To the Editor: New imaging and pathology investigations continually improve diagnostic accuracy. But tests must be used because they supplement clinical deduction, not because they are available, and the constellation of clinical features should not be ignored.

The case report by Chatterjee and colleagues is valuable for describing delayed diagnosis of meningitis, based on imaging which suggested subarachnoid haemorrhage and aspiration pneumonia.1 The 4-day history, examination (raised respiratory and heart rates, high fever) and results of initial investigations (neutrophilia, raised C-reactive protein level, lung consolidation) suggested a primary respiratory infection. The absence of a typical history of onset of subarachnoid haemorrhage is excused by the 5.5- hour hiatus before the patient was found semicomatose.

Subarachnoid haemorrhage was diagnosed because of density in the subarachnoid space on computed tomography (CT). The authors noted a 1980 report of this appearance in a patient with bacterial meningitis.2 They also noted only one previous report of purulent meningitis mimicking subarachnoid haemorrhage on CT scan (in 1994),3 but there is reluctance to publish “negative” outcomes. Aspiration as the cause of upper-lobe consolidation was unlikely. Bacterial pneumonia and chemical pneumonitis affect the lower lobe.4

Clinical findings were not consistent with subarachnoid haemorrhage, and meningitis was the differential diagnosis, so only the overweighted CT results prevented lumbar puncture on Day 0, which would have resulted in earlier, broader antibiotic therapy and possibly resumption of warfarin. By Day 1, it was too late to prevent permanent blindness (it was possibly too late on Day 0, but pupils were reactive at that time).

Even on Day 1, repeat cranial CT showed infarction but less evidence of bleeding; “a disparity between the amount of [alleged] subarachnoid blood and the patient’s clinical condition” was followed by magnetic resonance imaging then angiography and venography, instead of lumbar puncture as suggested by hindsight in the last sentence of the report. Shadows do not always equate with pathology.

Compare an article on the clinical diagnosis of meningococcaemia.5

Holistic care of Chatterjee et al’s patient included anticoagulation therapy. “It probably could have recommenced earlier” than after “a large pulmonary embolus” on Day 13 — perhaps, given the presence of long-term indications (lupus inhibitor, anticardiolipin antibody and previous thrombosis) and cerebral vessel inflammation causing infarction, on Day 1.

The main lesson, which we were all taught as students but needs career-long reinforcement, is in the penultimate sentence of the case report: “Investigations should not be interpreted in isolation from the clinical picture”.

In reply: We agree with Morgan that the symptoms, signs and laboratory investigations in our case report, although non-specific, supported a diagnosis of infection.1 The C-reactive protein level was not available for 24 hours. The unwitnessed drop in level of consciousness that occurred between 09:00 and 14:30 could have been secondary to meningitis or an acute cerebral event, and, while it is true that the lower lobes are predominantly involved in aspiration, they are not solely involved. Consolidation in other gravity-dependent segments, including the posterior segments of the upper lobes, can occur.2

The suggestion that “permanent blindness” could have been prevented is not supported. Fortuitously, an appropriate antibiotic to which the organism was fully sensitive was given from Day 0 (ceftriaxone). Adjunctive supportive care was quickly provided. In retrospect, anti-coagulation therapy could have recommenced earlier, but this remained a difficult decision in the context of the computed tomography findings. It remains unclear how this would have modulated the meningeal process, but it possibly contributed to the complication of pulmonary embolism.

We agree that there is a reluctance to publish what may be perceived as “negative” outcomes, but, educationally, these may be the most rewarding. This case was an uncommon presentation of a common disease, and we considered it sufficiently important to notify other practitioners. Of most importance in this era when technology in medicine advances exponentially, any investigation must be considered only an adjunct to, and not a replacement for, thorough clinical evaluation.