Abstract
Adverse drug reactions occur at a greater frequency in HIV-infected individuals. A 38-year-old Eritrean man was treated with outpatient co-trimoxazole for confirmed Pneumocystis jirovecii pneumonia, but was switched to clindamycin and primaquine due to nausea and vomiting. Following development of methaemaglobinaemia, he was recommenced on prophylactic co-trimoxazole. He was later found moribund with features resembling septic shock and required invasive respiratory support. The diagnosis of a rare, but severe reaction to co-trimoxazole did not become apparent until he was rechallenged with prophylactic co-trimoxazole after recovery from his initial severe reaction. In an era of polypharmacy and an increasing availability of novel drugs, this case is a timely reminder to clinicians of the ongoing need for pharmacovigilance, especially in HIV-infected individuals who may have unusual presentations of an adverse drug reaction.
INTRODUCTION
Co-trimoxazole is widely used for the treatment and prophylaxis of opportunistic pathogens in the setting of HIV infection. 1 Adverse reactions to co-trimoxazole are well documented with hypersensitivity, usually manifesting as a rash, occurring at some time in up to 60% or higher of HIV-positive individuals while on co-trimoxazole treatment. 2 Rarely, severe reactions indistinguishable from septic shock can occur with co-trimoxazole use which may be fatal unless the drug is discontinued. 3,4 We present a case of a life-threatening reaction to co-trimoxazole in an HIV-positive man to increase awareness of this unusual but serious presentation of hypersensitivity.
CASE REPORT
A 38-year-old Eritrean man with a history of mild asthma presented with a three-day history of dry cough, fever and dyspnoea. He had recently been diagnosed HIV-positive, with a CD4 count of 13 cells/mm3 and a viral load of 5171 copies/mL. Chest radiograph and arterial blood gases were normal. Pneumocystis jirovecii was confirmed from induced sputum and he was treated with outpatient oral co-trimoxazole 1920 mg four times a day. He was not administered corticosteroids. This regimen was changed to clindamycin and primaquine a week later due to nausea and vomiting. After a total of 21 days of therapy, he was asymptomatic but noted to have some peripheral cyanosis. An arterial blood gas (FiO2 0.21) showed a pO2 of 12.5 kPa and a methaemoglobin level of 20%; this was presumed secondary to the primaquine which was subsequently discontinued. His glucose-6-phosphate dehydrogenase levels were normal and he was commenced on prophylactic co-trimoxazole 960 mg once daily.
Two days later, he was found moribund and cyanosed and taken to hospital. He was pyrexial (39°C), tachycardic (140 beats/minute) and hypotensive (88/33 mmHg). On examination he had widespread bilateral crepitations and his jugular venous pressure could not be visualized. He did not have a rash. He had a pO2 of 4.68 kPa on FiO2 0.21. He had a neutrophil leucocytosis (20 × 109/L), elevated C-reactive protein (116 mg/dL), and elevated alanine aminotransferase (123 IU/L) and alkaline phosphatase (125 IU/L). Chest radiograph showed diffuse bilateral infiltrates. He had a normal electrocardiogram and echocardiogram. He was commenced on intravenous piperacillin–tazobactam, ciprofloxacin, co-trimoxazole 3840 mg twice daily and methylprednisolone 40 mg four times a day for presumed bacterial sepsis or recurrent Pneumocystis pneumonia. Due to clinical deterioration of his respiratory function, he was transferred to the intensive care unit (ICU) for mechanical ventilation. Following endotracheal intubation, copious fluid was suctioned from his airways with improvement in the appearance of his chest radiograph. Multiple blood cultures were sterile and atypical pneumonia serology was negative. Endotracheal aspirate was negative for Pneumocystis and acid-fast bacilli, and he was switched to atovaquone for pneumocystis prophylaxis. Magnetic resonance imaging of the brain and lumbar puncture were unremarkable.
Following clinical improvement over the next 72 hours, he was extubated and discharged to the ward, where he was recommenced on co-trimoxazole prophylaxis. One hour after the first dose, he developed a temperature of 40°C, pulse 130 beats/minute, blood pressure 75/34 mmHg (Figure 1), a diffuse erythematous rash with conjunctival injection and widespread chest crepitations. Serial chest radiographs showed rapidly worsening infiltration (Figure 1). Mast cell tryptase taken at six hours was within normal limits. He was transferred back to ICU, reintubated, treated with corticosteroids, inotropic support and broad spectrum intravenous antibiotics and was discharged to the ward 48 hours later after a full recovery.
Chest radiograph showing bilateral alveolar infiltrates and observation chart showing acute change in physiological parameters after patient was rechallenged with co-trimoxazole
As the temporal relationship between the administration of co-trimoxazole and the patients clinical presentation was initially vague, he had been treated presumptively for septic shock. The diagnosis of a severe drug reaction only became clear after he was inadvertently rechallenged with co-trimoxazole for the second time. On direct questioning, the patient revealed that he had taken his first dose of prophylactic co-trimoxazole only a few hours before his first admission, thus confirming the diagnosis.
DISCUSSION
Co-trimoxazole is a fixed-dose combination of two antibiotics, sulphamethoxazole and trimethoprim. Most adverse reactions occur in response to the sulphamethoxazole component. 5 While rare, there have been previously reported cases of a sepsis-like reaction in HIV-positive individuals treated with co-trimoxazole. 3,4 The classical features are fever, bilateral pulmonary infiltration and hypotension occurring rapidly after re-exposure to co-trimoxazole, usually within 30 minutes to four hours. 6 Other features that may be present include rash, conjunctival injection, raised inflammatory markers and deranged liver function tests. 6
Although anaphylactic/anaphylactoid reactions are well described to co-trimoxazole, 7 this sepsis-like reaction can be distinguished by the absence of clinical features such as urticaria, laryngeal oedema, bronchospasm and biochemical parameters such as normal levels of eosinophils, immunoglobulin E and, as with our patient, normal mast cell tryptase levels (Table 1).
Comparison of the features of drug-induced anaphylaxis and sepsis-like adverse drug reaction associated with co-trimoxazole
NSAID = non-steroidal anti-inflammatory drug; IgE = immunoglobulin E
While the exact mechanisms are yet to be elucidated, we speculate that this clinical syndrome resulting from reaction to co-trimoxazole may be under-recognized due to similarities to severe sepsis. With the increasing availability of novel drugs for treatment of HIV, continuing use of drugs for the prophylaxis of opportunistic infections and resulting polypharmacy, this case is a timely reminder to clinicians of the ongoing need for pharmacovigilance. HIV-infected individuals may present with unusual clinical manifestations and while attention may be focused on opportunistic infections or adverse reactions to new drugs, adverse reactions still occur to older, well used compounds and should not be forgotten.