breathe case presentations.qxd 26/07/2006 12:03 Page 5 CASE PRESENTATION Pulmonary embolism or Pneumocystis jiroveci pneumonia? Case report Table 1 Vital signs and F. Braiteh1,2 A 33-year-old male presented to the emergency laboratory test results at I. Nash3 department with a 5-day history of exertional dys- presentation pnoea, dry cough, lethargy and an ongoing fever of 38.9°C. He had been previously diagnosed Investigation Result Normal 1Medical Oncology, Division of with left-frontal oligodendroglioma during a range Cancer Medicine, The University work-up following a new-onset seizure 4 months Vital signs of Texas M.D. Anderson Cancer Temperature °C 36.6 earlier. After successful tumour resection and Center, 2University of Texas Respiratory rate cycles·min-1 22 adjuvant radiotherapy, the patient totally recov- Graduate School of Biomedical Heart rate beats·min-1 88 ered without any residual paresis. He was main- Sciences, Houston, TX, and Blood pressure mmHg 126/64 3Dept of Pathology, Hospital of tained on valproic acid and dexamethasone at a O2 saturation % 91 Saint Raphael, Yale School of -1 dose that was tapered down to 2 mg·day . Haematological counts and coagulation Medicine, New Haven, CT, USA. The physical examination was unremarkable. White cells ×109·L -1 6.8 4.0–10.0 The patient was haemodynamically stable but Platelets ×109·L -1 123 150–350 hypoxaemic and anaemic (table 1). Chest radio- Haemoglobin g·dL-1 9.3 12.0–16.0 Correspondence: graphy and computed tomography (CT) were Prothrombine time s 12.8 <13.0 F. Braiteh performed and the results are shown in figure 1. Partial prothrombine time s 22.2 22.0–35.0 1515 Holcumbe Blvd, Unit 10 Serum chemistry Houston a Sodium mmol·L-1 137 135–145 TX 7730 Potassium mmol·L-1 4.0 3.5–5.0 USA Chloride mmol·L-1 107 96–106 E-mail: Carbon dioxide mmol·L-1 21 22–30 [email protected] Anion gap mmol·L-1 9 4–10 Urea nitrogen mg·dL-1 26 8–18 Creatinine mg·dL-1 0.9 0.5–1.2 LDH IU·L-1 950 313–618 D-dimer μg·mL -1 >0.9 <0.9 Arterial blood gas analysis pH 7.52 7.48–7.52 PCO2 mmHg 26 35–45 PO2 mmHg 57 75–100 O2 saturation % 94 92–99 PA-a,O2 gradient mmHg 62 <10 b HIV ELISA Non- Non- reactive reactive LDH: lactate dehydrogenase; PCO2: partial pressure of carbon dioxide; PO2: partial pressure of oxygen; PA-a,O2: alveolar–arterial oxygen tension. Task 1 Interpret the chest radiograph and CT Figure 1 scan, and suggest a diagnosis. Chest radiography (a) and CT scan (b). Breathe | September 2006 | Volume 3 | No 1 83 breathe case presentations.qxd 26/07/2006 12:03 Page 6 CASE PRESENTATION Pulmonary embolism or Pneumocystis jiroveci pneumonia? Answer 1 Answer 2 The chest radiograph reveals diffuse bilateral The CT scan shows bilateral thrombi in the perihilar interstitial infiltrates, predominantly pulmonary arteries of the lower lobes (arrow) in the upper right lobe. The CT scan shows and in the right upper pulmonary arterial bilateral infiltrate in upper lobes, super- branches (window panel), confirming PE. imposed on a mild ground-glass appearance. The clinical presentation is suggestive of Intravenous heparin therapy was initiated. Pneumocystis jiroveci pneumonia (PcP) . The next day, a bronchoalveolar lavage (BAL) was performed and the results are shown in fig- Trimethoprim-sulphamethoxazole was initiat- ure 3. ed, but due to the development of an immediate a skin rash, treatment was changed to pent- amidine. As a result of significant hypoxia (table 1), the dose of dexamethasone was increased to 8 mg daily. The combined presentation of hypox- aemia and hypocapnia, along with the features of new right ventricular strain on ECG (right bun- dle branch block and right atrial enlargement) suggested pulmonary embolism (PE). A positive D-dimer latex assay (Diagnostica Stago, Parsippany, NJ, USA) was followed by spiral CT b scan pulmonary angiography (figure 2). Figure 3 Photomicrographs of BAL using a) Gomori methenamine silver and b) Giemsa stain. Figure 2 Spiral CT scan pulmonary angiography. Task 3 Task 2 Interpret the photomicrographs. Interpret the CT scan pulmonary angiography. 84 Breathe | September 2006 | Volume 3 | No 1 breathe case presentations.qxd 26/07/2006 12:03 Page 7 Pulmonary embolism or Pneumocystis jiroveci pneumonia? CASE PRESENTATION Answer 3 with lower-extremity paresis [15]. Other con- tributing factors in this population are related to Figure 3a shows clusters of brown-to-black chemotherapy and corticosteroids, which have cysts (4–7 μm) containing intracystic bodies prothrombotic effects [16], and the diverse pro- (arrows) with the characteristic appearance of coagulants [17] and fibrinolytic inhibitors "crushed ping pong balls". released by brain tumour cells and surrounding Figure 3b reveals small Pneumocystis jiroveci tissue [18]. Although heparin's efficacy and safe- (carinii) trophozoites (1–5 μm), where only the ty for DVT prophylaxis in elective neurosurgery is nuclei (stained purple) are visible (arrows). proven [19], guidelines for post-operative anti- coagulation and inferior vena cava filter place- ment remain undefined [20]. Lower extremity duplex ultrasound identified In the current study, where there had been a a deep venous thrombosis (DVT) extending from diagnosis of oligodendrioglioma, treatment with the middle of the calf up to the left distal corticosteroids and later presentation with femoral vein. Although it was hard to establish a hypoxia, the patient was at an increased risk for separate relationship between the presenting PE despite the absence of extremity paresis. The symptoms of this patient and those of PcP and suspicion of PE was lowered by the initial con- PE, it is believed that both entities were symp- sideration of a highly probable diagnosis of PcP. tomatic. The patient's condition improved over In an outpatient setting of a low-to-moderate the following days, and he was discharged home pre-test probability of PE, quantitative D-dimer on low-molecular-weight heparin. assay (ELISA or immunoturbidimetric assays) is a sensitive (overall sensitivity 93%) but not specif- Discussion ic test (overall specificity 51%) for establishing Approximately 250,000 patients are hospi- an accurate diagnosis, although it yields reliable talised in the USA each year because of venous information to rule out an acute PE [21]. Due to thromboembolism [1]. PE is frequently an occult its high negative predictive probability (between disease. In patients with DVT, the incidence of 94 and 100%), a negative D-dimer ELISA assay silent PE (40–50%) is five-fold that of symp- (<400–500 ng·mL-1) safely excludes the diagno- tomatic disease [2]. In cancer patients particu- sis of either DVT [21] or PE in low-probability out- larly, although half have been found to have patient settings [22, 23]. The D-dimer assay has some degree of PE on autopsy [3], only one in been proven in all disease outpatient settings, four cases has been clinically diagnosed before but in cancer patients, because of their chronic D- death [4]. PE mortality is elevated (17% at 3 dimer elevation, its specificity is lowered further, months) [5], and despite advances in diagnosis, yielding to a lower positive predictive value, prophylaxis and supportive care, the overall mor- although it does not affect its negative predic- tality has not improved much in the last few tive value. Increased D-dimer levels have been decades [6], or barely decreased [7]. The associ- documented in the plasma of patients with vari- ation of DVT with malignancy is strong, with a ous solid tumours, such as in the lung [24], lifetime risk of 20%. Cancer patients with the colon [25], breast [26], prostate [27, 28], thyroid most elevated risk of DVT and PE are those who [29] and cervix [30]. have been diagnosed with pancreatic adenocar- A positive test does not confirm the diagno- cinoma, advanced gastrointestinal adenocarci- sis, so an angiogram or nuclear scan is necessary noma, brain tumour or locally recurrent rectal for confirmation. In the current patient, the posi- cancer receiving radiation [8]. Compared with tive D-dimer test was followed by a CT scan early-stage breast cancer patients, those with angiogram that revealed bilateral PE. Since the metastatic disease have a three-fold increase in incidence of PcP in patients with cancer is the lifetime risk for thromboembolism (17.3 ver- increasing [31], the simultaneous occurrence of sus 5%) [9]. Neoplasm-independent risk factors two common complications (PcP and PE) in for DVT in cancer patients include all types of patients with highly prevalent disease (i.e. can- surgery [10], indwelling central vein catheters cer) is probably under-accounted for. Therefore, [11], the use of adjuvant haematopoietic growth in the event of a diagnosis of PcP, clinicians factors [12] and any increased immobility. should not exclude the likelihood of a coexisting Patients with brain tumours are at particular life-threatening diagnosis of PE, if the latter is risk of developing DVT (3–60%) [13]. Indeed, otherwise suspected. This is even more signifi- the incidence of PE in high-grade gliomas is ele- cant in patients with brain tumours who are also vated (24–60%) [14], particularly in patients at a higher risk of PcP. In fact, PcP occurs at a Breathe | September 2006 | Volume 3 | No 1 85 breathe case presentations.qxd 26/07/2006 12:03 Page 8 CASE PRESENTATION Pulmonary embolism or Pneumocystis jiroveci pneumonia? higher incidence in immunosuppressed patients month course of prednisone at 20 mg·day-1 may with AIDS, cancers, prolonged corticosteroid be enough to put the patient at an increased risk therapy or post-organ transplantation.