Cowardly Lion: Courage. What makes a King out of a slave? Courage.
What makes the flag on the mast to wave? Courage.
What makes the elephant charge his tusk in the misty mist or the dusky dusk?
What makes the muskrat guard his musk? Courage.
What makes the Sphinx the 7th Wonder?
What makes the dawn come up like THUNDER?! Courage.
What puts the “ape” in ape-ricot?
Whatta they got that I ain’t got?
Dorothy & Friends: Courage!
Cowardly Lion: You can say that again.

The Wizard of Oz, Movie Version, slightly edited for modern sensibilities.

The medical literature can, I suppose, can be divided into three categories. There are the slam dunk studies, where the outcomes are clear-cut. Penicillin treatment of syphilis and HAART for the treatment of HIV come to mind for positive studies. Anti-endotoxin monoclonal antibodies for negative studies pop to mind. Being a retired ID doctor, I tend towards infections as examples.

There are the stupid. Those studies that are unhinged from reality, like acupuncture or homeopathy, where you know any positive result is going all bias, all the time. But occasionally stupid studies get published in the literature. My favorite is Computed Tomographic Study of the Common Cold where 31 people got their heads irradiated to show that people with the cold had stuffy noses. And it made the New England Journal of Medicine!

But much of the medical literature is so often not as clear-cut as one would like. It is messy and often manifests Newton’s third law of the medical literature: for every study there is an equal an opposite study. Well, opposite in results, rarely equal in methodology.

The first law? Interpretations of the literature tend not to change with disconfirmatory studies aka the Law of Inertia.

The second law? Why F = ma, of course. Fatuous = (medical literature)(acerebral). I did learn something from that physics degree all those years ago.

Large swaths of the medical literature are filled with ambiguity and in reading and understanding it, requires lots of qualifiers. As an example, the best treatment for methicillin resistant S. aureus? I don’t know. As the Magic 8 Ball notes, Reply hazy, try again. Which antibiotic to use depends on many factors, but my bias is the answer is never vancomycin. I have long said that if your physician refers to vancomycin, or any antibiotic, as strong, powerful, or big gun, she is an idiot who knows nothing about the treatment of infectious diseases. I digress. ID stands for so many things with me.

The flu vaccine is one of those interventions whose effects are messy. Its effects depend on the vaccine, the host, the circulating strains of flu, prior vaccinations and disease. Synthesizing the literature either makes my brain hurt or identifies me as a first-rate intelligence. Probably the former.

An ideal vaccine would prevent all illness. Influenza vaccine is not ideal, so what would you like from the flu vaccine? Prevent illness, lessen the severity of illness, decrease spread of disease, decrease complications (stroke, heart attack, bacterial pneumonia, spontaneous abortion etc), decrease hospitalization and if hospitalized decrease transfer to the ICU and death. And, perhaps, lead to smarter, richer children.

Plow through all of that literature (as I write there are 41,622 results for influenza vaccine on the PubMeds. Reading one an hour would take 1734.25 days), which no one has, and you get the sense that, however imperfectly, the flu vaccine does all of the above. Just not every time in every person. I like to compare the flu vaccine to seat belts. Not perfect, but better to have them when needed than to go without.

Which brings us to Influenza Vaccine Fails to Stop Hospitalization and Death. Large Japanese Study Shows No Benefit on Hard Outcomes

You will probably be surprised to know that I am a fan of sarcasm and snark, although it is difficult sometimes to make either recognizable in writing. For example, my website is When I was doing my ID podcasts I would ask my kids to clarify some point regarding popular culture (from which, at 66, I am far removed) and my son would sarcastically say, “Being edgy again, Dad”? I would not consider myself to be edgy, but I thought it would be funny as a website name. I really should have gone with Slacker ID. Slacker Astronomy being my favorite podcast when I started all those years ago.

Anyway, the above headline is from Courageous Discourse, and since there is nothing whatsoever resembling courage in what follows, I assume the title is tongue in cheek. Like Standing Up For What’s Right.

The author quotes from the results:

The multivariable analysis showed a lower incidence of influenza in vaccinated individuals (hazard ratio [HR], 0.47; 95% confidence interval [CI], 0.43-0.51; P < 0.001), however the incidence of hospitalization for influenza did not differ significantly by vaccination status (HR, 0.79; 95% CI, 0.53-1.18; P = 0.249). Protective effectiveness against incidence waned quickly after 4 or 5 months.

and then concludes:

These data suggest the massive effort on vaccination in the general population is a waste of time and effort. If the frail and elderly get no overall direct reduction in hospitalization and death, influenza vaccination should be individualized based on pulmonary and systemic risks.

Oddly, he fails to courageously quote from the conclusion, which says:

Our study identified moderate vaccine effectiveness in preventing the incidence of influenza in the Japanese elderly. Vaccine effectiveness showed a trend of gradual attenuation. Clinicians should suspect influenza infection even in those vaccinated, especially in elderly individuals who had received vaccination more than 4 or 5 months previously.

The paper notes:

Multivariable analysis based on the time-dependent Cox regression model showed that vaccinated individuals had a lower incidence of influenza (HR, 0.47; 95% CI, 0.43–0.51; P < 0.001), while hospitalization for influenza did not differ significantly between the vaccinated and non-vaccinated groups (HR, 0.79; 95% CI, 0.53–1.18;P = 0.249).

Then those pesky qualifiers. The paper notes the vaccinated group were older and sicker, so the two groups were not really comparable. And we do not know what vaccination was used,  regular or high dose. In Japan, as best I can tell they have the high-dose vaccine for the elderly; I can’t tell if its use is standard.

Then the primary outcome was getting the flu.

Hospitalization and death were secondary end points and so one never really knows what to make of results one way or the other. There is a long history of secondary endpoints being unreliable. The authors note:

Second, confounding factors adjusted in the multivariable analyses were limited, especially for the secondary outcome of hospitalization. Those vaccinated may have been biased toward individuals who were more likely to be hospitalized according to the unadjusted vaccine effectiveness of hospitalization (HR, 1.21). On the other hand, the point estimate of the adjusted hazard ratio for hospitalization was less than 1, which suggested the effectiveness. However, the confidence interval was wide (0.53–1.18), and the results were not statistically significant due to lack of power. This is because the number of events for hospitalization was less than one-tenth of the number of events for incidence of influenza, regardless of age group. The results should be revalidated in the future with a larger number of event.

As the original authors suggest, it would be suspect to use the results of the paper to make decisions about vaccine effectiveness for hospitalization. I wonder why that was not courageously mentioned.

As to death, the older sicker vaccinated group had more deaths, the study makes no mention on the relationship between death, influenza, or the vaccine and no suggestion as to causality. The deaths are just listed on a table, not remarked on one way or the other.

I suspect the courageous author read the paper but in a cursory way, much in the way the proverbial drunk uses a streetlight.

Overall, this is par for the course for influenza vaccine papers that look at large populations and try and draw conclusions: modest efficacy that wanes with time.

But how does it fit in with the literature as a hole, er, I mean whole?

If you PubMed influenza vaccine and hospitalization, you get 9761 hits. Way too many to go through.

There was a meta-analysis of the high-dose vaccine in the elderly. The results were good as in fewer hospitalizations and deaths:

15 publications were meta-analyzed after screening 1,293 studies, providing data on 10 consecutive influenza seasons and over 22 million individuals receiving HD-IIV3 in randomized and observational settings. Across all influenza seasons, HD-IIV3 demonstrated improved protection against ILI compared to SD-IIV (rVE = 15.9%, 95% CI: 4.1-26.3%). HD-IIV3 was also more effective at preventing hospital admissions from all-causes (rVE = 8.4%, 95% CI: 5.7-11.0%), as well as influenza (rVE = 11.7%, 95% CI: 7.0-16.1%), pneumonia (rVE = 27.3%, 95% CI: 15.3-37.6%), combined pneumonia/influenza (rVE = 13.4%, 95% CI: 7.3-19.2%) and cardiorespiratory events (rVE = 17.9%, 95% CI: 15.0-20.8%). Reductions in mortality due to pneumonia/influenza (rVE = 39.9%, 95% CI: 18.6-55.6%) and cardiorespiratory causes (rVE = 27.7%, 95% CI: 13.2-32.0%) were also observed.

In diabetics?

influenza vaccination was associated with a lower mortality rate (Mantel Haenszel Odds Ratio (MH-OR), 95% CI: 0.54 (0.40; 0.74), p < 0.001). Patients who received influenza vaccination showed a lower risk of hospitalization for pneumonia (MH-OR, 95% CI: 0.89; (0.80; 0.98), p = 0.18). A sensitivity analysis using fixed effect model confirmed the results (MH-OR, 95% CI: 0.91; (0.87; 0.96); p = 0.001).

Community living elderly?

A meta-analysis of studies selected using predetermined criteria without language restriction. Conclusion: Influenza vaccine was effective in reducing influenza-like illness by 35% (95% confidence interval (CI) 19–47%), hospitalization for pneumonia and influenza by 33% (CI 27–38%), mortality following hospitalization for pneumonia and influenza by 47% (CI 25–62%); and mortality from all causes by 50% (CI 45–56%).

And the trivalent inactivated vaccine adjuvanted with MF59?

Our results suggest that MF59-TIV is effective in reducing several influenza-related outcomes among the elderly, especially hospitalizations due to influenza-related complications.

Again, in a recent meta-analysis, the flu does more than prevent flu hospitalization and death

receipt of influenza vaccination was associated with a 34% lower risk of major adverse cardiovascular events, and individuals with recent ACS had a 45% lower risk

So there is a large literature that shows efficacy of flu vaccination in preventing hospitalization and death as well as a variety of other complications.

My first thought would not be to courageously question the need for the vaccine, but wonder why this study was an outlier, besides low power. Is there something about the Japanese? There can be racial/regional differences in susceptibility to infections; I cannot find anything specifically related to the Japanese.

Prior studies have found benefit if vaccinations in elder Japanese:

The adjusted VE, determined using a case-control study, for preventing hospitalization for influenza A infection was 72.6% (95%!CI: 30.7–89.1%). In addition, the VE for preventing hospitalization of influenza patients with comorbidities was 78.2% (95%! CI: 41.1–92%).

Although there is not a robust number of studies, what is out there suggests the elderly Japanese may respond particularly poorly to influenza vaccine, with emphasis on may.  And I cannot find a robust literature specifically looking at the effects of the influenza vaccine specifically on the hospitalization and death in elderly Japanese.

I would conclude that when taken as a whole, the influenza vaccine does indeed decrease the odds of hospitalization and death in not just the elderly but all groups, as well has having multiple beneficial downstream effects for both individuals and populations.

My response from a single outlier would not be to courageously conclude that:

These data suggest the massive effort on vaccination in the general population is a waste of time and effort. If the frail and elderly get no overall direct reduction in hospitalization and death, influenza vaccination should be individualized based on pulmonary and systemic risks.

But instead wonder, huh? Why the outlier? Why was this study different? The methodology? The vaccine? The population?

Courage, well, more likely just good sense,  is realizing that a single study in a messy literature is hardly the basis for firm conclusions. Or any conclusion. Perhaps the solution lies in the needs of the Scarecrow rather than the Cowardly Lion.

Here are the links if you want to immerse yourself in my cowardly discourse on influenza.



  • Mark Crislip, MD has been a practicing Infectious Disease specialist in Portland, Oregon, from 1990 to 2023. He has been voted a US News and World Report best US doctor, best ID doctor in Portland Magazine multiple times, has multiple teaching awards and, most importantly,  the ‘Attending Most Likely To Tell It Like It Is’ by the medical residents at his hospital. His multi-media empire can be found at

Posted by Mark Crislip

Mark Crislip, MD has been a practicing Infectious Disease specialist in Portland, Oregon, from 1990 to 2023. He has been voted a US News and World Report best US doctor, best ID doctor in Portland Magazine multiple times, has multiple teaching awards and, most importantly,  the ‘Attending Most Likely To Tell It Like It Is’ by the medical residents at his hospital. His multi-media empire can be found at