Greetings and welcome to the month of March - I'm looking forward to viewing the Full Worm Moon Monday or Tuesday!
Covid Vaccine Efficacy Against Omicron in 5-11 Year-olds
Covid doesn't seem to be in the news much these days, the public is mostly tired of it. It was worth noting, however, a new report reinforcing the Pfizer vaccine's performance in this age group during the omicron era. Key points (though not new) are that a slightly longer interval of 8 weeks between 1st and 2nd vaccine doses is slightly better, though that benefit disappears after 3 months, and the vaccine provides good protection against severe outcomes for about 4 months, then starts to fade a bit. This and prior studies offer continued reassurance to families that covid vaccination, compared to no vaccination, continues to be beneficial for all age groups; the risk/benefit equation is a no-brainer.
Artificial Intelligence/Machine Learning (AI/ML) for Kawasaki Disease Management
When I was practicing full time, I'm pretty sure not a week went by that I didn't mention to someone at the hospital how much I hated Kawasaki Disease. I managed children with suspected KD for decades, mostly in the dark in terms of accurate diagnosis. The only way I could ever know if a child truly had KD is if they developed coronary artery aneurysms, and that outcome is a) present only in a minority of untreated children (thankfully); and b) really uncommon in those who were treated. We've been through multiple iterations of management guidelines, now very confusing and often requiring interpretation from an "expert." My interest in technology was piqued at seeing the words "artificial intelligence" in the title of this study. KD research in general suffers from the "garbage-in, garbage-out" problem - we don't have a true gold standard for diagnosis. For this study, the gold standard utilized for assessing accuracy of laboratory testing for KD diagnosis was the imperfect but accepted resource from the American Heart Association (AHA Guidelines). The lack of a valid gold standard for diagnosis is an unavoidable drawback in every study of KD, but with that caveat the researchers' modeling eventually came up with 3 biomarkers: C-reactive protein, NT-proB-type natriuretic peptide, and thyroid hormone uptake. Using AI-determined cutoff levels for those 3 tests, they developed a model with both sensitivity and specificity of 86% for diagnostic agreement for their patient cohorts with and without KD. (Note it does appear some of their KD patients would not have fulfilled AHA criteria, but that's another matter.)
Given the fact that KD is relatively uncommon and thus most clinicians initially evaluating children for KD do this infrequently, it would really help to have some non-subjective test result numbers to aid in diagnosis. So, this is a very important avenue of research. While 86% sensitivity and specificity sound like high numbers, they actually aren't that great in terms of narrowing down the diagnosis, particularly given that the gold standard is imperfect. I'll try not to bore you with the details of likelihood ratios, but for these numbers the positive likelihood ratio is 6.1 and negative likelihood ratio is 0.16 (the article itself didn't mention likelihood ratios, these are based on my own calculations.) Translated to the real world, if I thought a child I was seeing had a 50/50 chance of having KD based on my clinical evaluation, a positive result from the pre-test combination would raise that 50% chance to about 80%. Would I change my management based on a 50% chance versus 80% chance? Taking into account risks and benefits of treatment, I think I would treat for KD in both instances. On the other hand, if the result were negative, the 50% chance would drop to about 15%. That might be a level to maybe watch and wait, but again given the lack of a true gold standard for diagnosis I'm sticking with the existing algorithms with all their imperfections.
The authors detail how they hope to improve this model's predictive capabilities, and I look forward to seeing future studies from this group utilizing larger and better defined KD and control groups. For now, I wouldn't use this test combination outside of a research protocol.
Diarrhea in the News
I guess since covid is less newsworthy the press needed another illness for the spotlight. Diarrhea is the new poster child! A recent news story resulted in a call from one of my relatives asking how much to worry about norovirus.
Norovirus, scourge of cruise ships, is much more common in winter months. So, no surprise we're hearing about it the past several weeks. CDC reporting is mostly geared to number of outbreaks, rather than number of illnesses, so it's hard to get a handle on things. However, the outbreak number really isn't that big a deal now.
It is still a good idea to use common sense in being careful about norovirus. It is highly contagious, in part because the number of viral particles needed to cause disease is very small - about 100 or so, compared to around a billion live bacteria to be ingested to cause salmonellosis. This low "illness dose" is partly what leads to the recommendation not to rely on alcohol-based gels to protect you from norovirus and instead use the standard 20-second soap and water wash. It's not that alcohol gels (especially at a low pH) can't kill norovirus, it's more the numbers issue.
Speaking of diarrhea, another problem with low illness dose is shigellosis. Shigella infections have appeared in the news lately mainly for a problem of antibiotic resistance, termed extensively drug-resistant (XDR) strains. Usually shigellosis is a self-limited disease not requiring treatment, but antibiotic treatment can shorten duration of bacterial shedding in stool, limiting duration of contagion, and also offers some protection against severe disease which is important for immunocompromised folks and those with underlying chronic GI illnesses such as Crohn's disease.
The CDC document linked above states that only 5% of Shigella strains screened are XDR, but this does represent an increase over the past several years.
High risk groups for XDR shigellosis were men who have sex with men, people experiencing homelessness, international travelers, and people living with HIV. The total number of XDR strains was 239, and of the 232 episodes where information was complete, only 5% occurred in children. This is very different since shigellosis usually is a disease of young children. XDR strains remain susceptible to carbapenems (which would require IV therapy) and fosfomycin (oral but not approved under 12 years of age, though the drug has been studied down to newborn period). Note that the multiplex PCR packages for stool testing can detect Shigella but give no information on antibiotic susceptibility. If clinical suspicion for shigellosis is high (diarrhea containing blood and mucus, or a febrile seizure associated with a diarrheal illness), and you would consider treatment, order a standard stool culture.
Hippocrates
Yeah, that guy (or maybe a group of people) who came up with the oath. Most historians give him/them credit for first use of the diarrhea term. Of course I had to look that up, and it appears in his Aphorisms, Section VI, items 15, 16, 17, and 32. You might get a chuckle out of some of these. They seem to be in random order, but maybe there was some logic to this grouping that was more apparent in ancient times.