Pneumocystis jirovecii pneumonia (PJP) is a severe and potentially fatal opportunistic infection in immunosuppressed patients . Common clinical signs of PJP include fever, chest pain, dry cough, and dyspnea . Severe PJP may be associated with the Macklin effect, clinically presenting as spontaneous pneumomediastinum (SPM), subcutaneous emphysema (SCE), and pneumothorax (PNX) [3, 4]. Sherman et al.  and Villalona-Calero et al.  were the first to report that PNX and SPM complicated PJP in patients with AIDS. The Macklin effect, which is often the precursor of SPM in HIV patients, is relatively uncommon in patients without HIV—with an incidence rate of 0.4–4% . In this report, we reviewed the case of a 53-year-old man with pure red cell aplasia (PRCA) who developed a life-threatening complication of PJP due to severe respiratory distress syndrome. Multiple studies have established that non-HIV infected patients with SPM, SCE, and PNX are more prone to developing severe symptoms and complications [8,9,10].
In October 2019, a 53-year-old man with PRCA visited the emergency department of Xiangya Hospital, Changsha, China. His family history and psycho-social history were not special. The patient had experienced acute dyspnea for 5 days, accompanied by fever and cough without sputum. To treat PRCA, the patient had taken oral prednisone irregularly before the 5-day period. On admission, he was febrile (37.9 °C), tachypneic (30 breaths/min), and had a pulse oxygen saturation (SpO2) value of 80%. His blood pressure and pulse rate were within normal ranges. The breath sounds of both lungs were clear, without dry or wet rale. Furthermore, the heart boundary was normal, and there were no heart murmurs.
The main abnormalities in laboratory findings on admission were as follows: Total leukocyte, 15.2 × 109/L (normal reference, NR: 4.0–10.0 × 109/L); neutrophils, 14.6 × 109/L (NR: 1.8–6.3 × 109/L); lymphocytes, 0.3 × 109/L (NR: 1.1–3.2 × 109/L); albumin, 17.5 g/L (NR: 40.0–55.0 × g/L); lactate dehydrogenase, 656.0 U/L (NR: 120.0–250.0 U/L); glucose, 18.92 mmol/L (NR: 3.90–6.10 mmol/L); D-dimer, 0.55 mg/L (NR: 0–0.5 mg/L); procalcitonin (PCT), 0.61 ng/mL (NR: 0–0.046 ng/mL); C-reactive protein, 116.14 mg/L (NR: 0–8.00 mg/L); and erythrocyte sedimentation rate, 120 mm/h (NR: 0–21 mm/h). The liver and kidney functions were normal. Also, the HIV test was negative. We initiated blood culture and sputum culture upon admission. Because of a low SpO2 under nasal oxygen and poor consciousness, the patient was transferred to an emergency intensive care unit (EICU). He received a series of respiratory supportive maneuvers, such as high flow oxygenation for one hour and non-invasive ventilation for two hours. Then, initial mobile chest X-ray revealed air-space opacities in the bilateral middle and lower zones and hyperlucency at the edge of mediastinum structures, which is suggestive of pneumomediastinum (Fig. 1A). In addition, a chest CT scan was performed, which showed diffuse, patchy, ground-glass opacities in bilateral lungs, along with a large pneumomediastinum, small SCE in front of the sternum, and a small left PNX (Fig. 1B, C). However, the patient’s respiratory status did not improve with the respiratory maneuvers. As a result, he was placed on endotracheal intubation, starting with invasive ventilation in emergency scenarios. The ventilatory protocol included a pressure control rate of 16 breaths/min, a fraction of inspired oxygen (FiO2) of 100%, a positive end-expiratory pressure (PEEP) of 10 cmH2O, and a driving pressure of 20 cmH2O. At that time, the patient received an anti-viral treatment, with 75 mg of oseltamivir twice daily, accompanied by 4.5 g of piperacillin-tazobactam every 8 h. This broad-spectrum antibiotic therapy because of his compromised immune system resulting from a history of prednisone medication.
The patient’s hypoxemia (PaO2 < 50 mmHg) continued into the second day of hospitalization. The medical professionals were unable to perform prone positioning and lung recruitment on the patient because these therapeutic maneuvers would worsen his hypoxemia and facilitate the progression of the Macklin effect. Emergency veno-venous (VV) extracorporeal membrane oxygenation (ECMO) was thus initiated as an ultra-protective lung ventilation strategy. The general ECMO settings were as follows: flow rate, 3.5–5 L/min; sweep, 2.0 L/min; and FiO2, 100%. Ventilator settings were adjusted after initiating VV-ECMO, with decreases in FiO2 from 100 to 40% and a PEEP from 10 cm H2O to 5 cm H2O. A chest X-ray of two-sided “white lung” after ECMO performance was shown (Fig. 1D). The patient’s blood culture and sputum culture came out negative on the third day of hospitalization, confirming no bacterial infection. The G-test value was therefore elevated (600 pg/mL, NR: < 70). The patient tested negative for common viruses, necessitating the suspension of oseltamivir administration. Since traditional diagnostic assays were unrevealing, the patient resorted to the metagenomic next-generation sequencing technology for uncovering potential pathogens. Therefore, DNA sequencing was performed based on the BGISEQ-100 platform (BGI-Tianjin, Tianjin, China). High-quality sequencing data were generated after filtering out low-quality, low-complexity, and shorter reads. Using Burrows–Wheeler Alignment (BWA, http://bio-bwa.sourceforge.net/), the data mapped to the human reference genome were excluded. The remaining data were aligned to the Microbial Genome Database from NCBI, including bacteria, viruses, fungi, and protozoa (http://ftp.ncbi.nlm.nih.gov/genomes/). Finally, the mapped data were processed by removing duplicate reads for advanced data analysis. As a result, we detected a significantly higher infection by PJ in the patient’s bronchoscopy lavage fluid and confirmed the diagnosis of PJP (Fig. 2). Then, the patient was started on trimethoprim/sulfamethoxazole (SMZ-TMP) at a dose of 0.96 g every 6 h. On day 6 of hospitalization, the patient’s condition evolved with an improvement of the fever. Figures 3 and 4 showed the changes in laboratory findings following the VV-ECMO procedure. Except hemogram evaluated throughout the hospital stay, Fig. 3 also showed the alterations of neutrophil-lymphocyte count ratio (N/L) and platelet-lymphocyte count ratio (P/L), which were recognized as useful predictors of disease severity and worse prognosis. Figure 4 showed the kidney and liver functions of this patient. A temperate hepatic dysfunction on day 6 of hospitalization caused by PJ was relieved soon after protective treatment, without kidney injury. Serum lactate levels were increased at the early stage of PJP, and PCT as a marker of infection decreased gradually after the administration of SMZ-TMP.
Meanwhile, plasma D-dimer and fibrin degradation products, as two promising biomarkers to predict perioperative fibrinolysis and bleeding, increased during ECMO performance and decreased after ECMO withdrawal. As of day 9 of hospitalization, the patient had been on antimicrobial therapy for 6 days and showed progressive improvement in respiratory failure, with the disappearance of SPM and absorption of ground-glass opacities in X-ray (Fig. 1E). The ECMO cannulas were removed on day 10 of hospitalization. However, the patient needed prolonged ventilator support and tracheostomy. A repeat CT scan of the chest also showed resorption and disappearance of SPM, SCE, and PNX (Fig. 1F, G). On day 23 of hospitalization, the patient regained consciousness, with his invasive ventilation and tracheal tube removed successfully. No adverse or unanticipated events occurred. After 29 days of hospitalization, he was discharged. With two years of follow-up, the patient showed a good prognosis without recurrence. The patient could understand the duration and efficacy of the treatment.