Even though endemic outbreaks of common childhood diseases, such as measles, have been eliminated in some regions after prolonged mass-vaccination efforts, we are still being constantly reminded that reducing vaccination coverage of children in a community poses the risk of a reimported disease outbreak with potentially dire consequences to infants and immuno-compromised individuals. We are also being persuaded that implementing strict vaccination compliance will prevent an outbreak and protect vaccine-ineligible infants via the herd-immunity effect.
There is no question that a disease outbreak can happen in a non-immune community, if a virus gets there. The real question is, how well can high-vaccination compliance ensure herd immunity and protect a community from an outbreak?
Herd Immunity, a Key Principle
Herd immunity is not an immunologic idea, but rather an epidemiologic construct, which theoretically predicts successful disease control when a certain pre-calculated percentage of people in the population are immune from disease. A scholarly article on herd immunity states:
"Along with the growth of interest in herd immunity, there has been a proliferation of views of what it means or even of whether it exists at all. Several authors have written of data on measles, which "challenge" the principle of herd immunity and others cite widely divergent estimates (from 70 to 95 percent) of the magnitude of the herd immunity threshold required for measles eradication."
Herd immunity has been deemed instrumental in rapid disease eradication. Relying upon the meticulous work of Dr. A. W. Hedrich, who documented annual measles attack rates in relation to the proportion of naturally immune people in the 1900s-1930s, the United States Public Health Service had confidently announced in 1967 its intent to swiftly eradicate measles in the USA over the Winter by vaccinating a sufficient number of still susceptible children. Mass vaccination was implemented, but the expected herd-immunity effect did not materialize and measles epidemics did not stop in 1967.
The concept of herd immunity has been used to justify the idea of vaccinating children against a mild disease, who do not personally benefit from such vaccination, to protect a vulnerable but vaccine-ineligible segment of the population. For example, rubella is not dangerous for children. However, for pregnant women who have not become immune from rubella prior to pregnancy, a rubella infection poses a danger during the first trimester by increasing the risk of fetal developmental abnormalities (congenital rubella). Obviously, vaccination with a live-attenuated viral vaccine, such as the rubella vaccine, is contraindicated during pregnancy.
Perhaps with the good intention to immediately put an end to any risk of congenital rubella in their community, elementary-school children were vaccinated en mass against rubella in 1970 in Casper, Wyoming. Ironically, nine months after this local vaccination campaign, an outbreak of rubella hit Casper. The herd-immunity effect did not materialize and the outbreak involved over one thousand cases and reached several pregnant women. The perplexed authors of the study describing this outbreak wrote:
"The concept that a highly immune group of pre-pubertal children will prevent the spread of rubella in the rest of the community was shown by this epidemic not always to be valid."
The belief in herd immunity has no doubt been influencing vaccine-related legislation in many U.S. states and other countries. This notion is used as a trump card to justify and mandate legal measures aiming to increase vaccination compliance. An implicit assumption is that liberal vaccine exemptions somehow compromise this precious herd immunity, which the public-health authorities strive to establish and maintain via vaccination.
Herd Immunity, a Flawed Concept
Although the evidence for vaccination-based herd immunity is yet to materialize, there is plenty of evidence to the contrary. Just a single publication by Poland & Jacobson (1994) reports on 18 different measles outbreaks throughout North America, occurring in school populations with very-high vaccination coverage for measles (71% to 99.8%). In these outbreaks, vaccinated children constituted 30% to 100% of measles cases. Many more similar outbreaks, occurring after 1994, can be found by searching epidemiologic literature.
Before the 1990s, only a single dose of the measles vaccine was on the childhood schedule in North America. Frequent occurrence of measles outbreaks in highly vaccinated communities have been blamed by the medical establishment on what they thought was a failure-prone, single-shot vaccination strategy. The second MMR (measles-mumps-rubella) shot was introduced in the United States and Canada in the 1990s, followed by the elimination of the endemic measles virus from North America by 2002.
In 2011, an imported measles outbreak – and the largest in the post-elimination era – hit a community in Quebec, Canada with 95-97% measles vaccination compliance in the era of double vaccination against measles. If double vaccination is not enough to patch those alleged vaccine failures and ensure the elusive herd immunity, should we then look forward to triple (or, might as well, quadruple) MMR vaccination strategy to see how that might work out with respect to herd immunity? Or, should we instead re-examine the herd immunity concept itself?
The herd-immunity concept is based on a faulty assumption that vaccination elicits in an individual a state equivalent to bona fide immunity (life-long resistance to viral infection). As with any garbage in-garbage out type of theory, the expectations of the herd-immunity theory are bound to fail in the real world.
Ochsenbein et al. (2000) conducted an experiment in mice, in which they compared the effect of injecting mice with two preparations of the vesicular stomatitis virus (VSV). They immunized mice with either unmodified VSV (live virus) or ultraviolet light-inactivated VSV incapable of replication (dead virus). Then they tested the capacity of the serum from the two groups of immunized animals to neutralize live VSV over the 300 days following immunization.
The injection of the live-virus preparation induced long-lasting virus-neutralization capacity of the serum in mice, which persisted for the whole duration of the study (300 days). In contrast, the injection of the dead-virus preparation induced much lower levels of virus-neutralizing serum titers to start with. Virus-neutralizing serum titers reached a peak at 20 days post-immunization and then started to wane rapidly. They went below the level detectable by the neutralization test by the end of the study period (300 days). The conclusion of this experiment was that a procedure that attenuates or inactivates the virus also diminishes its ability to induce long-lasting virus-neutralizing serum titers upon immunization of animals.
Vaccines against viral childhood diseases are similarly prepared by first isolating the virus from a sick person, then rendering it artificially attenuated or inactivated to make a vaccine. The attenuation or inactivation of a wild virus to become a vaccine-strain virus is done to reduce the likelihood of it inducing the disease symptoms or complications, although this happens anyway in some cases. The process of attenuation, while making a vaccine virus "safer" than the original wild virus, as far as disease symptoms are concerned, also limits the durability of vaccine protection. In fact, all vaccines are by necessity either attenuated or inactivated microorganisms or their isolated pieces mixed with adjuvants; and, therefore, the protective effect of any vaccine is bound to wane sooner or later.
The protective threshold for measles-virus neutralizing serum titers in humans is known. Also known is the duration of time after vaccination with MMR when measles-virus neutralizing serum titers drop below the protective level in a segment of the population. 
The Boston University Measles Study
In 1990, a blood drive was conducted among the students of Boston University a month before the campus was hit with a measles outbreak. Due to these natural circumstances, researchers happened to have access to blood samples of many students who either got measles or were spared from the disease during the outbreak. The levels of measles virus-neutralizing serum titers were appropriately measured by the plaque reduction neutralization (PRN) technique, a month prior to and two months after the exposure. Pre-exposure PRN titers were then correlated with the degree of protection from measles: (1) no detectable infection or disease; (2) serologically confirmed measles infection with a modified clinical course of disease; or (3) full-blown measles. By the way, eight out of nine students who ended up getting full-blown measles, had been vaccinated against measles in their childhood.
The outcome of the Boston University measles outbreak study by Chen et al. (1990) was the following:
(a) In all previously vaccinated students who experienced full-blown measles, pre-exposure PRN titers were below 120;
(b) 70% of students whose pre-exposure PRN titers were between 120 and 1052, ended up having a serologically confirmed measles infection, but since their altered disease symptoms did not conform to the clinical measles case definition, they were categorized as non-cases during the outbreak; and
(c) Students with pre-exposure PRN titers in excess of 1052 were for the most part protected both from the typical clinical disease and measles infection.
During the outbreak, many students with pre-exposure PRN titers between 120 and 1052, who were officially categorized as non-cases, nevertheless had most of the viral-disease symptoms, including cough, photophobia, headache, and fever. These "non-cases" ended up with high post-exposure measles PRN titers, just as the disease cases did, suggesting that they were able to replicate the virus during their illness and possibly transmit it.
Subsequent Measles Vaccine Observations
A study by LeBaron et al. (2007) was conducted to determine the duration of measles virus-neutralization serum titers after the receipt of the second MMR shot. The study enrolled several hundred healthy Caucasian children from rural U.S. areas free of measles outbreaks for the duration of the study. About a quarter of these children generated relatively high titers in response to vaccination, although not nearly as high as the titers after a natural infection would be. The rest responded modestly, and some very poorly. The titers in all children, regardless of being high, moderate, or low, reached a peak in a month after the MMR booster, then came down in six months to the pre-booster levels and continued to decline gradually over the next 5-10 years of observation.
In the above study, only about a top quarter of children (called high responders) were able to maintain PRN titers in excess of 1000 units 10 years following their second MMR shot, received at the age of five. These children are therefore likely to still be protected from the measles infection by the time they are adolescents.
The least-efficient vaccine responders (bottom 5%) had their PRN titers fall below 120 units within 5-10 years after the second MMR shot. This percentage of vaccinated children is expected to have full-blown, clinically identifiable measles upon exposure when they get a bit older. This is the reason why vaccinated (and even twice-vaccinated) people show up as disease cases in numbers equal to or even exceeding the unvaccinated cases in communities with very high (>95%) vaccination coverage. Rapid loss of vaccine protection in low responders is the reason for the paradox of a "vaccine-preventable" disease becoming the disease of the vaccinated in highly vaccinated communities. Such disease cases (and outbreaks driven by them) are not due to random vaccine failures, they are anticipated vaccine failures.