Previously, I've written about the seriousness of measles, but recently I was in a discussion about what the actual measles death rate is. Is it one death per one thousand cases? Three per one thousand? One per ten thousand? One per blue moon? Depending on where you look, you might find all of the above. What's going on there? Let's take a look.
Reported vs. Estimated
There were about 400-500 deaths reported annually in the US during the decade prior to vaccination. Measles, like chickenpox, was contracted by nearly every child before adulthood, making the annual incidence of the disease similar to the birth rate, around 3-4 million cases per year.
Now, if we use 400 deaths per year and 4 million cases of measles per year just to make the math easy, that's one death per ten thousand cases of measles. Way less than the oft cited one per one thousand deaths, much less three per one thousand. You get the occasional antivaxer who does that shallow math and declares that the CDC is lying when they speak of a death rate in the 1-3 per thousand range.
But, that is comparing reported measles deaths to estimated measles cases, which is very naughty. When we look at reported measles cases, we find that measles cases were vastly underreported, because it was such a common disease. Out of the 3-4 million annual cases of measles, only about 400k to 600k were reported to the CDC. If we go with the 400k for easy math again, that's one reported measles death per one thousand reported measles cases.
So the question is, if measles cases were underreported so much, how much were measles deaths underreported? Measles deaths were commonly from secondary infections like pneumonia and encephalitis, so it would not be surprising if a number of deaths due to measles were not reported as measles proper and not recorded in the death statistics.
I don't know of studies looking at measles death underreporting in the United States before the measles vaccine. What we can look at, though, is measles reporting in the most recent large US outbreak, which occurred in 1989-1992 (and was the instigator for the second MMR vaccine recommendation). There were 57,859 cases of measles during this time, and 127 deaths, working out to almost exactly 2 deaths per 1000 cases of measles.
We would expect that in the 90's, when measles was not so ubiquitous, that reporting of cases and deaths was better than it would have been in the 50's, but there was some underreporting of both. Surveys indicated that a large number of measles cases during the outbreak were not seen by a doctor or were not reported, and the actual number of measles cases might have been 2-3 times more, though the nature of surveys limits the usefulness of this data.
On the other hand, one study in the Journal of Infectious Disease indicates that due to underreporting, the death rate of measles might have been as high as 2.83 deaths per 1000 reported cases of measles, hence the occasional mention of a death rate as much as 3 per 1000.
When I look at the historically reported cases/deaths and the deaths from the 90's outbreak, and factor in underreporting of both cases and deaths, I tend to go with the ballpark of 1 death per 1000 cases of measles. But to say the death rate may be as high as 1 per three hundred cases would be based on the above study.
What About Today?
But, that last outbreak was 20 years ago, and focused in urban areas that may have increased mortality risk factors. Would the US death rate be lower now? We can take a look at the 2011 European outbreak, which had over 26,000 cases and a death rate of around 1 per 3000, but varied highly by country. France bore the brunt with a death rate of around 1 per 2300.
UPDATE: Attentive readers have provided citations for the 2017 measles outbreak in Europe, with around one death per 1200 cases in Italy, and one death per four hundred cases throughout the European Union.
It would not surprise me to find that the United States has a death rate higher than France, given the challenges in our country regarding poverty and health care access. I hope it does not, but I don't want to have an outbreak of thousands of cases in order to find out.