There follows a guest post by Dr Ros Jones, a retired Consultant Paediatrician and member of HART.
We have heard a lot in the last few weeks about Yellow Card reports for any adverse effects of vaccination, so I shall seek here to give a little background to the system and where it can work well but where it can seriously fall short.
The Yellow Card system was introduced in 1964 following the thalidomide disaster as a way of formalising the reporting of adverse effects, especially for new drugs. Tear-out cards, printed on yellow paper, were inserted at the back of the British National Formulary (BNF), which acts as the bible for UK prescribing. This book, updated twice yearly, was given to every practising doctor. Any doctor or pharmacist seeing a patient with unexpected symptoms relating to a prescribed medicine could quickly complete one of the cards and send it to the regulatory authority. But already you can spot the problem here – the system depends on the health care professional recognising that the symptom might be related to a particular drug, so if the connection is not made then neither is the report. Take for example a busy orthopaedic SHO treating an elderly lady with a fractured hip. Will they think to report this as an adverse reaction to her blood pressure tablets? This matters. Studies show up to one third of hospital admissions are due to iatrogenic causes i.e., drug side-effects. Nowadays, the BNF is an online book and the Yellow Card system is also online, so perhaps even more ‘out of sight, out of mind’, especially if the ward is really busy at the time. If you ask colleagues whether vaccine adverse outcomes have been reported to MHRA, they often reply: “I’m not convinced it was the cause, it could have been due to anything.” But physicians are not responsible for deciding whether a clinical event was caused by a drug or was coincidental – that is the role of the MHRA.
All new drugs and vaccines are subject to trials, starting with animal trials usually involving a number of different species, then building gradually through small pilot studies on humans to establish dosage regimes (for example) and short term safety, before rolling out to large scale trials looking for both efficacy and longer term safety. In such trials, all adverse outcomes will be reported, with the control group acting as the base-line for any symptoms against which the new drug is compared. The system works well for reasonably common side-effects and here the size of the trials is important. You will see in drug information leaflets side effects listed as “very common: affecting greater than 1 in 10 people”, through to “very rare: affecting less than 1 in 10,000”. Generally speaking, “very rare” side effects are only listed if severe. Many drug and vaccine trials are only large enough to detect “uncommon” side effects and for any new drug it is only through post-marketing surveillance that rarer side-effects can be discovered. New drugs are marked in the BNF with a black triangle for two years, to remind doctors to complete yellow card reports. In addition, most drug trials will exclude certain groups – for example children, pregnant women and people with risk factors such as kidney and liver disease, so safety for these groups is very much dependent on animal studies or assumptions from other similar drugs. Species difference in adverse effects may occur too, so damage to the developing foetus may only be seen after a drug starts being used by humans. Certain age groups that are under-represented in trials, such as the very elderly, may also be at greater risk. If post-marketing surveillance reveals an unexpected problem then the drug licence may be withdrawn or modified (e.g. limited to certain age groups, as with the AstraZeneca vaccine).
Where an adverse event comprises something very rare, it may be obvious to the attending physician (or indeed the regulator) that this is a potential drug effect. So for example, thalidomide caused a limb-shortening birth defect, phocomelia. This was totally new to the obstetricians and midwives of the day, so it was quite quickly realised that this was a teratogenic effect. But because of the time lag between drug ingestion in early pregnancy and the discovery of the harm, despite withdrawing the drug, over 10,000 babies were affected across Europe. Turning to the recent concerns raised with SARS-CoV-2 vaccines, cerebral venous thromboses (CVT) are rare. A general physician might only see one case in several years, so when a number were reported, it soon became apparent that they might be linked to the vaccine. But if the side effect is a significant increase in a common condition, it may be much less obvious. An elderly patient admitted with a stroke would not ring alarm bells but if suddenly a lot more people were dying in a two-week period after receiving a new treatment, then this could be highly relevant. Another concern for the vaccines is that they have been granted a temporary licence ahead of the long-term safety reports. Over time, many of those volunteers in the control arm of the trials will have received a vaccination and the power of the randomised controlled trial in terms of assessing safety will be greatly reduced.
So what does a prescribing doctor make of all this? We are taught, “first do no harm”. This does not mean you cannot prescribe a drug with known side-effects (that would rule out most effective treatments!), but it does mean that you have a duty of care to consider the patient in front of you and make a judgement on the balance of risk, i.e., do the likely benefits of this drug outweigh the potential harms? With vaccines, this is made more complicated by the fact that the person in front of you is not ill. Thus, you are weighing up the benefit of preventing severe disease or death from a condition which they may or may not catch (indeed against which they might already have immunity) versus a risk of a significant adverse reaction. Clearly this calculation will be different according to susceptibility to the disease in question, so for SARS-CoV-2 vaccines, a safety threshold for a healthy young adult let alone a child would be set much higher than for a 75 year-old with diabetes.
What of the Yellow Card system – is it fit for purpose? I would argue that in the current situation it is not. The MHRA does not provide raw data for scrutiny but the publicly accessible US vaccine adverse effects reporting system (VAERS) illustrates the scale of the problem. From 2016 to 2020, non-Covid vaccines generated an average of 13 reported deaths per month, which, taking account of the number of vaccinations, equates to 0.56 deaths per million doses. In the first quarter of 2021, a total of 2,103 deaths have been reported following Covid vaccines from 130 million doses, giving a rate 16.2 per million doses, i.e., a thirty-fold increase on the background rate. This extraordinarily high rate has been apparent from the first month of Covid vaccine administration, and is not declining.
It is also notable that reports of CVT in the UK increased significantly after several cases had been reported in the press, thus raising awareness amongst medical staff and leading to a change in advice from regulators.
When rolling out a new drug to a whole population which uses a new and untried technology and which has been commenced under temporary licence without the benefit of long-term safety data, it is vital that any post-marketing surveillance is rigorous. As a minimum, every GP or hospital visit should record vaccination status; everyone receiving a vaccine should be given a pre-paid, pre-addressed card with a tick box for any side effects, to be returned to the MHRA 30 days after vaccination. Only then can even short-term safety be assured.