20. We recognise Mr P has significant concerns about his wife’s management, starting from when she was first prescribed insulin in the community, through to her hospital admission. We will address his concerns chronologically in order of events.
21. Each of our advisers have provided some clinical context we think it would be helpful to explain in the first instance.
22. Stevens-Johnson syndrome (SJS) is a rare but severe immune reaction, which affects the skin and mucous membranes. It often progresses to toxic epidermal necrolysis (TEN), which is a severe and life-threatening skin disorder.
23. SJS symptoms often typically begin gradually, often mimicking a common illness before escalating into more severe skin and mucosal issues. For example, with flu-like symptoms such as a fever, sore throat and cough followed by painful rashes, blisters and skin peeling. Triggers are primarily external factors, that provoke an abnormal immune response. Symptoms generally appear one to three weeks after exposure to a trigger, for example in drug related cases.
24. The initial (prodromal) phase lasts around one to three days and feels like flu or a viral infection. These nonspecific symptoms can make an early diagnosis extremely challenging, as they resemble common colds or infections.
25. The primary triggers for SJS are medications, accounting for over 80% of cases in adults. This can be triggered by almost any drug, but the medications with the strongest associations are antibiotics, anti-seizure medications, gout medications, pain relievers and anti-viral medications.
26. Infections can also be a trigger, but this is more common in children. Other factors which are less common include vaccinations, cancers, systemic diseases, contrast agents used in imaging scans, and certain gene variants. In about 10-15% of cases, no clear trigger is identified.
27. Humulin I, the type of insulin Mrs P was prescribed, is a genetically engineered insulin, most similar to the insulin created by the human body. This is why it is the most common insulin for patients to be started on, and it is very rare to experience an allergy to insulin.
28. An allergic reaction to an insulin like Humulin I, is most likely due an additive, rather than the insulin itself. Additives are used to make insulins longer acting. For example, protamine (which is a protein isolated from fish), zinc, or preservatives.
29. Local reactions are the most common reaction experienced from insulin, and they are generally mild with symptoms such as redness, swelling, itching or a rash at the injection site. They usually occur shortly after injection and may resolve on their own, or by changing the injection site.
30. Each of our advisers have explained they have never experienced or seen this type of reaction from insulin, and this is extremely rare. The world’s main SJS/TEN research network, called the RegiSCAR group, lists insulin in the category of ‘no evidence of association’ in their 2024 drug casualty table. This means there is no strong signal of association in all European data between insulin and SJS. We include this as we recognise this was an extremely rare set of clinical circumstances. It is likely that most diabetes specialists have never experienced this clinical scenario.
31. We recognise Mr P explains his wife started experiencing fever and flu-like symptoms after being prescribed insulin (Humulin I). This was prescribed on 30 September, and Mrs P contacted the community diabetic service on 3 October, as she had itchy skin.
32. The diabetes guidance is relevant here, and the Trend guidelines. The guidance sets out the management for patients who have type two diabetes.
33. The NMC code is also applicable here. It says:‘13 Recognise and work within the limits of your competence
To achieve this, you must, as appropriate: 13.1 accurately identify, observe and assess signs of normal or worsening physical and mental health in the person receiving care 13.2 make a timely referral to another practitioner when any action, care or treatment is required’.
34. Our diabetes nurse explains the decision to begin insulin, and this type of insulin specifically, was appropriate and in line with the diabetes guidance. This is because there were very limited treatment options available to Mrs P at this stage, due to her kidney problems and poorly controlled diabetes symptoms. Mrs P required insulin to control her blood sugar, and she would have been at risk if this did not take place. As set out above, this is the standard insulin to start patients on, has been around for a very long time and is generally regarded to be very safe.
35. The records explain the diabetic service advised Mrs P if the itching became unbearable she should stop, and that it would call her back the following day as a courtesy call. Our diabetes nurse adviser explains it was appropriate to have a telephone consultation at this stage, and the correct safety netting advice was given to stop if the itching continued.
36. Our diabetes nurse adviser explains there was an opportunity for the Trust to have asked more questions about the itching Mrs P was experiencing at this point, for example if it was localised to the injection site, or generalised, if it as red, the size of the area and so on. Gathering more details at this stage would have allowed for closer monitoring of Mrs P’s symptoms, to understand if things were getting worse at the next review.
37. We understand a multidisciplinary (MDT) meeting took place the following day. The outcome was to consider switching the type of insulin, if the itching persisted. Our diabetes nurse adviser explains this was appropriate management in line with the diabetes guidance. The Trust explained due to work commitments the nurse was unable to contact Mrs P that day to provide a follow up call, and this took place the following day.
38. The diabetic service called Mrs P on 5 October, and Mrs P reported feeling much better and less itchy. As a result, the advice was to continue the current treatment. This was the correct management based on the changing clinical context at that moment in time, as the issues appeared to have resolved.
39. We acknowledge later that afternoon, Mrs P contacted the service again to explain she had a swollen face and legs. The service called her back and advised switching the type of insulin, in line with the MDT recommendation. Our diabetes nurse adviser explains it was reasonable to advise to switch to another type of insulin, as this was the only suitable treatment that could be used to control Mrs P’s diabetes because of her poor kidney function.
40. We think this was correct advice, but recognise Mrs P was reporting red flag symptoms, and there was an opportunity for the Trust to explore this further, in line with the NMC code. Mrs P’s condition was worsening, as she had a new swelling of the face and legs, alongside her previously reported itchy skin. The Trust could have explored these symptoms in more detail.
41. As Mrs P was experiencing swelling of the face, we also think the Trust should have given safety netting advice for her to attend A&E at this stage in line with NMC guidance. We think this was a failing. We will go on to consider the impact of this later in the report.
42. We recognise Mrs P went on to self-present at A&E on 8 October and we have carefully considered her management throughout the admission.
43. Mr P has raised specific concerns about the decision to continue to administer insulin when Mrs P was on the acute medical unit (AMU). We recognise why Mr P would have worries about this.
44. Our ICU adviser explains when Mrs P was on AMU, she was experiencing very high glucose levels. She had also been given a steroid as part of allergic reaction treatment, which can increase glucose levels further. Her glucose levels had reached 24, which is very high. If high glucose levels are not treated, this can progress to a very severe condition called HHS (hyperglycaemic, hyperosmolar state). This can cause severe metabolic derangement of the blood, and potentially death, and needs to be treated.
45. This was discussed with an endocrine consultant, who advised to use a different type of insulin called NovoRapid. This is still termed an ‘insulin’, but is structurally different to that of Humulin I.
46. If a patient develops an allergic reaction to Humulin I, which is a type of isophane insulin (intermediate acting), it would be appropriate to use a different type such as NovoRapid, which is an insulin aspart (rapid acting). This is a suitable alternative to control blood glucose, as it has a different composition and structural difference.
47. Research shows in documented cases of hypersensitivity to human insulins, like Humulin I, switching to an insulin aspart like NovoRapid has successfully resolved all allergic symptoms, whilst maintaining blood glucose levels. Based on this, the decision to continue prescribe a different type of insulin was appropriate management in line with guidance. We understand why Mr P had worries about this decision, and hope this information provides him with reassurance this was a safe thing to do.
48. Mr P also has concerns the Trust caused a delay in his wife’s diagnosis and treatment whilst she was on AMU, which we have very carefully considered.
49. As set out above, Mrs P presented on Saturday 8 October and was admitted to AMU with severe allergic reaction symptoms. It was recognised she needed to be referred to dermatology. There was no dermatology cover available to the Trust over the weekend. Our dermatology adviser explains nationally there are hospitals that do not have evening and weekend cover, and this is usual in some areas.
50. Our dermatology adviser stresses the key thing about a drug reaction is stopping the drug, and it had been stopped at this point. After this, all that can be done to treat SJS is for supportive care to be provided.
51. The records from 8 October show Mrs P’s rash felt better and was less itchy. If it had been possible for dermatology to review her over the weekend, a diagnosis of SJS may not have been able to have been made at that point in any event. This is due to her presenting symptoms at that time.
52. As there was no cover available over the weekend, the Trust arranged for a dermatology review to take place when cover resumed on Monday 10 October. A plan was in place for if Mrs P deteriorated to be able to be reviewed overnight and for a possible admission to ITU if needed. This was appropriate from a dermatological perspective in the circumstances.
53. The dermatologist was contacted on the Monday 10 October, as planned. The dermatologist provided telephone advice to recommend 50/50 liquid paraffin every three hours, ointment and close monitoring for infections. Mrs P was not seen face to face.
54. The clinical picture appears to have developed at this point. Mrs P had been seen in the eye clinic and was experiencing issues with her vision, eyes and mouth. This is an indicator of SJS and would have pointed to diagnosis at this point. The SJS guidelines are therefore applicable here.
55. The guidelines set out if SJS is suspected, a patient should have a full physical examination, alongside investigations including a biopsy. They set out the need to establish a management plan, with a focus on fluids, nutrition and feeding.
56. Our dermatology adviser explains whilst the skin care advice was appropriate for Mrs P, she should have also been seen face to face at this point for a biopsy, with a clear management plan put into place, in line with SJS guidelines.
57. She was not reviewed by dermatology physically until 13 October, and we think this should have happened sooner. We think this is a failing.
58. As set out above, when SJS presents only supportive care can be put in place to try to do as much as possible to try to repair the skin. It is important to focus on the level of fluids and nutrition to allow the skin to heal. If Mrs P was seen face to face sooner, we think the gravity of the situation could have been understood, prompting escalation. An early diagnosis is crucial for SJS, and it is possible a diagnosis of SJS could have been confirmed sooner. We will go on to consider this impact of this in detail later in the report.
59. Mrs P was then seen by dermatology face to face on 13 October. A biopsy was then taken, and SJS was formally diagnosed. We have carefully considered if the Trust’s management was in line with the SJS guidance, from the point of diagnosis.
60. The SJS guidance sets out how to manage the condition, including initiating a primary management plan. If a patient is not able to maintain adequate nutrition orally, a nasogastric tube (NG) should be inserted for feeding. Continuous enteral nutrition (tube feeding) should be provided throughout the acute phase.
61. The medical records show on multiple occasions that Mrs P could not eat, or that she could eat only a small amount of soft diet, for example yoghurt, due to her mouth pain and mucosal involvement (where the mucous membrane lining or skin is affected). A dietician review was not requested, and the Trust did not consider inserting an NG tube to provide adequate enteral nutrition.
62. The SJS guidance also explains patients must be barrier nursed in a side room for controlled humidity, on a pressure relieving mattress. We cannot see this took place. Based on the above, we cannot see the Trust managed Mrs P’s condition in line with SJS guidance. We think this a failing and will go on to consider the impact of this.
63. We understand Mrs P was then referred to ITU for advice on 10 October. The advice from the Trust was that Mrs P could be cared for on the ward.
64. The intensive care guidelines are applicable here. They set out the different levels of care and criteria for each.
65. The first level is ward-based care, where a patient’s needs are met through normal care on a ward in an acute hospital. Level one is for enhanced care, where a patient needs more detailed observations and interventions. Level two is for critical care (also known as HDU care), for patients needing more increased levels of observations and interventions, and support more than level one. Level three is for patients needing multi organ advanced support (intensive care).
66. At the point of referral to ITU, Mrs P had severe hyponatraemia (blood sodium of less than 120 mmol/l). This is a life-threatening electrolyte disorder due to low sodium levels. She also had partially compensated metabolic acidosis, which can be seen on her venous blood gas analysis. This is a serious condition where there is too much acid in the blood. It meant that meant that her respiratory system had to breath deeper and faster to expel more carbon dioxide, to try and keep the acidity of her blood normal. She also had hyperglycaemia (high blood glucose). This is alongside the clinical picture of her widespread skin rash, with a painful mouth and difficulty eating.
67. Dermatology had also prescribed creams and ointments that required application every two to three hours, and careful monitoring. Her sodium levels needed to be closely monitored, alongside regular glucose monitoring, and nutritional support. Her widespread rash made frequent blood sampling more challenging. The medical records show the medical team could not take blood on one occasion, due to skin abnormalities.
68. Mrs P was presenting with the above issues without any signs of improvement, after receiving ward level care for two days.
69. As set out above, ITU advised Mrs P could remain at ward-based care, with fluids and oral sodium tablets for hyponatremia. Our ITU adviser advises oral sodium tablets are not included in any UK guidelines for managing severe hyponatremia.
70. Our ITU adviser explains although Mrs P’s vital signs were still stable, she had a complex clinical condition involving multiple systems of her body and was at risk with rapid deterioration. She required interventions to prevent deterioration beyond that of level one, alongside enhanced nursing and more frequent input.
71. Taking the above factors into account, the evidence supports Mrs P needed a level two critical care bed and admission to ITU from 10 October onwards in line with the intensive care guidelines. It would have been very difficult for Mrs P’s complex clinical, and nursing needs to be met on AMU, where generally one nurse has to look after at least four acute medical patients. We think there is a failing here.
72. On 13 October, the dermatology team contacted ITU again and to request an admission. The clinical picture had deteriorated even further at this point. ITU advised Mrs P could remain on AMU overnight, and she was transferred the following evening, on 14 October. Our ITU adviser explains the records suggest there was an unexplained reluctance to admit Mrs P to ITU. We think this was a second missed opportunity for Mrs P to have received level two care sooner. We think this is a failing.
73. To summarise, we think Mrs P should have been seen by dermatology face to face from 10 October and accepted for level two care on ICU at this point. We cannot see the Trust managed Mrs P’s condition in line with SJS guidance once she was diagnosed. This is because she was not given the right level of supportive care and intervention, including an appropriate feeding and fluids regime, or nursed in a side room.
74. We will now go on to consider the impact of the above on Mrs P, and if there is any evidence to suggest these failings led to her deterioration or death.
75. Firstly, we have carefully considered if presenting at A&E sooner, would have changed Mrs P’s management and treatment. It is important to recognise we do not know if Mrs P would have presented at A&E on 5 October if safety netting advice had been given. However, on balance we understand she was clearly worried about her condition, as she had been contacting the diabetes service for support. It is therefore reasonable to assume she is likely to have acted on the diabetic service advice and presented at hospital three days earlier.
76. As referred to earlier in the report, our dermatology adviser has explained the key thing about a drug reaction is stopping the drug, and it had been stopped at this point on 5 October. After this, all that can be done to treat SJS is supportive care, for example ITU intervention.
77. Looking at the clinical picture when Mrs P did present at A&E three days later, on 8 October, it was reported her rash felt better and was less itchy. As her condition was in the early stages at this point in time, on balance we think it is unlikely a diagnosis of SJS would have been made if she had presented sooner, from 5 October. It is highly unlikely she would have met the criteria for an ITU admission at this stage.
78. Whilst we do recognise this safety netting advice should have been given, and we acknowledge the chance she could have attended hospital sooner, we do not think this meant a diagnosis and treatment would have been able to be picked up and started sooner. This is due to her presentation at that time, balanced with the rarity of this situation. It is possible Mrs P could have been seen in A&E and sent home, as her symptoms were still nonspecific at that stage, and had not fully progressed. Her condition appeared to start to deteriorate further a few days later, and this is when more information became available.
79. Once Mrs P did present to hospital, as set out above, we think there was delays in her care and treatment being provided. We think a diagnosis and supportive treatment on ITU could have been in place from 10 October. We have also carefully considered the impact of this.
80. By the time Mrs P was admitted to ITU four days later, on 14 October, the records explain her skin was like tissue paper, she had rashes all over her body, most parts developed into fluid filled blisters, alongside eroded skin to the lip and eyes. Our ITU adviser explains this indicates an advanced stage of SJS, and she was extremely unwell at this point.
81. We understand she went on to suffer a cardiac arrest later that night and treatment was subsequently withdrawn. Our ITU and dermatology advisers have very carefully considered if the evidence suggests the failings we have identified lead to Mrs P’s deterioration.
82. Mrs P could have had better supportive care, and closer monitoring of her skin and she was not given the best chance of treatment. This is balanced with that is a very rare and serious condition, and even with this level of monitoring, and an earlier escalation to ITU, the outcome is very likely to have been the same.
83. SCORETEN is a severity of illness score for toxic epidermal necrolysis (SCORETEN). This estimates mortality risk in patients with SJS or TEN. A patient scores one point for each of the criteria present. The criteria are: • age above 40 • presence of cancer • heart rate above 120 • skin detachment more than 10% • serum urea more than 10 mmol/l • blood glucose level above 14 mmol/l • serum bicarbonate level less than 20 mmol/l
84. At the time of confirming her diagnosis, Mrs P’s score was recorded to be ‘3-4’ with a predicted morality of 35.3% - 58%.
85. With a score of ‘1’, morality is predicted to be 3.2%. With a score of ‘5 and above’, this increases steeply to 90%. Our dermatology and ITU advisers explain there is an extremely high mortality risk associated with SJS.
86. Recognising Mrs P had a high predicted morality from SJS/TENS alone, Mrs P was also presenting with a range of other comorbidities alongside this, which may not be fully captured by the risk model. She had kidney disease, very high blood sugar and was a very unwell patient. Taking all of this into account, even with closer monitoring and treatment, we cannot say the sad outcome could have been avoided.
87. We are mindful there were missed opportunities for Mrs P to have had more supportive treatments, which could have alleviated some of her symptoms, such as pain and suffering. It is highly unlikely she would have survived this, but we recognise she was denied the best opportunity to do so. This care and treatment may have helped to correct some of her abnormal physiological paraments like acidosis, low sodium, high glucose, her poor nutritional state and with managing the extent skin damage.
88. Our ITU adviser explains patients with SJS/TEN need significantly elevated calorie requirements for several reasons. The first is due to the body being in a hypermetabolic state. This is where the body burns calories at an abnormally high rate and has an increased energy expenditure. The body also goes into a hypercatabolic state, which is where bodily tissues, particularly muscle, protein and fat break down.
89. This is because TEN triggers a systemic inflammatory response. There is a release of cytokine (a protein that regulates immune responses and inflammation) into the blood and of stress hormones. This raises the basal metabolic rate (BMR) need to support repair in the body. The basal metabolic rate is the minimum number of calories your body needs to perform essential life sustaining functions, like breathing and cell production.
90. The extensive skin barrier loss from the condition also disrupts thermoregulation, the process that allows the body to maintain a stable internal body temperature. This causes increased heat loss, and more energy is used to maintain body temperature.
91. The second is due to the extensive loss of nutrients and fluids. One of the main characteristics of TEN is large scale skin sloughing (the loss of skin). This results in substantial loss of proteins, electrolyse and fluids due the wounds. This creates a negative nitrogen balance in the body.
92. Protein catabolism is the process where proteins are broken down in the body, to produce energy and repair cells. This process becomes higher than the rate of protein synthesis, which is the process where cells build proteins, and is crucial for cell growth. This leads to a low albumin level, which is a protein in the blood and muscle breakdown. These losses of protein in the body mean a higher calorie intake is needed so there is enough protein in the body for healing to take place.
93. The third is there is an increased demand of energy needed for wound healing and tissue regeneration to take place. Regenerating the epidermis (the outer layer of the skin) requires a substantial amount energy. There is also a risk of getting potential secondary infections, such as sepsis, which raise the amount of energy needed for the body to sustain its vital functions. Without adequate calories, patients are at risk of delayed healing, prolonged hospitalisation and higher mortality.
94. TEN also affects mucous membranes, e.g. the mouth and throat, causing painful ulcerations and malabsorption. Malabsorption is a digestive disorder where the small intestine cannot properly absorb nutrients. This limits oral feeding, emphasising the need for enteral or parenteral nutrition.
95. As Mrs P didn’t get this level of care and support, she was not given best chance of combatting the resulting autoimmune response she had.
96. We also recognise what an incredibly difficult and distressing time it was for Mrs P’s family. We do not underestimate how difficult it must have been to see Mrs P in pain and so seriously unwell. It is likely they would have felt more supported if Mrs P was more closely monitored and treated, with more intensive care. Whilst this understandably would always have been a distressing time, we think if better symptom control has been in place some of this distress to them could have been avoided. Learning Mrs P could have had better support and monitoring at the end of her life is likely to exacerbate this.
97. We have looked to see what the Trust has done so far to acknowledge this and put things right.
98. We recognise the Trust carried out an internal investigation, and we acknowledge this is detailed and thorough. It has acknowledged on 5 October Mrs P’s swollen face was a red flag indicator, and she should have been given safety netting advice.
99. It also acknowledges a diagnosis of SJS could have been made sooner and escalation to ITU could have been sooner. We are reassured to see this, and it is clear the Trust has taken this seriously. We understand as a result of its investigation it was considering dermatology cover at weekends, and to increase knowledge and management of SJS.
100. The investigation sets out learning would be shared to the serious incident group, feedback would be given to staff, the report would be presented to clinical governance sessions and submitted to the quality board
101. The Trust has created a detailed action plan which sets out developing a best practice process to manage potential allergic reactions. At the time, these actions were ongoing or scheduled for six months’ time. We would be grateful if, as part of our recommendations, the Trust could provide us with any action that has happened since then in line with this, and any updates.
102. We acknowledge the work the Trust has done so far. After careful consideration, we think the action plan could go further than this, as it does not include any specific detail about following SJS guidance, for example with regards to management of fluids and nutrition we have highlighted. There is no specific mention of the more supportive treatment Mrs P could have had, or the impact this had on her. We therefore ask the Trust to take action.