By Helen Petousis Harris 17/10/2021

Recent misleading media headlines about vaccines being administered incorrectly in the absence of evidence do little to help public confidence in vaccines. Spoiler alert, vaccines are not being administered incorrectly. The topic of this blog is based on what could be an important scientific question – is one of the most widely used injection techniques a risk for serious vaccine reactions?

Most vaccines are given by injection into muscle and techniques used to achieve this vary with a range of preferences. Over the past few months there has been a growing conversation about the way vaccines are injected and the potential for serious events, in this case, myocarditis.

Myocarditis is inflammation of the heart muscle, and it can affect the way the heart performs. The most common cause is a viral infection, sometimes it might be a reaction to a drug or part of another condition. Myocarditis can be mild or severe. Fortunately, most cases resolve without permanent complications. Diseases that can cause myocarditis include influenza, rubella and…. COVID-19, so just to be very clear, vaccination is an important tool against myocarditis.

Do the mRNA COVID-19 vaccines cause myocarditis?

To date around 2 billion doses of mRNA vaccines have been administered worldwide and under close observation in many nations. In July 2021 the WHO Global Advisory Committee on Vaccine Safety (GACVS) confirmed a plausible relationship between these mRNA vaccines and myo or pericarditis. Data to Oct 2021 indicated that most of these cases are mild, only require conservative treatment of anti-inflammatories and possibly colchicine, and people recover.

The increased risk of myocarditis following the receipt of mRNA vaccines is VERY SMALL. However, it is an important consideration, and we need to minimise this risk where possible. The risk varies a lot depending on age, gender, and dose, with the most likely to be affected are young males. See here for some analysis on risk.

There are a range (nine so far) of mechanisms proposed to explain how these vaccines might cause myocarditis and 8 of these have been summarised in a review. The 9th is related to inadvertent intravenous vaccine injection. The latter idea does not fully explain this phenomena, for example it does not provide for the fact we see most cases in young males after the second dose.

From here on I will focus on this intramuscular injection technique hypothesis.

Dissecting the question

Let’s dissect the big question into bite-sized pieces by asking a series of smaller questions.

  • Have any other vaccines caused myocarditis?
  • Can the vaccine be administered intravenously by mistake?
  • Is there evidence for biological plausibility?

Have other vaccines cause myocarditis?

As myocarditis is often associated with a viral infection there is a biologically plausible reason that some vaccines might be able to do this too, in particular, live vaccines. This has been observed in the past with smallpox vaccine, which is a live vaccine given by a process called variolation into the skin (not injected into muscle). The study was among military personnel in 2002/3. Of over 230,000 vaccinees there were 18 cases of probable myocarditis that occurred with 30 days of vaccination. This is an incidence of 7.8 per 100,000 which is 3-4 times what one might normally expect by chance in the military population, but not higher than the 1-10 per 100,000 in the rest of the population.

A 2018 study looking at adults who received at least one live virus vaccine over an 11 year period were reviewed for the onset of myopericarditis within 42 days post vaccine. Out of 416,000 people there were no cases of myocarditis and one probable case of pericarditis. So, no increased risk from this study.

A review of reports to the US spontaneous reporting system over a 38-year period 1990-2018 found 708 reports and the estimated reporting rate was less that 0.1 cases per one million doses of vaccines. No signal here.

The key point is that only the live smallpox vaccine has been associated with an increased risk for myocarditis and it is not administered by intramuscular injection.

Can the vaccine be administered intravenously by mistake?

Firstly, a bit about how vaccines are injected.

Intramuscular injection technique

Most vaccines are given into muscle tissue for a variety of reasons. One is that it generally results in fewer injection site reactions than if the vaccine is given into the subcutaneous layer. Another reason is that the blood vessels are small and diffuse. In my 23 years of interest in vaccines I have heard every argument for every approach and even carried out a randomised trial to put three of them to the test back in 2010.

Just to be clear that as a scientist and not a clinical person I have never given an injection to a human. That said, I have given 84 subcutaneous injections to my kitten when he was very ill with a kitty coronavirus-related illness (FIP). He made a full recovery BTW. That said, I did explore this topic including a clinical trail as a component of my doctoral thesis so am not entirely without insight.

Part one of the injection process is to plunge the needle into the muscle using a 90-degree angle (basically straight in).

The second act may occur using three general approaches.

  1. Push the dose in rapidly and withdraw.
  2. Push the dose in slowly and withdraw
  3. Pull back on the plunger, check to see if any blood flashes back into the syringe, then push the dose in. This is called aspiration.

The rational for aspiration was that there might be a risk of injecting into a vein with the meds ending up in the wrong place. Pulling back to check there is no blood was in order to avoid this happening. Because flashbacks of blood hardly ever happen and the process prolonged the injection, it was abandoned by many countries’ National Immunisation Technical Advisory Groups as the recommended technique as well as the WHO in favour of the more pragmatic plunge and push approach.

There is no evidence to indicate that this shift in common practice has had any impact on vaccine safety.

Hitting a blood vessel

The blood vessels in the muscle tissue used for vaccines are generally pretty small, and there is a difference between nicking a vessel and delivering the entire 0.3ml of vaccine into it. To get the load away you would have to get the tip of the needle precisely into the vessel and have it remain there for the entire injection. The area of the deltoid muscle used for injections is noted to be free of arteries and major blood vessels (although there are variations and site of injection is important in order to avoid major nerves and blood supply). The diameter of the needles used to vaccinate are between 0.33mm and 0.25mm (23gauge -25gauge) depending on type. The diameter of capillaries in muscle tissue are between 2-10 micrometres, or 10 times smaller than the needle. Not like the veins that snake up your arm and are used to take blood from.

For years vaccinators have debated the best location and injection technique for giving vaccines. My aforementioned minor contribution on the matter was a clinical trial of three injection techniques and FWIW found no difference when aspiration was used compared to not for injection site reactions or pain on injection, also supported by a systemic review of randomised trails and more recently [here]. With respect to the value of abandoning or retaining the practice of aspiration, the consensus has been to abandon it as having no added value and at the same time adding to the injection process and pain. Not aspirating is based on current best evidence. However, I would add that there are variations in vaccinator technique with some individuals veering away from best practice as is evidenced by the location of the plaster one sometimes sees on peoples arms!

We do not actually know the frequency of ‘hitting a vein’ apart from that ‘it hardly ever happens.’

Is there any evidence for biological plausibility?

A recent study examined the effect of intravenous administration of mRNA vaccine compared with intramuscular administration in mice. In these animals, intravenous administration was associated with the development of more severe subclinical myocarditis than intramuscular administration. The mice remained well but displayed histopathalogical evidence of myopericarditis. It is thought that this model might be helpful in understanding the mechanism of post vaccine myocarditis.

Before you go off and use this study as definitive proof that an injection technique that does not use aspiration is the cause of post vaccine myocarditis there are some important points to keep in mind.

  • These are mice not humans so while useful we must be cautious in generalising to humans
  • The animals were given many times the human dose – 0.25 mcg per gram weight (or 250 mcg per kilo) compared with the human dose of 0.4 x 10-3 per 70kg person. In other words, vastly more than a human dose. [Here for discussion]
  • Although the mice lost some weight after vaccination they remained healthy and active regardless of the route of vaccine administration. There was no increased death of the mice
  • These data suggest that is plausible that the rare inadvertent intravenous administration of vaccine could increase the risk of myopericarditis. However, there is also plenty of criticism about the choice of (BALB/c) mouse model etc, and interpretation of histological findings, particularly around the potential misinterpretation of common background lesions. In other words, this experiment has been heavily criticised.

Just a brief note on another study that examined the effect of intravenous administration of the AstraZeneca viral vector vaccine (ChAdOx1 nCov-19) on the potential for formation of blood clots. They used an in vitro model by incubating healthy human platelets with the AstraZeneca and Pfizer vaccines and a mouse model where both intravenous and intramuscular administration of vaccine was used. When the vaccine was administered into the vein platelets were observed to clump together with the vector and become activated, but not when it was injected into muscle.

There are no human experimental studies, and these models are not necessarily reflections of what would happen in humans given a normal dose of vaccine into a vein. That said, the animal results do not refute the hypothesis.


Firstly, we do not really know if using an injection technique that avoids aspiration results in intravenous administration of the drug. It is assumed to be a possibility given clearly the needle hit a vein. How often do intramuscular injections end up in a vein? And by this, I mean not a nick in a vein but actually delivering the payload into the vein. This could do with further clarification if it is at all feasible.

Secondly, The most pragmatic way to address whether aspiration reduces the risk for myocarditis following mRNA vaccine is through observational studies.  These could compare the patterns of mRNA vaccine -related myocarditis between populations receiving the vaccines through different injection techniques (yes there are some complex methodological considerations to address multiple confounders). There are two reasons that this might be an important thing to do. Firstly, because not addressing this contributes to loss of vaccine confidence. Secondly, because although this hypothesis is unlikely to be the sole cause of the observed increased myocarditis risk associated with these vaccines, it cannot be excluded as a contributor therefore out of an abundance of caution it seems prudent to address it.

Finally, in the presence of COVID-19 there is far less risk for myocarditis associated with the mRNA vaccine  than there is for the disease.