Purpura fulminans is a life-threatening condition in which there is extensive cutaneous haemorrhagic necrosis in the setting of disseminated intravascular coagulation (DIC). Three sub-types of purpura fulminans have been described in the literature: acute infectious purpura fulminans (often due to overwhelming sepsis commonly due to Neisseria meningitidis bacteraemia), neonatal purpura fulminans (due to hereditary protein C or S deficiency) and idiopathic purpura fulminans (either drug-induced or of unknown aetiology). Here, we describe a case precipitated by trimethoprim-sulfamethoxazole, and review the relevant literature for possible mechanistic explanations.

Case presentation

A 29-year-old non-binary patient (pronoun “they") presented to the emergency department with left-sided neck swelling, painful swallowing, discharge from the left ear, subjective fevers and a rapidly spreading itchy rash over the arms, torso and upper legs. The patient thought they were taking oral flucloxacillin in the community prior to presentation. The patient had a background of chronic mild neutropaenia of unknown cause, atopic eczema, recurrent skin furuncles and previous episodes of otitis externa secondary to methicillin-susceptible Staphylococcus aureus, primarily treated with oral flucloxacillin.

On examination, the patient was febrile. They had an indurated left pinna with mild cellulitis, mild preauricular and level II cervical adenopathy, keratinous debris within the left auditory canal and an extensive maculopapular rash covering the arms, torso and upper legs.

Blood tests revealed isolated severe neutropaenia (0.22x10[[9]]/L from a baseline of 1.4x10[[9]]/L), a normal eosinophil count (0.01x10[[9]]/L), and an elevated C-reactive protein (CRP) of 72 mg/L (normal <5 mg/L). A swab of the left auditory canal showed no pathogens. The patient was presumed to have recurrent otitis externa, as well as a cutaneous eruption secondary to flucloxacillin. They were prescribed ciprofloxacin ear drops, and intravenous cefazolin; the latter was switched to meropenem after receipt of the very low neutrophil count.

Within 48 hours of admission, the patient’s fevers had settled, the rash had significantly faded, the CRP had fallen to 23 mg/L and the neutrophil count had risen to 0.8x10[[9]]/L. The patient was discharged on a course of daily oral trimethoprim-sulfamethoxazole for combined treatment of potential otitis externa and recurrent staphylococcal skin boils.

The following day (approximately 8 hours after commencing the course of trimethoprim-sulfamethoxazole, and 28 hours after their last dose of meropenem), the patient re-presented to the emergency department with facial swelling, posterior headache, nausea and vomiting, acute generalised asthenia and an extensive and rapidly progressive purple rash. Collateral history from the patient’s general practitioner revealed that the patient had, in fact, taken trimethoprim-sulfamethoxazole (rather than flucloxacillin) for nine days leading up to the initial admission, and that was their first exposure to trimethoprim-sulfamethoxazole.

On examination this time, the patient looked unwell, and was hypotensive (80/40mmHg) and tachycardic with facial oedema, but had no oral mucosal ulceration. The left pinna was dermatitic without evidence of cellulitis. There was minimal ear canal oedema with some keratinous debris present, but no obstruction or purulence. There was no evidence of middle ear or mastoid infection. An extensive, and in many places coalescent, purpuric rash was noted over the chest, abdomen and back, and this later progressed to include the limbs and digits (Figure 1).

Blood tests revealed evidence of disseminated intravascular coagulation ([DIC]; INR 1.9, D-dimer >20,000μg/L, and platelets 40x10[[9]]/L), acute kidney injury (creatinine 106μmol/L), modest mixed-pattern liver enzyme derangement and a CRP of 24mg/L. A computed tomography (CT) scan of the brain and cervical spine showed mild stranding around the left lower external ear and mild adenopathy in the parotid space and left neck with no drainable collections. There was no intracranial pathology or skull base bony erosion. The patient’s trimethoprim-sulfamethoxazole therapy was ceased. Due to concerns about septic shock as the cause of purpura fulminans, the patient was commenced on broad-spectrum intravenous antibiotics (meropenem, amoxicillin and vancomycin), and admitted to the intensive care unit for vasopressor therapy. Over the course of the next 48 hours, the patient’s clinical condition significantly improved, and they were weaned off vasopressors.

Multiple microbiological investigations returned with negative results, including four sets of blood cultures, urine microscopy and culture, bilateral ear swabs and HIV serology. The patient underwent punch biopsies of the skin lesions which revealed features consistent with DIC, but no other pathology (see Figure 2).

Antibiotic therapy was de-escalated in a step-wise manner, and ceased altogether by day 6. Subsequent bone marrow examination three weeks later showed increased granulopoiesis, but otherwise no abnormality. The patient was given a probable diagnosis of trimethoprim-sulfamethoxazole-induced disseminated intravascular coagulation, and advised to strictly avoid this agent in the future.

On clinical review 2 months later, the patient’s skin had returned to normal aside from mild discolouration in the region of the previously most extensive purpuric changes of the right forearm. The left ear remained indurated and mildly dermatitic, similar to its appearance during the patient’s hospitalisations. A subsequent flucloxacillin challenge did not cause a cutaneous reaction.

Discussion

Only a handful of case reports have suggested a causal link between a drug exposure and purpura fulminans. These cases were reportedly caused by non-steroidal anti-inflammatory drugs (2 cases,[[1,2]]), levofloxacin (1 case[[3]]), paracetamol (2 cases; one patient had chronic alcoholism,[[4]] and the other had concomitant Klebsiella bacteraemia[[5]]), tocilizumab (1 case[[6]]), granulocyte colony-stimulating factor (1 case, although the patient had concomitant staphylococcal sepsis[[7]]), contaminated cocaine (1 case[[8]]) and phenytoin (1 case[[9]]). Trimethoprim-sulfamethoxazole has been implicated in many cutaneous reactions, including purpuric rashes related to immune thrombocytopaenia. However, a scoping review of the published literature revealed no cases of purpura fulminans triggered by trimethoprim-sulfamethoxazole. The World Health Organization collaborating Centre for International Drug Monitoring (VigiAccess)[[10]] has reported a total of 128,908 reactions against trimethoprim-sulfamethoxazole. The main categories of reactions are: skin and subcutaneous tissue disorders (40%), “general” disorders (10%), gastrointestinal disorders (9%), blood and lymphatic system disorders (6%), and immune system disorders (4%).

Several mechanisms of drug-induced purpura fulminans have been proposed. The offending drug may act as a hapten, triggering a systemic immune response to an antibody-drug complex.[[3]] Some patients appear to have developed a leucoerythroblastic vasculitis precipitating small vessel thrombosis and purpura fulminans.[[3]] A T-cell mediated mechanism has been proposed in other cases due to absence of detectable pathological antibodies or immune complexes.[[9]] Drug-induced liver injury may cause acquired protein C or S deficiency,[[4,5]] or enhanced release of coagulation factors, leading to a hyper-coagulable state.[[6]] Finally, drug excipients or contaminants may cause direct vasculopathy.[[8]] Some authors have refrained from speculating on a potential mechanism, but have argued for a causal link between drug exposure and purpura fulminans based upon close temporal association and absence of feasible alternative causes (e.g., infections).[[2,11]]

Based upon published criteria[[11]] for causality of associations between drug exposure and adverse events, we believe exposure to trimethoprim-sulfamethoxazole was by far the most likely cause of pupura fulminans in our patient’s case. The patient initially presented with a generalised maculopapular rash, fever and worsened neutropaenia 9 days after commencing trimethoprim-sulfamethoxazole. The rash and fever promptly subsided after changing the patient to meropenem and topical ciprofloxacin. Having been clinically well on discharge, the patient then re-presented critically unwell with purpura fulminans 8 hours after re-exposure to trimethoprim-sulfamethoxazole—following which they again had relatively prompt clinical improvement on cessation of this agent. The patient’s calculated Naranjo score was 5 (“probable adverse drug reaction” with score of 5–8).[[12]]

The patient has not had a recurrence of purpura fulminans since this event, indicating that the cause was acquired and transient rather than due to a heritable coagulopathy. Although difficult to exclude entirely, microbiological testing did not suggest an infectious cause, and in retrospect, the appearances of the patient’s ear and neck were in keeping with chronic changes from prior recurrent otitis externa.

View Figures 1 & 2.