Prescribe the damn aspirin. Then think about it.

ACS is not just about sorting out the meds and organise the PCI ASAP, then moving on. If the patient is still alive and sitting front of you following an MI, we need to remember they are at high risk right now of reinfarcting as well other complications.

With this in mind, we should get the treatment on board as soon as possible. If the diagnosis is clear from the history/ECG/troponin, then of course this is no problem. Similarly, if another diagnosis e.g. gastritis seems to be far more likely and there is a negative ECG/troponin (although a negative troponin does not in itself rule out an MI until 12 hours from the symptoms) then it seems reasonable to hold off treatment.

The problem arises when you have those in between patients. I recently saw a 59 year old man with a background of gastritis presenting with chest pain lasting 10 minutes and starting three hours ago. There were no other cardiac features of the pain, but it was different from his usual gastric pain, and more severe. His ECG and initial troponin were normal. Should we start the ACS meds now?

It’s tricky. Perhaps we could classify the ACS drugs into high risk and low risk drugs. The main risk with the ACS drugs is bleeding, and we are most worried about intracranial and gastrointestinal bleeds, as these can have high morbidity/mortality. The high risk drugs, like clopidogrel and heparin and thrombolysis, could be given once we are confident we are dealing with ACS and they meet the relevant criteria within the ACS guidelines. The low risk ones (aspirin, morphine, GTN) can be given immediately provided that there are no contraindications. In fact, in the UK, the prehospital treatment of suspected ACS includes aspirin 300mg provided that there is no allergy. A history of gastritis is not a contraindication in these circumstances.

I had a look at the evidence to support aspirin in MI and also its risks. In the ISIS-2 trial, the main risk of aspirin at 160mg OD for a month was a 0.6% increase in minor bleeding, with no increase in major bleeds (defined as intracranial or gastrointestinal bleeds). There were 804/8587 (9.4%) vascular deaths among patients allocated aspirin tablets vs 1016/8600 (11.8%) among those allocated placebo tablets, which is an absolute risk reduction of 11.8% – 9.4% = 2.4%. The number needed to treat to prevent one vascular death I make as 100/2.4 =  41. That’s pretty good compared to most interventions we do.

I don’t think we should be giving oxygen in ACS by default. Oxygen is a vasoactive substance. It causes vasoconstriction everywhere but in the lungs, where it causes vasodilation. Overenthusiastic administration of oxygen in ACS is bad. It decreases coronary blood flow through vasoconstriction, as well as worsening reperfusion injury through the formation of reactive oxygen species. A recent Cochane review also found no benefit and possibly some harm from giving oxygen to MI patients compared to room air. The BTS states:

Most patients with acute coronary artery syndromes are not hypoxaemic and the benefits/harms of oxygen therapy are unknown in such cases.

NICE  says oxygen only to keep sats > 94% (or at 88-92% in those at risk of retaining CO2).

Let’s get back to that 56 year old man waiting for us. We should check that the initial dose of aspirin has been given, and if it has not been given for whatever reason, we need to get it in there ASAP. The benefit in terms of lives saved for aspirin is as good as the benefit thrombolysis offers. We can also give the other low risk meds like morphine and PRN GTN if there are no contraindications, and hold the high risk meds pending the 12 hour troponin. This way, the patient is getting most of the benefits and little of the harm until we know more about his condition, which will allow us to make a better informed risk/benefit decision on the rest of the meds. I disagree with a school of thought I have seen which does nothing until ACS is confirmed.

I now like to think about groups of ACS patients based on something I read from Dr Tabas at the San Francisco General Hospital:

1.        Definite ACS

Diagnostic ECG changes and/or diagnostic troponin rise

 2.        Possible ACS

Good history, possible ischemic changes on ECG, trace troponin rise (or complicated by other causes of chronically elevated troponin rise e.g. renal failure, heart failure)

3.        Not ACS

Not a convincing story, no evidence from ECG or troponin.

As a caveat, we need to bear in mind patients who may be having silent MIs – for example women, the elderly, diabetics, or post-operative and critically ill patients. Depending on the presentation, following a normal ECG and initial troponin they may justify a 12 hour troponin to rule MI out in the absence of any specific ACS-sounding symptoms, especially if they have CV risk factors.

It’s that middle group (possible ACS) which is the trickiest. I think that as long as the prescriber know the contraindications and risks of the low risk drugs and how they apply to the patient in front of them, then they should be given more routinely and more quickly in possible ACS cases.

How do we classify ACS?

ACS is cardiac sounding pain at rest, lasting at least 15 minutes. The first split is into ST Elevation ACS and non ST Elevation ACS based on the ECG. The universal definition of myocardial infarction with ST elevation on the ECG is:

New ST elevation at the J point [which is where the QRS complex joins the ST segment] in two contiguous leads with the cut-points: ≥0.1 mV in all leads other than leads V2–V3 where the following cut points apply: ≥0.2 mV in men ≥40 years; ≥0.25 mV in men <40 years, or ≥0.15 mV in women.

This reflects the fact that V2/V3 are the most likely leads to have a little bit of ST elevation in normal health. They are also the leads most likely to demonstrate high take off. High take off is a normal variant that usually occurs in the young athletes and represents early repolarisation. There is an upwards curving shape to the ST elevation and the ST segment does not evolve over hours or days.

Non ST elevation ACS is then classified into NSTEMI and unstable angina on the troponin levels, using the cut offs at your local lab.

I’ll write up some more stories from the nights. As a teaser for the next post, this is the ECG of a farmer in his early 50s who had two collapses in 3 days. He had never had a fall or collapse before, and his only medical problem was osteoarthritis of the hip (a classic occupational risk, with all the jumping from tractors). A previous ECG 6 months ago was normal. What does the ECG show and how could that explain the collapses?

Right bundle branch block

Your first patient has an INR of 10

“Was the INR really 10?” asked the CT2 as we left the handover meeting.
“Yeah. It actually said >10 which means its probably 20.”
“Or 100”
“Or she’s just a pool of blood now.”

The patient was sitting up in her chair, doing puzzles. I was relieved. The single biggest cause of death from warfarin is intracranial bleeding, following by GI bleeding. I saw the job ahead of me as:

1. Working out if she’s having any major or minor bleeds
2. Working out a cause for the INR to shoot up
3. Finding out about why she is on warfarin, her usual dosing and targets
4. Screening for any other urgent medical problems.

Major bleeds are intracranial, gastrointestinal and interarticular, or any bleed that causes haemodynamic instability. Minor bleeds are all the others.

She had no features of raised ICP. This means no headache, no vomiting, no focal neurology, no seizures, normal neurological examination including cranial nerves and pupils. There was no Cushing’s reflex apparent in her obs (which I think of as the brain being so shocked it does the opposite of shock: high blood pressure with a low pulse). I did not check for papilloedema. Those last two signs are pretty late.

She had no features of a GI bleed, although she had bleeding haemorrhoids. The blood was fresh, and was only present on the tissue.

She had no pain or stiffness in her joints. No joint swelling on examination.

In terms of minor bleeds, she had a couple of oldish bruises. She had the haemarrhoids as mentioned earlier, but no mucous membrane bleeding e.g. when brushing teeth. No haematuria or epistaxis.

This was a patient with an INR > 8 without any significant bleed (haemarrhoids alone do not count). You could now look up what to do in the BNF:

The main adverse effect of all oral anticoagulants is haemorrhage. Checking the INR and omitting doses when appropriate is essential; if the anticoagulant is stopped but not reversed, the INR should be measured 2–3 days later to ensure that it is falling. The cause of an elevated INR should be investigated. The following recommendations (which take into account the recommendations of the British Society for Haematology(2)) are based on the result of the INR and whether there is major or minor bleeding; the recommendations apply to patients taking warfarin:

Major bleeding—stop warfarin; give phytomenadione (vitamin K1) 5 mg by slow intravenous injection; give dried prothrombin complex (factors II, VII, IX, and X—section 2.11) 25–50 units/kg (if dried prothrombin complex unavailable, fresh frozen plasma 15 mL/kg can be given but is less effective); recombinant factor VIIa is not recommended for emergency anticoagulation reversal

INR > 8.0, minor bleeding—stop warfarin; give phytomenadione (vitamin K1) 1–3 mg by slow intravenous injection; repeat dose of phytomenadione if INR still too high after 24 hours; restart warfarin when INR <5.0

INR > 8.0, no bleeding—stop warfarin; give phytomenadione (vitamin K1) 1–5 mg by mouth using the intravenous preparation orally [unlicensed use]; repeat dose of phytomenadione if INR still too high after 24 hours; restart warfarin when INR <5.0

INR 5.0–8.0, minor bleeding—stop warfarin; give phytomenadione (vitamin K1) 1–3 mg by slow intravenous injection; restart warfarin when INR <5.0
INR 5.0–8.0, no bleeding—withhold 1 or 2 doses of warfarin and reduce subsequent maintenance dose

Unexpected bleeding at therapeutic levels—always investigate possibility of underlying cause e.g. unsuspected renal or gastro-intestinal tract pathology

So my options were vitamin K orally (1-5mg) or intravenously (0.5mg). I preferred the oral option, as I would rather not put a needle in a patient with an INR of 10 if it were not necessary. Of course, this meant that there were no oral versions available on AMU. I had to go to maternity, where they keep ampoules for prophylaxis and treatment of haemorrhagic disease of the newborn. These ampoules can be given IV, IM or oral.

Vitamin K takes a while to work its magic, as the clotting factors have to be made all over again now that vitamin K is back. The onset of action is slower if given orally, but even when given IV, it takes 2 hours to have any effect and a maximum effect within 6–12 h, compared with 12–24 h for oral vitamin K. This means that if the patient is having a major bleed, you need to replace the clotting factors (prothrombin complex, or fresh frozen plasma if none available) to have an immediate effect and then give vitamin K. Put another way, minor bleeds are distinguished from major bleeds in that a minor bleed won’t kill you in the time it takes for IV vitamin K to have its effect.

So far, so textbook. It was more interesting trying to work out why a compliant lady, who had been on warfarin for at least 20 years and never missed a dose or anticoagulation clinic, had suddenly developed this INR. She denied any alcohol over the past week. She did mention a recent diagnosis of “colitis” 3 weeks ago but was unable to elaborate further. This week she had had diarrhoea for 6 days, initially passing stools 6 times a day with urge inconitenence, but now just 4 times with plenty of warning and feeling much better in herself.

I figured this must be mild UC which was resolving, given she was down to 4 stools daily. I stopped the metronidazole she had been prescribed the day before as the most common cause of a deranged INR is a drug interaction, and metronidazole is a well known inhibitor of the enzymes that metabolise warfarin.

The INR the next day corrected to 2.1 and she was discharged with close INR monitoring in the community.

There are no guidelines on inpatient hyperglycemia

Following on from the COPD blog post, the same patient later had a series of interesting capillary blood glucose levels. Bear in mind he was a known type 2 diabetic on gliclazide.

5:30am: Glucose = 2.7

Action by nurse: Orange juice given. To repeat in 3 hours.

8:30am: Glucose = 16.1

Action: Normal breakfast given. Sugary drinks avoided (e.g. no sugar in tea).

12:30pm: Glucose unrecordable (high). Ketones 0.7.

The patient was on a reducing dose of prednisolone for his COPD exacerbation, which we could not really stop. We therefore were thinking of how to control his glucose levels, bearing in mind his gliclazide and risk of hypos. This was after all a patient who had 2.7 earlier today.

I was wondering at this stage if we really need to do anything right now about his hyperglycemia. It is likely to be transient, and a brief chat with the patient revelaed that this happens “every time I am on steroids”. He has had a chest infection too, which would worsen glycemic control. I decided that we needed to know his fluid status and whether or not he was symtomatic. He was passing clear urine and felt no thirst or other signs of dehydration (hyperosmolar hyperglycemic state). He also had no vomiting, abdominal pain or confusion (? DKA, as type 2 diabetics can still get this).

The diabetic nurse was on leave. I could not get through to an endocrinologist. I had to work out a plan.

I suggested hourly BMs initially. This did not impress the nurse, who felt more tight glucose control was necessary. I did some reading, and found that transient, asymptomatic hyperglycemia does not always need treatment. Given his previous level of 2.7 and the fact he was on gliclazide, I figured that the benefits, if any, of lowering glucose in an asmyptomatic patient with a clear, transient cause probably did not outweigh the real risk of a hypo.

Then I read this.

I discussed this with the CT2, and took the advice in this article into account and we came up with the following plan:

Measure capillary glucose before each meal. If >20,  give 4 units Actrapid before the meal. Repeat capillary glucose at 30 mins (= onset of action) and 2 hours (= peak action) and if hypoglycemia ever suspected post Actrapid. If <20, do not give Actrapid.

See diabetic nurse/team as soon as possible.

Any thoughts on this plan?

Edit 16/5/2013: For any hyperglycaemic patient in hospital, first rule out the emergencies of DKA and HHS. Then, look for an underlying cause e.g. dose error/missed doses, infection, MI, stroke and pancreatits. Then, if the patient is otherwise well with the hypergylcemia, refer to the diabetes team to consider tweaking the antidiabetic meds. You don’t usually have to give any insulin in these situations. As a consultant endocrinologist told me, acute hyperglycemia by itself does not kill.

Hypoxia kills. Hypercapnia happens.

“Steady on his O2 there. He’s a COPD patient; we don’t want him to stop breathing.”

A 73 year old man with COPD not on home oxygen and usually well-controlled had come in with an exacerbation. His ABG showed a clear type I respiratory failure with a normal base excess, and predominantly emphysematous changes on the xray. His saturations were 93% on 60% oxygen. I overheard doctors and nurses worried that he was on 60% oxygen, and they moved him down to 35% oxygen. They were concerned that he would stop breathing.

The hypoxic drive theory as it is taught is debunked  yet its legacy remains. The BTS states this – search their article for “hypoxic drive” and see for yourself. That’s not to say that if you give oxygen to a COPD patient, you wouldn’t increase the PCO2. This can and does happen. The reason behind this has nothing to do with decreased respiratory drive. Let’s dive into the physiology of COPD.

Our respiratory drive comes about 85% from chemoreceptors in the medulla and 15% from chemoreceptors in the carotids and aorta. The ones in the medulla sense the pH of the CSF, which is a surrogate for CO2 levels. That’s why any metabolic acidosis can also generate an increased respiratory rate despite normal pO2 and pCO2.

The peripheral chemoreceptors are more interested in oxygen. If oxygen levels drop below 60 mm Hg, then they get excited in most people. When high flow oxygen is given, it is true that this drive may be reduced so much as to be virtually switched off. However, most of the respiratory drive comes from the central chemoreceptors, so even if this were switched off, the patient should still breathe. In fact, if the patient were at all ‘dependent on the hypoxic drive’ and this was switched off, the lack of ventilation and subsequent CO2 rise (as pCO2 and ventilation are inversely proportional) should switch the medulla chemoreceptors into overdrive. Studies have also shown that the minute ventilation is not decreased in COPD patients given high flow oxygen. So why does the pCO2 rise?

There are two changes that occur when high flow oxygen is given to a COPD patient:

1. V/Q mismatch worsens

Hypoxic vasoconstriction is nature’s way of identifying poorly performing alveoli and not wasting valuable blood on them. Suppose an alveolus is underventilating for whatever reason. As oxygen levels in the alveolus drop, the smooth muscle of the pulmonary capillaries around that alveolus constrict, reducing the flow to the underperforming alveolus. This means that the blood is free to seek out better performing/ventilated alveoli, leading to better ventilation perfusion matching.

It’s like the UK government cutting funding to sports in which Team GB failed to get a medal during the Olympics. The resources (blood) could be used for greater effect if spent on sports where we have real medal chances (well ventilated alveoli).

Suppose now that high flow oxygen is given. Suddenly, no alveolus is so hypoxic that it can’t get any blood. In fact, the blood now finds its way to all the alveoli, and is not necessarily directed to the crème de la crème.

It’s like the UK government saying that any sport that gets a copper medal gets funding, then introducing a copper medal at the Olympics for anyone who finishes in the top 8 positions.  Whilst this may seem great for the diversity of sports represented (more alveoli getting blood), it means that the funding for the top performing sport has to be drastically reduced, which is not the most effective use of resources.

The remaining alveoli with actually decent ventilation are no longer getting the most blood, and poorly ventilated alveoli are now being used to try and get rid of CO2. And that won’t work.

2. The Haldane effect

As pO2 increases, this decreases the affinity of Hb for CO2. This means less CO2 is bound to Hb, leading to an increase in the partial pressure of CO2 (without an increase in the total CO2 content of the blood).

Both these effects would also happen if Mo Farah were given high flow oxygen. The reason he doesn’t go into Type 2 respiratory failure is firstly because he probably has better ventilated alveoli than the typical COPD patient and secondly because he would compensate for any increase in the pCO2 by increasing his minute ventilation, which would rapidly normalise his pCO2. It is the very fact that he can compensate for any rise in pCO2 that makes him safe to have high flow oxygen. In any patient with a reduced ability to increase the work of breathing/ventilation rate, high flow oxygen could potentially lead to dangerously raised pCO2 levels. Note how this does not just mean the select group of COPD patients who retain CO2, but also those with any possible ventilation problem e.g. chest wall deformities, neuromuscular problems and the morbidly obese. The BTS guidance reflect this:

“For most patients with known chronic obstructive pulmonary disease (COPD) or other known risk factors for hypercapnic respiratory failure (eg, morbid obesity, chest wall deformities or neuromuscular disorders), a target saturation range of 88–92% is suggested pending the availability of blood gas results.”

So, giving high flow oxygen does not lead to apnea, although high oxygen flow rates can lead to a rise in CO2. A rising pCO2 can be looked for with both blood gases and clinical assessment of the patient (drowsiness, jitters and twitches, bounding pulse and raised blood pressure etc.).  There will nearly always be some rise in the pCO2 in a COPD patient as more oxygen is given, for the two reasons given above. Even if the pCO is rising, this in itself is not an indication for NIV or invasive ventilation. It is only when the pH drops below 7.35 that ventilation is indicated. (See page vi8 of the BTS guidance).

The warning signs of raised pCO2 will generally happen slowly, and a rising pCO2 leading to acidosis can be managed by NIV or intubation as necessary. A patient who has died of hypoxia cannot be managed.

Bottom line: make giving enough oxygen your first and foremost priority in any acutely unwell and/or hypoxic patient, no matter what.

Bloody alcohol

Today, a patient had a macrocytic anaemia, low platelets and an oddly low neutrophil count despite having clear signs and a chest xray suggestive of a lobar pneumonia. To my on-the-ball consultant, this was explained by chronic alcohol abuse.

Pretty much all doctors, nurses and even orthopaedic surgeons know that alcohol abuse can cause a macrocytosis of the red blood cells. But why? And what else does it do to the blood?  Are there any blood tests that highlight alcohol abuse?

I’m not quite sure what audience this review about alcohol and the blood is aimed at, but it’s very readable whilst explaining things from first principles. Alcohol basically has a bone marrow suppressing effect, potentially leading to:

1. a macrocytic anaemia

2. a suppressed white cell count, especially neutrophils

3. a much reduced platelet count.

Paradoxically, the patient is at increased risk of both bleeds (e.g. intracranial bleeds) and clots. The bleeding risk is because of a decrease in the effectiveness of the platelets, as well as reduced platelet numbers. The patient could also have alcoholic liver disease and its associated clotting factor production failure. The clotting risk is more complex. Fibrinolysis (i.e. getting rid of unwanted clots) is a balance between tissue plasminogen activator (tPA), which wants to make plasmin and so lead to fibrinolysis, and plasminogen activator inhibitor (PAI-1), which stops plasmin production. At moderate levels of alcohol consumption, tPA is stimulated. This may explain the protective effect of moderate alcohol intake against myocardial infarctions. At higher levels, PAI-1 is also stimulated, which would increase the risk of inappropriate clots not being mopped up.

There are multiple reasons for alcohol to affect the red blood cells, which can be split into:

1. Vacuoles messing with the membrane – these generally disappear after 3-7 days of abstinence. These are found in most alcohol abusers. Anything that messes with the red cell membrane can lead to macrocytosis. Cell membrane failure is in fact the mechanism behind macrocytosis in liver disease (cholesterol metabolism is impaired, leading to increased cholesterol and phospholipids deposition in the red cell membrane).

2. Sideroblastic anaemia – one third of alcoholics have some form of sideroblastic anaemia. This is because alcohol screws up an enzyme in the haem sysnthesis pathway, leading to iron (in the form of ferritin) wandering around with nothing better to do than form ferritin rings called sideroblasts in the cell. This prevents the cell from maturing further (and is usually a microcytic anaemia).

3. Co-existing B12/folate deficiency – alcoholics don’t eat proper food.

4. Don’t know – this is the most common explanation I found for the cause of macrocytosis in alcohol. It takes a good month or two from abstinence for the cells to recover to normal size, so it is likely to affect RBCs early in their development.

Are there ways of detecting alcohol abuse in the blood? Again, medical students, doctors and most orthopaedic surgeons would know something about gamma GT. I have heard many people give their own opinion on it being incredibly useful or totally worthless, so I decided to look into the relationship between gamma GT and alcohol.

Gamma GT is a membrane bound enzyme. It is raised in most liver diseases, especially ones that feature biliary obstruction. A WHO/ISBRA study found that 52% of chronic alcohol users had an elevation in the level of gamma GT. In other words, 48% did not. The test was more sensitive for alcohol abuse in men.

The real problem with this test is the lack of specificity. If you are  presented with a set of LFTs and asked to comment on the gamma GT, the trick is to look at the AST:ALT ratio. The greater the AST relative to the ALT, the more likely  it is to be an alcoholic issue (AST:ALT > 2 as a guide). The greater the ALT relative to the AST, the more we think about viral hepatitis, NASH and Wilson’s (which has an ALT:AST >4). Many chronic liver diseases have roughly equal levels of AST and ALT. An elevated gamma GT in a patient with nothing to suggest biliary obstruction, combined with an elevated AST:ALT and a FBC suggestive of alcohol abuse as discussed earlier would make me more confident of alcohol abuse than an elevated gamma GT in a patient with obstructive jaundice and an elevated ALT compared to AST.

I also learnt that it takes about 2-3 weeks for gamma GT to come back down following abstinence. To me, gamma GT when applied to alcohol abuse seems more like a tumour marker in how we should use it. Combined with a clinical history and having thought about other causes of it being raised, we can use it to monitor progression in a patient diagnosed with the condition. We need to avoid using it in isolation to make a diagnosis.

Socially, physically, mentally. There aren’t many more ways in which a drug can be damaging. It amazes me how alcohol abuse is tolerated as a way of life in this country. If I were in charge, tobacco would be illegal, alcohol abuse necessitating police or health action punishable with a fine and psilocybin/cannabis would be used under controlled conditions with carefully selected patients for the medical benefits. Just my thoughts.

Edit 16/5/2013: Professor Nutt has been doing plenty of interesting research on psilocybin as a potential antidepressant.

On Call Week: Liverpool Care Pathway, STAT

Apologies for the delay since my last post. It was busy.

There were many learning experiences I want to share. Perhaps the most poignant one was writing up the Liverpool Care Pathway for an elderly woman with multiple co morbidities and a likely terminal episode of pneumonia. The consultant made a decision to start the LCP, and I was asked to write the relevant medications.

My stream of consciousness went something like this: What dose of morphine should we use? I don’t know if she is opiate naive…does it matter at this stage? Can an F1 write big doses of morphine up? Oh, she’s on oxygen. Does that have to stop? It’s symptom relief only…but does the oxygen make her feel less breathless? Maybe we could monitor her sats and see if she needs it…but hold on, is that invasive monitoring? She looks dry, and I’m pretty sure the LCP says that actually you can use artificial hydration, I remember a case study on the GMC website. How can I tell if she is agitated, or in pain? Do the family decide…

As ever, when in doubt with a prescription, I speak to a pharmacist. These are the most amazing people in the hospital. They are fountains of knowledge raining on the gasping fish out of water that is the F1 asked to make a decision about which drug to use.

“Hi. I need help.”

“What is it?”

“I’ve been asked to prescribe the LCP to a terminally ill patient. How do I write up a syringe driver? How do I decide the rate and dose? How do I…”


She produced a LCP booklet that was about 30 pages long. I started reading from the front, which felt like opening a new TV and reading the company propaganda from Sony when it’s pretty clear that the things I really need are going to be later in the booklet. I never understand why companies insist on selling you the product in the first few pages after you have clearly just bought it.

I skipped forward to the prescribing section. There were five pharmaceutical targets to the care:

1. No pain

2. No vomiting/nausea

3. No agitation/restlessness

4. No respiratory distress

5. No respiratory secretions

The management of each of these depended on whether or not the patient was already experiencing the symptom. If the patient had the symptom, an appropriate drug should be written up as a syringe driver with a PRN subcut breakthrough dose. If the patient were not yet experiencing this symptom, then a PRN dose should be prescribed. As I began attempting to fill in the prescription chart, I realised I could not write anything unless I knew what the particular needs of my patient were. I decided to see the patient.

The patient was a tall woman who had clearly lost weight recently. Folds of skin hung loosely off bones that are not meant to be so easily visible. She was breathing rapidly but with shallow breaths, with each inspiration accompanied by what sounded like basal crepitations but amplified and coming out the mouth. She did not seem to be in pain, but how do you tell in a patient who is not verbalising or even vocalising? She was as peaceful as he could be with all the secretions in her respiratory tract, and her respiratory rate was around 16.

I asked the nurse looking after her what she thought about her symptom control over the last few hours. I also spoke with her daughters to work out what symptoms (if any) were bothering her most in the last few days.

For pain, we looked through the notes, including her previous prescriptions, and there was no suggestion of any pain nor any history of painkillers above PRN paracetamol. I decided to use PRN diamorphine 10mg s/c for pain control based on  the LCP recommendation.

The patient had not been vomiting or expressed any nausea. PRN haloperidol 1mg s/c was the medication of choice for this scenario. As the patient appeared to be opiate naive, and 10mg diamorphine is a pretty big first dose of opiate, there was a significant risk of inducing vomiting in her last few moments alive, which needed antiemetic cover.

The advantage of haloperidol was that is was also useful for agitation and restlessness. The patient had none of that at the moment, so a PRN dose was all that was needed. The other option recommended was midazolam, which would be more sedating. As her daughters were around and presumably wanted to speak with her, I preferred the less sedating option.

I have since learnt that respiratory distress is often tied to anxiety in the dying patient, and both should be treated together. Relaxation exercises and physiotherapy, as well as basic treatment like sitting the patient up if tolerated, can be helpful. Medically, morphine can be used PRN. There was no need to write up any additional medication in this case. We decided that oxygen was not needed, as the mask was probably uncomfortable and as the respiratory rate did not increase with the oxygen off, the patient probably was not in respiratory distress. I realise that I am making the assumption that respiratory rate and respiratory distress have the usual relationship that they have in non palliative medicine. If anyone has anything to add on this I would be really grateful.

Finally, the respiratory secretions. The patient had symptoms, so needed hyoscine hydrobromide. I gave this as a subcut syringe driver over 24 hours, with a PRN top up as needed.

I didn’t have long before I was back to my usual on call routine of pretending to be in 3 places at once. But just for a few minutes, I felt like a doctor who was independent from the chaos in the rest of the hospital. I was making the care of my patient my one and only priority, and it was rewarding.

Patient UK 

Tees, Esk and Wear NHS Trust’s implementation of the LCP