First report of paracoccidioidomycosis reactivation as a complication of immunosuppressive therapy for acute severe colitis in a caving enthusiast

Background

Paracoccidioidomycosis (PCM) is a fungal disease endemic to Latin America. It is caused by Paracoccidioides sp, which are present in the soil of the American continent from Mexico to Argentina and is primarily hosted by humans and armadillos. It is associated with exposure to rural and agricultural environments, and affected individuals can remain asymptomatic or develop PCM in an acute (juvenile) or chronic (adult) form. During an individual’s lifetime, dormant lesions may be reactivated.1 2 Smoking, alcohol excess and immunosuppression have all been implicated in reactivation of the disease.2 3 Men are affected significantly more than women (11:1), with the majority of cases being reported in Brazil.2 3 The organism Paracoccidioides brasiliensis and Paracoccidioides lutzii are sensitive to a large number of systemic antifungal agents. Azole derivatives, sulphonamide derivatives and amphotericin B are commonly used in clinical practice.4 Although it can be a multiorgan condition, in 25% of patients with the adult type of disease the lungs are the only organ system involved.

Immunosuppressive therapies are widely used to reduce inflammation in patients with active ulcerative colitis (UC).2 5 Unfortunately, this strategy, particularly the use of anti-tumour necrosis factor-alpha therapy, has been linked to the development of severe opportunistic infections, including tuberculosis, endemic mycoses and latent viral infections such as varicella-zoster virus, cytomegalovirus (CMV) and hepatitis B.6 7 Immunosuppressive therapy leading to the reactivation of PCM is uncommon. Covre et al highlights that there were 12 such documented cases between 1976 and 2016.2 In total, only three cases have been reported after the use of immunobiological agents.8

This, therefore, represents the first documented case of PCM reactivation after immunosuppressive treatment in inflammatory bowel disease (IBD). The case highlights a rare but potentially serious complication of immunosuppressive drugs and the need to include it as a differential diagnosis in an unwell patient with respiratory symptoms who has received such therapy, particularly if there has been a history of travel to an endemic area.

Case presentation

A 52-year-old male paramedic presented with loose, bloody stools with a frequency of 20 times per day. He had a history of recto-sigmoid UC diagnosed 7 years previously and had had no prior surgery. He had had rescue therapy with infliximab (IFX) 3 years prior to good effect, continuing for 22 months. He had been taking azathioprine 50 mg one time a day based on metabolite levels and response, plus adalimumab 40 mg every other week, which was started 6 months prior for a flare. He had also become steroid-dependent over the previous few months, flaring when reduced below 20 mg prednisolone daily. As he was prescribed three immunosuppressants, he was taking chemoprophylaxis in the form of co-trimoxazole (960 mg every other day). Flexible sigmoidoscopy showed mild recto-sigmoid colitis with no evidence of CMV on biopsies. Stool cultures were repeatedly negative for routinely tested pathogens, including Giardia sp, Cryptosporidium sp, Salmonella sp, Shigella sp, Escherichia coli 0157, Campylobacter sp, Clostridium difficile and norovirus. He was treated for 8 days with intravenous hydrocortisone and discharged on a weaning dose of 40 mg daily of oral prednisolone after symptoms settled. C-reactive protein (CRP) had fallen from 41 mg/L to 13 mg/L (normal <5 mg/L).

However, the patient was re-admitted 13 days later with worsening symptoms of bloody diarrhoea and frequency. CRP was 70 mg/L, abdominal X-ray showed thumb-printing and repeat flexible sigmoidoscopy again showed moderate colitis now extending into the descending colon (figure 1). Biopsies again were negative for CMV. Intravenous hydrocortisone was recommenced and azathioprine was increased to 75 mg one time a day as thiopurine metabolites were suboptimal.

Potential further management in the form of colectomy was put to the patient. His decision was for a trial of IFX, despite discussion of the risk of serious infection and the lack of data for its use as rescue therapy in this context and clinicians’ advice for colectomy. He had previously experienced induction of remission with this. It was administered at a dose of 5 mg/kg on day 3 of admission as rescue therapy. The last dose of adalimumab 40 mg was 21 days prior to that. By day 8, he was feeling much better, CRP had fallen to 13 mg/L and he had passed his first formed stool in 2 months. He was converted to oral steroids and planned for discharge the following day.

However, on day 9, he became feverish up to 39.5°C and short of breath. Despite 3 days of antimicrobials, including intravenous fluconazole, there was little improvement. CT of the chest, abdomen and pelvis revealed multiple consolidative lesions in both lungs (figure 2). Travel history was revisited and it transpired he had spent time in Costa Rica and Uzbekistan 1 year previously. Beta-D-glucan, a marker for invasive fungal infection was raised at >80 pg/mL on four occasions (normal range<80 pg/mL). Bronchoalveolar lavage showed no significant pathogens. Extended investigations as detailed in table 1 were negative for: Aspergillus antigen, Aspergillus PCR, panfungal PCR and Pneumocystis jirovecii (PCP). Quantiferon and HIV tests were also negative. As the patient was not responding to treatment, antifungals were empirically changed to an echinocandin.

Taking the travel history into account, testing was widened to include dimorphic fungi. This revealed a positive test for PCM on immunodiffusion. Antifungal treatment was then changed to an azole antifungal (initially itraconazole, later changed to posaconazole, then voriconazole, due to difficulties with tolerance and absorption of the different compounds).

Given his life-threatening infection, no further IFX was given, azathioprine was discontinued and steroids weaned. His bowels deteriorated again and a subtotal colectomy was carried out on day 26. This was complicated by a splenic bleed and postoperative wound dehiscence, requiring two further operative interventions. He made a very slow recovery over 70 days on the intensive care unit, requiring intubation, ventilation and renal replacement therapy. Colonic biopsies seen in figure 3 were re-examined and were negative for PCM. He was eventually discharged home with oral voriconazole, to be continued for 6–12 months. His PCM immunodiffusion was negative after 5 months of voriconazole.

Differential diagnosis

Given the findings of fever, shortness of breath, raised inflammatory markers and bilateral consolidation on the CT scan, the main differential diagnoses would be other infectious causes of pneumonia.

Treatment

Guidelines for treatment of PCM are primarily based on two prospective, open randomised trials.4 Queiroz-Telles et al studied 53 patients in a randomised open-label study comparing voriconazole and itraconazole, which demonstrated that both were as effective and well tolerated as each other.9 Shikanai-Yasuda et al observed 42 patients that were randomised to itraconazole, ketoconazole or sulfadiazine and this demonstrated no superiority of any one regimen over the other.10

The case was initially managed with empirical broad-spectrum antimicrobials, including a third-generation cephalosporin, a carbapenem, metronidazole and doxycycline. Fluconazole was also used without significant clinical improvement. Once the raised beta-D-glucan indicated the likely presence of a fungal infection, not responding to fluconazole, antifungal therapy was changed to anidulafungin (an antifungal of the echinocandin group). Once a diagnosis of PCM was made, the anidulafungin was first switched to itraconazole, then posaconazole and later voriconazole, as there were issues with tolerance of the different agents, manifesting as deteriorating liver function tests, nausea and vomiting. Absorption of the oral preparations can be challenging and this was closely monitored through serum levels. Eventually, satisfactory levels were achieved with voriconazole, and the patient was discharged with close follow-up, with a plan to continue the antifungal treatment for 6–12 months.

It was also noted that this patient was taking prophylaxis in the form of co-trimoxazole 960 mg on alternate days when he developed PCM reactivation. There are several reasons why this might have failed. Morejón et al demonstrated that 10/25 patients with coexistent HIV and PCM developed PCM, despite being on co-trimoxazole prophylaxis of 960 mg daily.11 Furthermore, Shikanai-Yasuda et al recommended a dose of 960 mg two times per day to treat PCM.4 This shows that prophylaxis can fail in immunocompromised patients and that 960 mg on alternate days was likely not a sufficient dose. Given there have only been three other cases of PCM linked with biological therapy, none of which mentioned the use of antimicrobial prophylaxis, little is known of what the optimum prophylaxis should be in this cohort of patients.