For several years, the University of Ottawa’s Kumanan Wilson has been looking for ways to overcome parental distrust of vaccinations. In a December 2011 study, he and several other researchers examined reactions to MMR vaccine. The study begins: “The measles, mumps and rubella (MMR) have been used extensively in children and have been demonstrated to be safe and effective in preventing disease.” Was this intended to acknowledge that these once common childhood infections were generally benign and conferred natural immunity against re-infection? Or was the word, “vaccines” carelessly omitted in the authors’ rush to repeat their oath of allegiance to vaccine dogma? If the latter, one wonders if the study might be somewhat biased and selective in its reporting despite the authors’ declared “no competing interests”.
They state: “We sought to examine the population wide side effects of these vaccinations on the combined endpoint of emergency room visits and hospital admissions in selected periods post-vaccination.” The study examined MMR injected at 12 months along with meningococcal C vaccine, and MMR injected at 18 months along with a pentavalent booster of diphtheria, acellular pertussis, tetanus, polio and Hib. The author’s data was derived from Ontario records showing paediatric emergency room visits and hospitalizations from April 1, 2006 to March 31, 2009. As we shall see, their choice of a self-controlled case series (SCCS) design (and a “modified” one at that) has badly compromised their declared objective that, “The overall goal of this study was to determine the risk of serious adverse events in all children vaccinated in Ontario at 12 and 18 months of age with recommended pediatric vaccines.”
An SCCS design can have some useful characteristics, eg with cases serving as their own controls, characteristics such as age and race cannot vary and thus cause confounding. However, even the authors have admitted to possible limitations of their study. An obvious one is confounding due to the co-administration of other vaccines at the time MMR was injected. However, they refer to one of their previous studies which, they say, “demonstrated the safety of the pentavalent vaccine” and claim that, because it’s not a live vaccine, meningococcal C vaccine “should create inflammation in the immediate post-vaccination period as opposed to one week later [as happens with MMR].” They must have forgotten that meningococcal C vaccine contains aluminum, an adjuvant which can cause inflammation for years after injection. Another of the study’s flaws is that doctors visits were not included as endpoints. And apparently the authors didn’t consider the likely possibility that some of the 12 and 18 month endpoints they recorded and graphed reflected an ongoing stream of longer term adverse events resulting from any or all of the numerous prior injections recommended from 2 months to 12 months.
The authors note that, “because it is a live vaccine the MMR vaccine has the potential to cause adverse events one to 2 weeks following vaccination.” But the MMR monograph states: “Excretion of small amounts of the live attenuated rubella virus from the nose or throat has occurred in the majority of susceptible individuals 7 to 28 days after vaccination.” The authors did retrieve and graph data re combined endpoint events for periods of 37 days for each of the 12 month and 18 month sets of vaccines (see graphs on pages 3 and 4 of the study). The 37 days included data from 7 days prior to vaccination to 30 days after. They designated a three-day period between the seventeenth and twentieth days post-vaccination as a transition “washout period” and each day from the twentieth day to the twenty-eighth day post-vaccination as a ‘control interval’. The remaining days between the day of vaccination and the seventeenth day post-vaccination would normally each have been designated a ‘risk interval’. However, – and this is where more serious “selection” comes in – the authors inform that, “we modified the standard self-controlled case series approach by looking for an elevation in risk during each vaccination day up to day 17.” (emphasis added)
Among the risk intervals up to 17 days after the 12 month vaccinations, only nine days were found to have statistically significant elevations in counts of the combined endpoint. For that after the 18 month vaccinations, only three days’ counts were statistically significant. This meant that, especially for the 18 month vaccination period, many days which had near-statistically-significant elevations were omitted from analysis. Essentially, the authors creamed off counts from the few elevated risk intervals, compared them to those of the control intervals and ignored counts of the non-elevated risk intervals and washout period (which likely also included some hospitalizations and ER visits possibly due to vaccinations). Just accounting for elevated risk intervals and control intervals alone, Table 3 on page 7 shows totals of 11,167 ER visits and 686 hospitalizations for the 12 month vaccination period; 4288 ER visits and 269 hospitalizations for the 18 month period. Considering that records of 413,957 children were examined for the study, the grand total of 16,410 events calculated from the Table 3. figures provides a hint of how likely it is that some possible vaccine adverse events were not discovered by this study. Also acknowledged in the table were “five or fewer deaths” during both the elevated risk interval and the control interval. It’s not disclosed if there were any deaths during the statistically insignificant days rejected as risk intervals or during the ‘washout period’. It’s also not disclosed how few “fewer” might be. This unfortunate lack of disclosure is due to regulations disallowing counts lower than 6 for reasons of privacy.
Deaths were not included as endpoints in the SCCS since, “a subject dying effectively truncates their follow-up potentially biasing the results”. (In other words, if the deaths occurred during the elevated risk period or during the control period any day up to the second-to-last day, their inclusion as one of the endpoints could have effectively increased the SCCS relative incidence results since, at some point, there would have been no controls with which to compare them.) Is this valid excuse for omission of death statistics one reason why the use of SCCSs is becoming more common for evaluating drugs? Also, a case was omitted from the analyses if they’d been injected with another vaccine at the same time as or up to 28 days after the 12 and 18 month vaccinations. Yet another problem with the SCCS design is that it did not allow for complete quantification of endpoints. If one case was associated with more than one endpoint, only the endpoint which occurred first was used for the study analysis (eg if a case attended an emergency room and was subsequently admitted to hospital, only the emergency room visit was used). In order to counter the bias which this omission of a second endpoint could have introduced, the authors conducted secondary analyses which found that, in fact for both vaccination periods, ER visits were more prevalent than hospitalizations and the increases in the latter were statistically insignificant.
The graphs both show combined endpoints decreasing as the days of vaccination are approached and increasing immediately after, the lowest counts being on the vaccination days. Presumably this is due to most children being vaccinated when they are relatively healthy. Nevertheless, the 12 month vaccination day combined endpoint count was almost 200; the 18 month one was about 125. These counts suggest that some children may have been vaccinated at times when their health was compromised. If they were, was it because the zeal to vaccinate overtook common sense? And why do the graphs show so many days outside of the elevated risk periods (even some prior to the days of the 12 and 18 month vaccinations) with combined endpoint counts of 300 to 600 or more? Were most of these due to non-vaccine events or, as suggested previously, were they a reflection of ongoing and cumulative poor health resulting from previous vaccinations?
The authors analyzed the types of cases represented by the risk interval endpoints of the 12 month vaccination and compared their prevalence and severity to those of the control intervals. On average, they found no difference in severity. The greatest increase was for febrile seizures, fevers and viral rashes. And, “There was an increase in presentation for multiple conditions during the risk period compared to the control period.” The 18 month vaccination period showed a relatively reduced increase in possible vaccine adverse events compared to the 12 month period. The latter showed an increase in one or more ER visits per child of approximately 598 per 100,000 vaccinations which translates into an increase of approximately one child experiencing at least one ER visit per 168 children vaccinated. The authors suggest the difference between the elevated risks of the two vaccination periods might be due to an immunizing effect of the 12 month MMR vaccine on the children who’d received the second dose at 18 months. They inform, “The top diagnoses for the presentation to the emergency room during the 12 month risk interval would all be consistent with a mild viral illness.” and deduce, “The explanation for this effect is likely the controlled replication of the virus [measles, mumps or rubella vaccine virus] creating a mild form of the illness the vaccine is designed to prevent.”
But if the viral illness was “mild”, why were the patients taken to ERs? The authors suggest, “The increase in ER visits we observed could be a result of insufficient information being provided to parents who may not expect their child to develop a reaction a week after vaccination. In particular, the likelihood of this risk may be underestimated by physicians.” If this is true, perhaps the reason doctors are not providing sufficient risk information is because, just like their patients, they too have been poorly informed and think MMR and vaccines in general are “safe and effective”. With this mindset plus fear mongering propaganda they’ve heard about the risks of infectious disease, is there any wonder why parents would rush their children to hospital at the first signs of one erupting? But older doctors and their patients might be better informed and less fearful. Older doctors who’ve lived during the pre-vaccine era know that once-common infectious diseases were generally mild and inconsequential, similar to the viral illnesses experienced by the vaccinated children in this study.
And how will one ER visit per 168 MMR-vaccinated children be seen by governments the next time provincial medical officers of health ask for another dose of MMR to be added to taxpayer funded vaccination programs? Despite his study’s results, lead author Kumanan Wilson parroted to the Cape Breton Post: “MMR vaccine has been proven to be highly effective. And it’s been proven to be very safe.” Perhaps he thinks that endless repetition and lack of education will help sway funding bodies just as it apparently has some parents and doctors.