

14/ It is certainly possible some got this on the airplane - the index patient person was actively shedding and coughing + no masks. And strongest evidence here that there might have been transmission on airplane is several cases adjacent to business class.

Joseph Allen @j_g_allen

15/ if you take out those people who had symptoms well beyond incubation period (meaning, they were unlikely to have gotten this from the flight), does Figure 1 that shows the dense clustering (the strongest argument the authors make), look different?

Joseph Allen @j_g_allen

15/ eg, Case #4, 10, and 11 are only ones w/ symptoms within 6 days of flight. Does Figure 1 look different if it only shows these 3 cases as 'airplane-associated' cases?

Joseph Allen @j_g_allen

15/ One way to explore this more is look more closely at the time-activity data and symptom onset in relation to seat location on airplane, but I don't see a way to link the data from the seating chart (fig 1) to time-activity data in appendix)

Joseph Allen @j_g_allen

16/ genomic sequencing would answer this...

Joseph Allen @j_g_allen

17/ Taking a step back You can get this virus anywhere. There been millions of travelers since this flight yet one report (similar to SARS). Could be more, sure, but overall rare. And this report is not as airtight as the headline and abstract (and news reports) make it seem

Joseph Allen @j_g_allen

18/ Big picture: I stand by statement I made in op-ed in May. -time on airplane is lower risk than other parts of travel -masking on airplanes right now is a must -take precautions -only travel if really necessary -STAY HOME WHEN SICK washingtonpost.com/opinions/2020/…

A Marm Kilpatrick @DiseaseEcology

While the overall argument in thread by @j_g_allen has merit (air travel is less dangerous than article suggests), he is incorrect about incubation periods (see Fig from meta-analysis: medrxiv.org/content/10.110…). 6.7% of inc periods are >14d & 2.5% are 18+d. twitter.com/j_g_allen/stat… pic.twitter.com/qG41yRIhT2

Joseph Allen @j_g_allen

1/ Stay in your seat and remain calm. This report is making the rounds. I read it. (including appendix, which I'm guessing few read, b/c if you did....) wwwnc.cdc.gov/eid/article/26…

2020-09-19 20:41:59

拡大

A Marm Kilpatrick @DiseaseEcology

Thus, none of the cases can be ruled out due to long incubation periods. All are quite plausible. Similarly, @j_g_allen argues that being on cruise ship or hotel could have led to infection. Possibly, but need data on infection in those settings to assess. None was given.

A Marm Kilpatrick @DiseaseEcology

Finally, the absence of reported cases on flights is not evidence that transmission is not occurring. Many (most?) case investigators in US don't attempt to identify detailed source of travel-related cases.

A Marm Kilpatrick @DiseaseEcology

Clear @j_g_allen is taking some heat for publishing op-ed suggesting air travel was low risk (washingtonpost.com/opinions/2020/…) due to 3 recent papers show evidence of in-flight transmission: wwwnc.cdc.gov/eid/article/26… wwwnc.cdc.gov/eid/article/26… wwwnc.cdc.gov/eid/article/26…

A Marm Kilpatrick @DiseaseEcology

Strongest actual argument for safety of air travel is last article above: when masks are worn and precautions are taken risk is low. There were 6 asymptomatically infected passengers on that flight & they infected only one person who took her mask off while in bathroom.

A Marm Kilpatrick @DiseaseEcology

(Was low transmission due to lower infectiousness of asymptomatically infected people? That's a topic for another thread, but here's a paper by @nataliexdean et al w/ some suggestive evidence: medrxiv.org/content/10.110…)

A Marm Kilpatrick @DiseaseEcology

What is needed to quantitatively assess real risk of air travel is large scale study w/ 100s of flights (not 1 or 2). Flights to country that has strict 14d quarantine would be ideal so post-flight exposure is low. Pre-flight risk still a possible counfounder, of course, and ...