A 92 year old man with moderate dementia presented to the emergency department after a fall that morning in his nursing home. Radiographic investigation confirmed left sided rib fractures but no pneumothorax. He was discharged back to the nursing home. He returned 4 days later with increasing dyspnoea and continuing chest wall pain, despite paracetamol and ibuprofen analgesia. Examination findings included bilateral coarse crepitations that were worse on the left side, hyper-resonance on the left side, and decreased air entry. The anteroposterior chest radiograph was repeated and demonstrated a pneumothorax of approximately 80%¹ on the left side with a small haemothorax (fig 1A). There was no history of a further fall; however, the patient remained confused and noncompliant.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Anteroposterior radiographs: (A) first, (B) second, (C) third.

The medical team felt that standard treatment for traumatic pneumothorax (28 Fr drain in the fifth intercostal space) would not be well tolerated, owing to his confusion. A different approach was proposed and was accepted by the family. Thus, 800 ml of air were successfully aspirated from the second intercostal space, midclavicular line, using a 16 Fr cannula. This was confirmed by improvement in documented air entry and of oxygen saturation (on 15 litres of oxygen) from 93% to 96%.

The repeat radiograph demonstrated a fluid level in the left lung field extending up to the third rib posteriorly. This appearance was suggestive of a massive haemothorax (fig 1B). The patient’s pulse, blood pressure, respiratory rate, and oxygen saturation remained stable, before and after this investigation.

Repeated clinical examination by three senior doctors, (two consultants and a specialist registrar) confirmed dullness and decreased air entry only at the left base. This contradicted the radiological findings that suggested mid or upper zone involvement of the haemothorax. The chest radiograph was repeated the next day (fig 1C) and showed almost complete re-expansion with a small haemothorax at the left lung base. No physiotherapy or other interventions occurred. The patient’s respiratory rate and oxygen saturations remained unchanged.

DISCUSSION

Manual aspiration is a proven management for spontaneous pneumothorax.^2,3 The technique of choice for a traumatic pneumothorax is intercostal tube thoracostomy. This was not deemed feasible or safe in this confused elderly man. Outpatient management of spontaneous pneumothorax by small gauge chest drain with a flutter valve has also been used successfully.^4,5

On review of the literature there have been no similar cases reported where radiological appearances have deteriorated in this manner following chest aspiration.

Possible explanations for the radiographic appearance of apparent massive haemothorax in our patient include:

1. Secondary pulmonary oedema due to re-expansion.^6–9 This is uncommon. In our case, there were no clinical signs of pulmonary oedema.

2. Failure of expansion of the lung with a raised left diaphragm following aspiration. There is one previous report in the literature of failure of expansion.¹⁰ Our patient’s chest was resonant to percussion, with good air entry in the middle and upper zones, after the aspiration.

3. Our favoured theory: Archimedean*. As the lung re-expanded, the serosanguinous collection in the costodiaphragmatic gutter (from the fractured ribs) was displaced superiorly in the pleural space, giving the appearance of an acute haemothorax.

We believe this case clearly demonstrates the old adage of treating the patient and not the radiograph, and we use first principles (as all scientists should) to support our hypothesis.