I think I finally get hyponatremia

My final day of shadowing highlighted everyone’s favourite (and most common) electrolyte abnormality: too little salt. The patient was an 81 year old man who came to AMU having been referred by his GP for “anorexia, 2 stone weight loss and low BP”. The admission clerking elicited progressive dysphagia for the past few weeks. On examination, there was a palpable bladder and the patient was clinically dehydrated. There was however bilateral pitting oedema to the mid shins.

An endoscopy was arranged in light of the weight loss and progressive dysphagia. More interesting from a learning point of view was the Na value of 110 last night. I always groaned at the list of differentials to work through for hyponatremia. Is there a simple way of classifying the causes that we can apply clinically?

Emedicine helped me here. It taught me that Na doesn’t cross the cell membrane, so if there is a low concentration of Na in the blood, it must be because either:

a) Na is being lost from the blood to somewhere else


b) the blood is diluted

Going through option a), there are only a few ways Na can leave the blood. Whichever route it takes, it will take some water with it. Provided that relatively more Na is lost than water, the patient will become hyponatremic. Na can either leave:

a) through the kidneys


b) somewhere else

If it leaves through a) the kidneys, then the concentration of Na in the urine will be stupidly high given the circumstances. The value which is taken as stupidly high is anything above 20mmolL. The causes include renal disease, diuretics (especially loop) and Addison’s (which acts like a dose of spironolactone, leading to a lack of sodium-potassium exchange in the distal convoluted renal tubule).

If it leaves through b) somewhere else, pretty much any orifice or break in the skin that allows fluids through will lead to Na loss with it e.g. vomiting, diarrhoea and burns. The kidneys, which should be working normally, will do everything in their power to preserve the precious Na, leading to a low urine Na concentration (<20 mmol/L).

In both cases, there is a loss of fluid from the body. This leaves the patient dehydrated.

This means that if the hyponatremic patient is hypovolemic, we should think about how Na and water managed to leave the body, particularly focused on whether it was lost renally or elsewhere (GI tract and skin).

What about if the blood gets diluted? This can happen if:

a) too much water is taken in


b) water can’t get excreted


c) the patient is in a state of fluid overload, with relatively more water than Na being retained

Option a) occurs in water intoxication or psychogenic polydipsia. It is pretty rare. The urine concentration of Na would be appropriately low (<40 mmol/L).

Option b) occurs with a concentrated, water-deprived urine ([Na] >40mmol/L). The most celebrated cause is the 3rd greatest bane of medical school after enzyme inducers and the causes of clubbing: the (pretty rare) syndrome of inappropriate anti diuretic hormone, SIADH.

The causes of SIADH initially seem overwhelming. This may be because Interleukin 6 stimulates the release of ADH (or vasopressin, same thing). Since IL-6 is part of the acute phase response as well as chronic inflammation and general crapness, lots of inflammatory pathologies can cause SIADH. The common ones are cancers (especially small cell lung cancer, which may be because of its neuroendorcrine origin) and chest infections (especially Legionella).

SIADH can also occur whenever there is disruption of the pituitary stalk, as the hypothalamus controls the pituitary and the pituitary gets confused without mission control telling it what to do. This means that many CNS disorders, especially trauma, meningitis and stroke, can cause SIADH.

There is also a list of drugs which can cause SIADH. In keeping with the CNS-ish aetiology, I think of them as ‘neuro-psychs’. The main neuro ones are anti epileptics, especially carbamazepine and gabapentin. The psych ones are drugs we give them (SSRIs, amitriptyline, antipsychotics) and drugs they given themselves (ecstasy, opiates).

Option b) also occurs in hypothyroidism, but the exact mechanism is unclear. People with hypothyroidism have a reduced cardiac output, which may lead to reduced stimulation of the carotid sinus baroceptors, leading to extra ADH.

Option b) also apparently occurs in glucocorticoid deficiency. I have not yet found the mechanism behind this, so please do comment if you find/know it.

This means that in the euvolemic hyponatremic patient, we should think of causes of excess water ingestion or impaired water excretion, using urine [Na] to distinguish between them. Consider SIADH, hypothyroidism and glucocorticoid deficiency. 

Finally, option c). The states of fluid overload are the major organ failures. A high urine [Na] (>20mmol/L) implies renal failure as this is a stupid thing for the kidney to do when the body is hyponatremic, so it must be broken. If the urine [Na] is appropriate (<20mmol/L), then consider heart or liver failure, or nephrotic syndrome.

This means that in the fluid overloaded hyponatremic patient, we should think of organ failures.Use urine [Na] to help establish whether or not the problem is renal failure.

I’ll post what the final cause of this man’s hypovolemic hyponatremia was when all the tests come back. He was on diuretics so that’s by far the most likely explanation. His last U&Es this afternoon were Na 113, K 3.9, Urea 7.8 and Creatinine 70.

Clinically, he has no symptoms of severe hyponatremia i.e. GCS is normal with no seizures or confusion. Previous results show this was a chronic hyponatremia, so the correction has to be slow to avoid central pontine myelinolysis. I will write up what I learn about the treatment of hyponatremia tomorrow.

Any comments, especially on any further differentials and how glucocorticoid deficiency causes hyponatremia, will be really appreciated.


Day 2: Atypically typical angina

Day 2 felt a lot easier than Day 1. I knew which platform the train left from.  I knew that my team were a friendly and forgiving lot. Most reassuringly, I knew that I didn’t have much to live up to (the previous F1 hadn’t turned up for about 3 months).

I clerked more new patients in at AMU. At 3pm, a 68 year old woman was referred from her GP for two reasons: 1) atypical chest pain for the last month and 2) a pulse of 48.

This is a woman who had a known background of angina for the past few months, but was adamant that this pain was different. She described 5-10 minute episodes of poorly localised non pleuritic pain in the front of the chest radiating down the left arm and to the jaw. It was like “being wound up” and she made Levine’s sign (a clenched fist over the chest as she described the pain). It was relived by her GTN spray. What then was different to her normal anginal pain? Two things: 1) it was much worse than normal (8/10 severity, compared to 5/10 normally) and 2) it occurred almost exclusively in the morning, and was definitely not related to exertion, nor relieved by rest.

This was odd. A cardiac systems review also showed that her shoes were not fitting her properly for the last month because of bilateral leg swelling. In addition, she was getting short of breath on walking 20 yards, whereas a month ago she went for long leisurely walks.

Palpitations? “Yes doctor. Lasts for maybe 30 seconds, happens anytime, maybe once a day.” *Taps out slow regular beat*

CV risk factor wise, she had hypertension, hypercholesterolemia, a father who died of an MI at 58, an ex smoker for 10 years with a 45 pack year history and no known diabetes. She also had a diagnosis of angina.

Her previous medical history was also relevant for a recent diagnosis of COPD, made 2 weeks ago in the GP by spirometry, as well as rheumatic fever 40 years ago and CKD stage 3.

Her medications included atenolol 25mg BD, aspirin 75mg OD, GTN PRN, alendronic acid 70mg once weekly, bendroflumethiazide 2.5mg OD, losartan 50mg OD, omeprazole 10mg OD as well as a Combivent inhaler.

She lived alone in a bungalow and was fully independent. She wasn’t particularly worried and only came here because her GP told her to.

On examination, she looked comfortable at rest. HS I was quiet and II was very loud. There was also a 2/6 probably pansystolic murmur which was heard throughout the precordium with no radiation to the carotids or axilla.

Her JVP was not raised at 3cm. There was bilateral pitting oedema to the lower third of the shin, with no tenderness or erythema.

The lungs were clear, including the bases.

ECG showed sinus bradycardia (48) with a borderline first degree heart block.

My impression at this stage was that this was weird. Whatever else was going on, there was probably some fluid overload. Given the worsening shortness of breath, and her established ischaemic heart disease, this was probably heart failure, although a worsening of her kidney disease was possible. There were no signs or symptoms of liver disease, which I thought would be the next most likely cause.

The question is why would her heart failure deteriorate over the last month, and also cause ischaemic pain in the mornings. I figured that the ischaemia could occur when the heart failure is at its worst i.e. on lying flat for prolonged periods, as would be the case in the mornings. This could also explain why her idea of ‘rest’ (lying flat) did not help, and exertion (i.e. getting upright) may not be so bad for her. I thought of the causes of heart failure, which could be split into intrinsic problems with the heart (supply) and increased demand on the heart.

Intrinsic problems of the heart include ischaemia (ACS is the main concern), arrhythmia (this patient was in sinus bradycardia) and valvulopathy (this patient had a pansystolic murmur and a history of rheumatic fever).

Increased demand could be due to anaemia, thyrotoxicosis or sepsis.

In this patient, I thought that the most likely cause would be the bradycardia as a result of the beta blocker. You would expect most of the other causes to lead to a tachycardia as a response to an inadequate cardiac output. This would also be consistent with the mild AV block developing, although we had no old ECGs to compare to. The consultant agreed with this 🙂

I suggested the following investigations in the plan:

FBC (anaemia)
U&E (renal function)
WCC differential (any infection leading to increased cardiac output demanded)
Troponin (?ACS)
Continuous Telemetry ECG
Echo (LVEF and valve assessment)

With the following immediate management:

Stop beta blocker
Furusemide 40mg PO
Salbutamol nebulised 5mg

The consultant added a respiratory nurse review and a heart failure nurse review. He added LFTs to look for another potential cause of fluid overload. He also reminded me that I forgot the chest xray in all the excitement.

I noticed that the AMU nurses did an ECG and CXR on most patients. I will now think of AMU clerkings as history, examination, impression, ?CXR, ?ECG, investigations and management. It’s pretty easy to get so focused on the differential and immediate management that the basics can get forgotten.

Day 1: A straightforward SOB case, right?

2pm, 26.07.12. The first patient I saw as a newly qualified doctor was a 26 year old woman with shortness of breath. A glance at her triage notes highlighted a PMH featuring asthma. As I waited for the consultation room to be ready, I wrote down my differentials at the top of the clerking sheet:

  1. Asthma
  2. Pneumonia
  3. PE
  4. Pneumothorax
  5. Panic attack
  6. (DKA)
  7. (Acute heart failure)

I looked at the obs. Temp 37.7 Pulse 122 RR 22 BP 108/80 O2 sats: 98% on air. An early warning score of 2. I felt more nervous as a doctor than I ever felt as a medical student.

“So, what’s brought you in today?”
“Lots of things.”
“What’s bothering you the most?”
“Well, I’ve got this headache.”

This was not what I  expected. Still, this was the patient’s concern, so I had to focus on it.

“Can you tell me more about it?”
“Well, it’s different to my migraines…”
“You suffer from migraines?”

And so I became a statistic. One of the rubbish 77% of doctors who fail to fully let the patient explain their story before a focused question. I was aware of this at the time, and consciously tried to return to her trail of thoughts.

“Yes Doctor.”
“Sorry to interrupt, I just didn’t see it in your notes and wanted to check. You were telling me about your headaches…”

This led to an exploration of no less than 7 presenting complaints:

  1. Headache for 1 month – occipital area, bilateral, worse on coughing, not affected by posture, doesn’t disturb her sleep, 7/10 severity, constant, not throbbing, no aura, no focal neurology, no head injury, never had anything like it before. No rash, neck stiffness or photophobia.
  2. SOB – worsened over the past month, was playing sports with no problem one month ago. Now can’t walk 400 yards. No obvious trigger, not related to work (she works in an office). Actually gets worse during the day, and is worst before sleeping. Constant, but getting worse. Not a particularly sudden onset.
  3. Cough for 1 month. Yellow phlegm, no haemoptysis. No diurnal variation. Quit smoking 2 weeks ago to help; no effect.  No GORD symptoms, no post-nasal drip.
  4. Chest pain for 1 week. non pleuritic. In centre of chest, upper third of sternum. Not affected by pressure. No radiation, sweating, nausea or vomiting. Bilateral. No relationship with food or exertion.
  5. Pins and needles +/- numbness for 1 week. Hands get pins and needles only, whereas the feet can get both. Episodes of minutes then resolves fully.
  6. Constitutional symptoms – 1 week of night sweats and fever. Also loss of appetite (no weight loss) and extreme fatigue. Currently spending >70% of the day in bed. No mood disturbance. Wants to get on with life. Active person before.
  7. Postural dizziness – worse on waking in the morning/sitting up. This patient was actually referred from her GP for the above symptoms and a standing BP of 66/48.

Of note, she had URTI symptoms a month ago which resolved in a week. The GP first treated for asthma 2 weeks ago with a salbutamol inhaler and a 5 day prednisolone course. The patient came back to the GP last week, and was given 500mg amoxicillin as a 7 day course.

She had no previous medical history.

Her mother was diagnosed with hypothyroidism the day before.

She lived with her family and partner. The parents returned from Cuba 6 weeks ago, and the father had a severe productive cough for 2 weeks on return.

On examination, the striking findings were a diffuse, polyphonic end expiratory wheeze and generalised MRC 4/5  in all limb movements.

I could not think of a unifying diagnosis or a plausible combination. I decided to focus on the SOB and cough, as together these offer a fairly compact list:

  1. Asthma exacerbationIn favour: Common things are common. Explains SOB and cough.Against: The diurnal variation was different to that expected in asthma (which is usually worse in the early hours, as the smooth muscle tone of the airways is greatest then in both asthmatics and non-asthmatics). No trigger. The patient had no personal or family history of atopy. It also does not explain the hypotension and would be unusual to cause such severe constitutional upset with asthma.
  2. PneumoniaIn favour: Cough productive of yellow phlegm, SOB and constitutional upset could be explained. Again, common things are common. Could even explain the hypotension if severe. Previous URTI.Against: There was no response to high dose amoxicillin.(This could be explained by patient factors and disease factors. Patient factors include non compliance. Disease factors include bacterial resistance, an atypical pneumonia or an imported infection from the father.)
  3. PEIn favour: A diagnosis that must be considered in any acute breathless patient. Female. Explains cough and possibly chest pain.Against: If present, chest pain would usually be pleuritic. No risk factors. No sudden onset. No calf swelling or tenderness on examination.
  4. PneumothoraxIn favour: If asthmatic, becomes more likely. Explains SOB and possible cough.Against: Does not explain constitutional upset. Does not usually progress over a month.
  5. Addison’s DiseaseIn favour: Possible subclinical Addison’s, which had been unmasked by the recent infection. Infection explains SOB, cough and URTI symptoms earlier on. Addison’s explains constitutional upset, generalised weakness and postural hypotension. Family history of autoimmune disease.Against: Less common (5 cases per million per year incidence in Western countries).



WCC was raised at 12
Neutrophils were raised at 86%
CRP was raised at 14

Everything else was unremarkable.

Chest xray was unremarkable.
ECG showed sinus tachycardia.

My plan was salbutamol 5mg nebulised, ipratropium 500mcg nebulised and a 5 day course of oral prednisolone, I also wanted the amoxicillin switched to doyxycycline or clarithromycin.

I had a lot of internal debate about a d-dimer. On the one hand, if it were positive, we would be obliged to do a CTPA, otherwise it hasn’t really affected management. It would be likely to be positive in any case as the patient is probably infected  and/or has a PE. Given this, then the result is almost definitely positive, which means we should skip the test and go straight to CTPA if clinical suspicion remains.

I presented to the registrar. He felt it was infection and wanted to change to doxycycline. He wanted the patient discharged.

I then presented to the consultant. He was more suspicious of PE, and requested a d dimer and to admit the patient. I was asked to write up therapeutic clexane.

After much paranoid recalculating of the dose based on her weight, my first ever prescription as a doctor was 150mg Clexane SC OD.

Questions I would be grateful for your help on:

  • Why might a pneumonia not respond to amoxicillin?
  • What are the commonest presentations of PE? What are the subtler ones?
  • The U&Es were unremarkable in this patient, and did not even tend toward the lower end for Na or higher end for K. Does this essentially rule out Addison’s? I am aware of the proper investigations for Addison’s; I would really appreciate knowing if well-within-normal U&Es makes Addison’s very unlikely.

Thank you for reading this and I look forward to your comments. I will be sharing my cases here regularly and also turning them into SBAs on drcrunch.co.uk/sba.htm