An intern (the equivalent of an FY1 here in Malawi) asked me to review a patient with significant respiratory distress but minimal evidence of a chest problem. He just said that things ‘didn’t quite add up’. His ‘gut feeling’ would turn out to be faithful, as I think they usually are.
As I entered the HDU of the adult male medical ward (a 6-bed area which is the only place that oxygen can be administered) I found the patient, a 22-year old male in severe respiratory distress and pouring with sweat. I was first struck by his gasping in severe pleuritic pain, and then the beads of sweat that were literally appearing and running down his face as I watched. The poor guy looked very sick, and terrified.
I think it need not be intimidating to manage such a sick patient as all is really required is a comprehensive ABCDE assessment, as opposed to ‘clever’ history-taking (not that ABCDE assessments are easy, they’re simple but not easy).
This 22-year-old male presented with ‘laboured breathing’ and abdominal distention. His A&E clerking had established a diagnosis of chest infection, but as aforementioned the intern was not entirely satisfied with this diagnosis.
The patient’s first complaint was of abdominal distention, ‘hardening’, and ‘sensations’ in the left upper quadrant of his abdomen starting 2/52 ago. Over the proceeding days he developed sharp pleuritic LUQ pain radiating to his upper back and left shoulder. Subsequent to this he developed shortness of breath but with no cough or phlegm.
He reported a 2/52 history of ‘red’ urine. Also the patient had been newly diagnosed with HIV on this admission (at this hospital HIV tests are a routine occurrence if the patient has not had one in the preceding 6 months, and I saw new cases daily). There was no history of loose stools, vomiting, dysuria, or headache. He was opening his bowels and passing urine as normal. He denied fever but was sweating profusely.
There was no personal or contact history of TB (probably a meaningless negative finding given the ubiquity of co-existent TB/HIV in Sub-Saharan Africa). There were no regular medications or allergies, and no past medical or surgical history. The patient admitted to drinking and smoking but did not quantify amounts.
Despite the obvious respiratory distress (the use of the word ‘respiratory’ here is misleading, I shall explain later) this patient had a clear airway, there was no audible wheeze or stridor.
There was profound tachypnoea and he was obviously having serious pleuritic discomfort. Every time he attempted to breathe his chest movement was abruptly stopped, leaving him breathing very shallowly with a respiratory rate of 50 breaths per minute.
Chest expansion was poor but equal and there was use of accessory muscles (his scalenes and sternocleidomastoids were tenting up in his neck with every breath). Saturations were 94% on high-flow oxygen. Percussion note was resonant throughout but dull in the bases. Breath sounds were present but minimal and vesicular. There were no wheezes or crackles.
His hands were warm and reasonably perfused but with a capillary refill time of around 3 seconds. Pulse rate was 100 beats per minute and regular. Blood pressure was 110/80 mmHg. Jugular venous pressure was difficult to assess due to the profound respiratory distress and accessory muscle use. Heart sounds were normal with no added sounds and no murmurs.
This patient has mild haemodynamic compromise, and also is pyrexial, sweating and tachypnoeic – so he will lose fluids rapidly.
The patient was fully alert, oriented and obeying commands (GCS 15/15), he was moving all four limbs. Blood sugar testing was not available. Temperature testing was not available, but his profuse sweating indicated pyrexia.
If an unwell patient has a reduced level of consciousness and no blood sugar testing is available one can empirically treat as hypoglycaemia – but no IV glucose available.
There was no evidence of injuries, bleeding, wounds, or peripheral swelling.
Given the history I paid careful attention to the patient’s abdomen, and found it to be rigid with particular tenderness and rebounding in the left upper quadrant.
CLINICAL PROBLEM LIST FIRST – THEN CLINICAL DIAGNOSIS
Summary of clinical problems:
- (use of accessory muscles / severe pleuritic discomfort / clear chest)
- (consider: hypoventilation / VQ mismatch / SIRS / sepsis)
- GENERALISED PERITONITIS
- (historical and examination evidence of LUQ pathology / diaphragmatic irritation)
- (potential intra-abdominal source of sepsis / inflammatory splenomegaly)
Now we must consider a differential diagnosis that explains all of the above clinical problems, otherwise our diagnosis is incomplete or incorrect. Here an intra-abdominal problem seemed the likely explanation.
- LEFT-SIDED INFRA-DIAPHRAGMATIC INFLAMMATORY PATHOLOGY, suggested by:
- LUQ pain referred to the shoulder tip (this phenomenon of referred pain is due to the shared innervation of the diaphragm and the C3-5 dermatomes)
- Severe pleuritic discomfort and shortness of breath.
With regards to specific aetiology the following diagnoses could be considered:
- INFLAMMATORY SPLENOMEGALY / PERI-SPLENIC ABSCESS / SPLENIC RUPTURE
- (the postero-superior surface of the spleen abuts the diaphragm);
- (this is a likely diagnosis I feel and could occur with malarial splenic sequestration i.e. with Plasmodium vivax, which actually is rare in Malawi – most cases are falciparum, apparently owing to the Duffy negativity of the native population);
- (many other bacterial, viral or parasitic infections can however cause inflammatory splenomegaly);
- (splenic rupture is possible but given the reasonable haemodynamic status the bleeding must have been tamponaded at the time – but may have later ruptured the splenic capsule leading to haemorrhagic shock and cardiac arrest)
- PERFORATED VISCUS
- (i.e. stomach or colon – may occur with typhoid but the absence of a GI prodrome or constitutional symptoms makes this less likely)
- ABDOMINAL/THORACIC AORTIC DISSECTION OR RUPTURE
- (the presence of back pain and abdominal pain may suggest this but the patient’s reasonable haemodynamic state, lack of radio-radial delay, normal chest X-ray, and a more likely alternate cause make this less likely)
With a diagnosis of generalised peritonitis (from suspected splenic pathology) the patient was referred to the surgeons.
I returned after the weekend to find that the patient had been reviewed by the surgeons who had identified his abdomen was ‘tense’ and ‘rigid’ (their words) but given his level of respiratory distress that the problem was not intra-abdominal and that he should be managed for pneumonia and/or acute asthma.
The patient spent the next 24 hours being treated with salbutamol nebulisers and aminophylline (he had no wheeze on examination therefore I don’t think this treatment was indicated), he later deteriorated and was sent to critical care where he was intubated. He then had a cardiac arrest and sadly died.
The above CXR shows no evidence of a significant respiratory problem (i.e. pneumothorax, collapse, effusion, consolidation). Of note the right hemi-diaphragm is a little raised. This might be explained by the abdominal rigidity tamponading the abdominal viscera (esp. the liver) and preventing their downwards displacement with respiratory movement. Also note that the patient’s respiratory effort was minimal owing to the severe pleuritic discomfort.
Of course a saddle pulmonary embolus can account for severe respiratory distress in the context of a clinically and radiologically clear chest, but a saddle pulmonary embolus is primarily a haemodynamic problem. This patient’s problem was not primarily haemodynamic. Also in my experience it is only intermediate pulmonary emboli that cause pleuritic chest pain (owing to pleural infarction resulting from a more distal embolus). Saddle pulmonary emboli result in severe circulatory failure, and pleural infarction is unlikely to occur before much more clinically obvious circulatory collapse.
Above all, the history and examination findings were not consistent with massive pulmonary embolus.
Given the lack of radiological resources in this hospital I believe an urgent explorative laparotomy was indicated. You might argue that this is not the judgment call for the non-surgeon to make, but with such profound and unequivocal signs of generalised peritonitis the decision seems less nuanced to me. This was a clinically obvious surgical emergency and I feel the onus is upon the surgeons to explain why a laparotomy is not indicated. Unfortunately the surgeon here acknowledged the clinical features of peritonitis but then made an entirely different diagnosis.
I have fed back this information to the surgical department but the reception was underwhelming to be diplomatic. It appears that there is no culture of reflection or root-cause analysis as compared to our (I have since come to realise) quite amazing National Health Service.
In a busy, cramped, under-resourced environment mistakes like this are inevitable. A difficult truth is Malawian doctors do not enjoy the privilege of having high-quality medical education and post-graduate training boards like we have so perhaps these situations are much more likely (I hope this does not come across as lacking in respect for our colleagues over there, I feel it is an important factor to recognise and one I feel very grateful for at home, since being to Malawi). Ultimately though, we all make mistakes and they should form the foundation for reflection and improvement in practise. The occurrence of mistakes is not the problem, but how we deal with them. Therefore at the very least we may all learn from this event.
- Worried? Ask for HELP!
If you ever walk away from a patient with that ‘something doesn’t quite add up’ feeling then always ask for a senior opinion.
- The diagnosis is in the history
As is usually the case the diagnosis was in the history – and it appears everybody here was distracted by the patient’s tachypnoea & accessory muscle use and ignored the preceding 2/52 history of abdominal pain and features of phrenic irritation.
- ‘Chest pain’ and pleuritic pain do not always equate with a chest problem
The spleen underlies the thoracic cage in the LUQ and therefore splenic pathology may present with ‘chest pain’, also the spleen abuts the diaphragm so can cause diaphragmatic irritation and pleuritic discomfort. Also consider gastric, pancreatic, renal and hepatic causes.
- Interpret clinical signs in context, not isolation
Many other practitioners deemed that this patient was suffering from acute asthma or pneumonia because he had a ‘silent chest’ and was ‘using accessory muscles’. Both of these facts were correct in isolation but the conclusion drawn was incorrect. The patient’s chest was ‘silent’ because his diaphragm was splinted by a subphrenic inflammatory pathology and he was taking feeble breaths (try applying your stethoscope to a friend’s chest and see what you can hear if they maintain 10% inspiratory effort). The patient was using his accessory muscles because his diaphragm was splinted.
Accessory muscle use suggests increased work of breathing, which may result from airway obstruction, but not necessarily. In this case it was due to splinting of the diaphragm. There was no element of airway obstruction identifiable here.
- Tachypnoea is a highly sensitive but poorly specific clinical sign
Tachypnoea (like ‘chest pain’) is not synonymous with respiratory pathology so be wary of making assumptions, always perform a comprehensive history and examination and consider metabolic, and non-respiratory causes. Many if not most patients who are critically unwell will have a raised respiratory rate, and some patients with profound respiratory distress will have a non-respiratory problem.
It may appear premature to ask the question “Is Clinical Assessment A Dying Art?” based on the single adverse event described above. Here there was a failure to recognise an obvious clinical problem, and I am sure most registrars would not have missed this. That said, I believe this reflects a wider problem in clinical surgery/medicine in the UK. Many practitioners are familiar with hearing “I’ll see the patient after the scan” in response to a surgical referral, or with the post-take ward round consisting of ‘continue plan’. It is a very common occurrence also for the medical registrar to be called to surgical wards to deal with problems that are simply routine post-operative complications. I can recall common referrals for cranial diabetes insipidus in neurosurgical patients (i.e. post-hypophysectomy, a clinical neurosurgical problem and recognised post-operative complication) and respiratory depression from post-operative opioid use (in fact I almost exclusively saw this on surgical wards so would expect surgeons to be good at dealing with this). Because these problems do not require a scalpel they are often labelled as ‘medical’ and therefore the physician is called upon to deal with them. The Royal College of Physicians has released a bulletin urging non-medical specialties to be wary of misusing the on-call medical registrar service to this end (1).
Owing to delays in clinical recognition patients may languish and ultimately require avoidable treatment escalation (i.e. ITU), or worse a cardiac arrest. Furthermore, this reluctance to acknowledge the value of clinical assessment also creates a barrier against gaining surgical assistance for medical patients without prior radiological evidence (‘get a scan first, assess the patient later’).
It is surely difficult to influence this problem as it exists in today’s medical practice. We must acknowledge that the skill of clinical assessment, be it medical or surgical, is vital for timely and safe patient care. This skill is nuanced, challenging, and its development requires diligence and hard work.
An eminent surgeon once told me that to succeed in surgery you need to do one thing. To “make the right decision, and make it fast.” Many of these time-critical decisions should take place on the ward and therefore are clinical decisions that require a clinical assessment.