Pneumocystis polymerase chain reaction and blood (1→3)-β-d -glucan assays to predict survival with suspected Pneumocystis jirovecii pneumonia

J Infect Chemother 20 (2014) 109e114

Contents lists available at ScienceDirect

Journal of Infection and Chemotherapy journal homepage: http://www.elsevier.com/locate/jic

Original article

Pneumocystis polymerase chain reaction and blood (1/3)-b-D-glucan assays to predict survival with suspected Pneumocystis jirovecii pneumonia Yasufumi Matsumura a, Yutaka Ito b, *, Masaki Yamamoto a, Aki Matsushima a, Miki Nagao a, Shunji Takakura a, Yoshitsugu Iinuma c, Satoshi Ichiyama a a

Department of Clinical Laboratory Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan Department of Respiratory Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan c Department of Infectious Diseases, Kanazawa Medical University, Ishikawa, Japan b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 9 July 2013 Received in revised form 15 August 2013 Accepted 5 September 2013

Pneumocystis polymerase chain reaction (PCR) and blood (1/3)-b-D-glucan assays are known to be useful for the diagnosis of Pneumocystis pneumonia (PCP). However, their impact on the outcome of clinically suspected PCP patients has not yet been elucidated. Between January 2008 and July 2011, we prospectively observed 190 immunocompromised patients who had ground-glass opacity on chest computed tomography scans and were suspected to have PCP. The blood b-D-glucan levels of these patients were measured, and PCR was used to detect Pneumocystis jirovecii in the respiratory samples. The 30-day mortality rates and related factors were assessed. The 30-day mortality rate of all included patients was 21.6%. Both b-D-glucan-positive (10.1%) and PCR-positive patients (15.0%) had significantly lower mortality rates than b-D-glucan-negative (28.1%) or PCR-negative patients (30.1%). All of the 13 definite PCP patients had positive PCR and b-D-glucan results, received anti-PCP treatments, and survived. Among the 72 patients who were negative for microscopic detection of P. jirovecii but received anti-PCP treatments, positive PCR results (odds ratio [OR], 0.14; 95% confidence interval [CI], 0.02e0.74), a high Sequential Organ Failure Assessment score (OR, 1.42; CI, 1.08e1.88), and positive b-D-glucan levels (OR 0.25, CI 0.06e1.02) were associated with mortality rates using stepwise logistic regression analyses. A positive Pneumocystis PCR or b-D-glucan test was a candidate predictor of survival in patients who were suspected of having PCP, even though negative for visual detection by microscopy. Ó 2013, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

Keywords: Pneumocystis jirovecii pneumonia Immunocompromised host Polymerase chain reaction Blood b-D-glucan

1. Introduction Pneumocystis pneumonia (PCP) is one of the most prevalent infections in human immunodeficiency virus (HIV)-infected patients [1]. PCP also occurs in non-HIV immunocompromised patients receiving transplantation, immunosuppressive therapy, and antitumor chemotherapeutic agents [2]. The standard method for the laboratory diagnosis of PCP is microscopic visualization of Pneumocystis jirovecii in respiratory samples. However, non-HIV patients often develop PCP with a lower fungal burden than

* Corresponding author. Department of Respiratory Medicine, Kyoto University Graduate School of Medicine, 54 Shogoin-Kawahara-cho, Sakyo-ku, Kyoto 6068507, Japan. Tel.: þ81 75 751 3830; fax: þ81 75 751 4643. E-mail address: [email protected] (Y. Ito).

patients infected with HIV [3]. A significant number of patients without identified microscopic P. jirovecii have a clinical presentation compatible with PCP following complete resolution of the pulmonary infiltrates after a full course of PCP treatment [4,5]. However, it has long been debated as to whether PCP exists in the absence of positive microscopy, and the condition has been treated on a case-by-case basis [4,6]. Polymerase chain reaction (PCR) assays of respiratory samples are highly sensitive and have been used for the diagnosis of PCP. PCR assays have 94e100% sensitivity and 79e96% specificity in the diagnosis of microscopically positive PCP (definite PCP) when compared with controls of pneumonia [7e9]. However, they lack specificity in distinguishing disease from colonization [10]. Quantitative real-time PCR assays have been reported to be more promising for the diagnosis of PCP than conventional PCR [5,11].

1341-321X/$ e see front matter Ó 2013, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jiac.2013.09.004

110

Y. Matsumura et al. / J Infect Chemother 20 (2014) 109e114

However, DNA sequences targeted for PCR and the cutoff values used in these assays have not been standardized. Because respiratory samples are sometimes difficult to obtain, especially in critically ill patients, rapid diagnostic serological tests are needed. Of the various blood markers tested, it has been reported that the blood (1/3)-b-D-glucan (b-D-glucan) level is the optimal test for PCP diagnosis [12e15]. Although these studies have shown that b-D-glucan assays have 92e100% sensitivity and 65e94% specificity for definite PCP, b-D-glucan is not specific for Pneumocystis. Furthermore, false positive results due to factors such as the administration of immunoglobulin, bacteremia, hemodialysis, surgical gauze exposure, and certain antibiotics have been reported [16]. Non-HIV-infected patients with PCP have a mortality rate of 20e 40% [2,17], which is a higher rate of mortality than HIV-infected patients (10e20%) [2,18,19]. Although PCR and b-D-glucan assays have good diagnostic performances for definite PCP, they have not yet been elucidated as predictors of the outcome of clinically suspected PCP patients. In this study, we investigated the mortality rates among immunocompromised patients suspected of having PCP due to the presence of ground-glass opacity (GGO) on chest computed tomography (CT) scans. The significance of positive Pneumocystis PCR and b-D-glucan assays as predictors of 30-day mortality among patients suspected of having PCP without microscopic evidence of P. jirovecii was determined. 2. Patients and methods

usual recommended dose for more than 10 days or until death. Adjunctive steroid therapy was defined as an increase in the dosage or initiation of corticosteroid treatment with more than 5 days of administration. Hypoxia was defined as an arterial PaO2 <70 mmHg at room air or a requirement for supplemental oxygen. Invasive fungal infection was diagnosed according to the European Organization for Research and Treatment of Cancer criteria [21]. Thirtyday mortality rates of patients stratified by b-D-glucan, Pneumocystis PCR, and anti-PCP treatment were calculated. Factors associated with 30-day mortality rates among patients receiving antiPCP treatments who were negative for microscopic detection were determined by univariate and multivariate analyses. 2.3. b-D-glucan assay Plasma b-D-glucan was measured with the WAKO b-glucan test (Wako Pure Chemical Industries, Tokyo, Japan). b-D-glucan values of more than 11 pg/ml (commercial cut-off value) were considered to be positive. The assay was routinely performed at our institution on the same day that the plasma was obtained. 2.4. Pneumocystis PCR Respiratory samples were centrifuged at 1000 g for 10 min. The pellet was resuspended in 200 mL of the original fluid, and DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). Molecular detection of P. jirovecii was performed by single-round PCR amplification of the mitochondrial large-subunit rRNA [22].

2.1. Patients and study design 2.5. Statistical analysis This study was conducted at Kyoto University Hospital, a large 1182-bed university hospital located in Japan between January 2008 and July 2011. We, the infectious disease physicians, prospectively intervened with patients suspected of having PCP (suspected PCP). Immunocompromised patients (those who received immunosuppressive agents including corticosteroids, anti-tumor chemotherapy, or organ transplantation or who had malignancies or HIV infection), patients who had clinical presentations such as fever, cough, sputum, dyspnea, or leukocytosis, and patients who had GGO on their chest CT were enrolled in this study. These patients were tested for blood b-D-glucan, Pneumocystis PCR, and Gomori methenamine silver staining of bronchoalveolar lavage fluid (BAL) or induced sputum (IS). When a patient had a positive b-D-glucan result, preemptive treatment for PCP was recommended to the attending physician, but not always given. If b-D-glucan was negative, empirical treatment for PCP was performed clinically after a discussion with the attending physician. Adjunctive steroid therapy was administered when the patient was hypoxic. Antimicrobial or antifungal treatments were performed clinically by the attending physicians. The Ethics Committee of Kyoto University Graduate School and the Faculty of Medicine (E-1121) approved this study and waived the need for obtaining informed consent from each patient. 2.2. Variables, definitions, and outcomes Clinical information collected from the medical chart review included the age, sex, underlying diseases, immunosuppressive therapies and prophylaxis for PCP during the month prior to sample collection, clinical symptoms, laboratory values, Sequential Organ Failure Assessment (SOFA) score (severity of illness scoring system to predict mortality) [20], anti-PCP treatment, and the final diagnosis of pneumonia. A definite diagnosis of PCP was established by microscopic identification of P. jirovecii using Gomori methenamine silver staining. Anti-PCP treatment was defined as the administration of trimethoprim-sulfamethoxazole or pentamidine at the

Categorical variables were compared using Fisher’s exact test. Continuous variables were compared using the ManneWhitney U test. To determine the association of independent variables with the risk factors for 30-day mortality rates, all variables with a P value of less than 0.10 from univariate analyses were subjected to further selection using a forward stepwise logistic procedure. The variables included in the SOFA score (hypoxia, mechanical ventilation, thrombocyte count, and total bilirubin) were excluded from the multivariate models. The goodness of fit of the last model was evaluated by the HosmereLemeshow test. A P value <0.05 was considered statistically significant. All statistical analyses were performed using Stata version 11.2 (StataCorp, College Station, TX, USA). 3. Results 3.1. Clinical diagnosis and treatment of pneumonia A total of 190 suspected PCP patients with a newly found GGO were included. These patients provided 56 BAL samples and 134 IS samples. Microscopic identification of P. jirovecii was made in 9 BAL and 4 IS samples. The characteristics of these patients are shown in Table 1. Eighty-five patients received anti-PCP treatments, including 13 definite PCP patients (trimethoprim-sulfamethoxazole in all the patients) and 72 patients without microscopic detection (trimethoprim-sulfamethoxazole in 68 patients and pentamidine in 4 patients). The remaining 105 patients did not receive anti-PCP treatments. Under our b-D-glucan-based treatment strategy, 57 of the 69 b-D-glucan-positive patients (83%) received anti-PCP treatments, whereas 93 of the 121 b-D-glucan-negative patients (77%) did not. The final diagnoses of patients stratified by the b-D-glucan and Pneumocystis PCR results are shown in Fig. 1. All of the 13 definite PCP patients had positive PCR and b-D-glucan results, received anti-PCP treatments. Of the 72 patients receiving anti-PCP treatments without microscopic detection, 6 patients had invasive fungal

Y. Matsumura et al. / J Infect Chemother 20 (2014) 109e114 Table 1 Clinical characteristics of patients who were suspected to have Pneumocystis pneumonia. Characteristics

With PCP treatment (N ¼ 85)

Without PCP treatment (N ¼ 105)

P-value

Age, years Male sex Underlying disease Collagen vascular disease Solid malignancy Hematological malignancy Organ transplantation HIV infection Corticosteroids Immunosuppressive agents Other than corticosteroids Anti-tumor chemotherapy PCP prophylaxis Hypoxia Mechanical ventilation Pneumothorax SOFA score Laboratory findings Neutrophils/mm3 Lymphocytes/mm3 Thrombocytes/mm3 C-reactive protein, mg/dL Albumin, mg/dL Creatinine, mg/dL Blood urea nitrogen, mg/dL Total bilirubin, mg/dL Lactate dehydrogenase, IU/L b-D-glucan, pg/ml Positive b-D-glucan Positive Pneumocystis PCR Positive respiratory culture Invasive fungal infection Antibiotic treatment Antifungal treatment

64 52

(55e72) (61%)

64 63

(54e72) (60%)

0.71 0.88

50

(58%)

54

(51%)

0.38

17 14

(20%) (16%)

31 14

(30%) (13%)

0.18 0.55

7 6 62 39

(8%) (7%) (73%) (46%)

11 0 75 39

(10%) (0%) (71%) (37%)

0.63 0.007 0.87 0.24

17

(20%)

26

(25%)

0.49

7 62 24 6 2

(8%) (73%) (28%) (7%) (1e4)

17 76 41 5 3

(16%) (72%) (39%) (5%) (1e4)

0.13 1.00 0.13 0.55 0.58

6200 700 15.8 5.7

(4000e8800) (400e1100) (11.0e24.9) (2.0e9.8)

6300 700 14.9 7.1

(4400e10,200) (300e1200) (6.3e26.6) (2.6e14.1)

0.48 1.00 0.49 0.24

2.5 0.7 18

(2.9e3.3) (0.6e1.0) (14e27)

2.8 0.7 20

(2.4e3.5) (0.5e1.2) (14e30)

0.98 0.92 0.55

0.6 373

(0.4e0.8) (285e502)

0.7 303

(0.5e1.1) (236e414)

0.04 0.01

26.9 57 74

(5.7e84.6) (67%) (87%)

4.3 12 33

(4.3e4.3) (11%) (31%)

<0.001 <0.001 <0.001

19

(22%)

28

(27%)

0.51

6 52 34

(7%) (61%) (40%)

12 91 40

(11%) (87%) (38%)

0.33 <0.001 0.88

Data are presented as the numbers (%) or medians (interquartile range). Some patients had two or more underlying diseases or conditions. Ten patients with hematological malignancies had undergone stem cell transplantation. b-D-glucan values of more than 11 pg/ml (commercial cut-off value) were considered positive. PCP, Pneumocystis pneumonia; HIV, human immunodeficiency virus; SOFA, Sequential Organ Failure Assessment; PCR, polymerase chain reaction.

infections (probable pulmonary aspergillosis, n ¼ 5; candidaemia, n ¼ 1), 4 of whom were positive for b-D-glucan. Antifungal drugs were administered in 29 patients (40%): 10 (14%) as directed therapy (invasive fungal infections, n ¼ 6; oral or esophageal candidiasis, n ¼ 4), 14 (19%) as empirical therapy that was discontinued within 7 days, and 5 (7%) as prophylaxis. Antibiotics were administered in 45 patients (63%). Of these patients, 14 patients harbored a potential causative organism of bacterial pneumonia. Twelve patients were positive for b-D-glucan but did not receive anti-PCP treatments. One patient received pentamidine but discontinued at the eighth day due to adverse effect. Five patients had invasive fungal infections, other four patients were negative for PCR, and the other two patients had bacterial pneumonia which rapidly responded to antibiotics. Among the 69 patients who were positive for b-D-glucan, 1 patient receiving anti-PCP treatments (intravenous immunoglobulin

111

administration) and 2 patients without anti-PCP treatments (intravenous immunoglobulin administration or taking food containing an edible mushroom, Sparassis crispa) had a potential cause of false positive for b-D-glucan. The remaining patients did not have Pseudomonas or Alcaligenes bacteremia, had surgical gauze exposure, underwent hemodialysis using cellulose membranes, and received intravenous administration of amoxicillin-clavulanate. 3.2. Thirty-day mortality rates The 30-day mortality rate of all included patients was 21.6% (Table 2). All of the 13 definite PCP patients survived. b-D-glucanpositive patients had a significantly lower mortality rate than b-Dglucan-negative patients (10.1% vs. 28.1%, P ¼ 0.003). This difference was also observed if only patients who received anti-PCP treatment were analyzed (10.5% vs. 39.3%, P ¼ 0.003). PCRpositive patients had a significantly lower mortality rate than PCR-negative patients (15.0% vs. 30.1%, P ¼ 0.01), and this difference was also observed if only patients who received anti-PCP treatment were analyzed (12.2% vs. 72.7%, P < 0.001). In contrast, no significant difference was observed in patients without anti-PCP treatments (21.2% vs. 23.6%, P ¼ 1.00). All of the 18 patients with invasive fungal infections received adequate antifungal treatments. Of the 6 patients who received anti-PCP treatment, 1 b-D-glucan-positive patient with pulmonary aspergillosis died. Of 12 patients without anti-PCP treatments, 5 patients were positive for b-D-glucan (Fig. 1). One b-D-glucan-positive patient with candidaemia and 2 b-D-glucan-negative patients with pulmonary aspergillosis or sinus aspergillosis died. 3.3. Risk factors for 30-day mortality rates in patients receiving anti-PCP treatments without microscopic detection The results of univariate analyses of factors associated with 30-day mortality rates in 72 patients receiving anti-PCP treatments without microscopic detection are listed in Table 3. Positive b-D-glucan (odds ratio [OR] 0.24, 95% confidence interval [CI] 0.08e0.77) and positive PCR (OR 0.06, CI 0.01e0.29) results were associated with survival. Among 44 positive b-D-glucan patients, b-D-glucan levels were not associated with mortality, hypoxia, mechanical ventilation, pneumothorax, or SOFA score. Factors associated with mortality were hypoxia, mechanical ventilation, high SOFA score, low thrombocyte counts, high blood urea nitrogen level, high total bilirubin, antibiotic treatment, and antifungal treatment. The multivariate model in Table 4 shows that 30-day mortality was independently associated with positive PCR results (OR 0.14, CI 0.02e0.74) and with a high SOFA score. Positive b-D-glucan results did not reach statistical significance (OR 0.25, CI 0.06e1.02), but the model without PCR results did reach an OR of 0.21 and a CI of 0.06e 0.80 for positive b-D-glucan and an OR of 1.50 and a CI of 1.17e1.94 for the SOFA score. 4. Discussion In this prospective observational study, we first demonstrated that positive Pneumocystis PCR and b-D-glucan results were significantly associated with survival among immunocompromised patients who were suspected of having PCP but did not have definite PCP. The results remained significant after adjustment of other prognostic factors. The chest CT is a useful test for diagnosing pneumonia in immunocompromised patients. Because extensive GGO on a chest CT is the most frequent pattern in the early phase of PCP [23], we included patients presenting GGO on a chest CT, in addition to clinical suspicion of PCP due to immunocompromised status and

112

Y. Matsumura et al. / J Infect Chemother 20 (2014) 109e114

190 immunocompromised patients who had clinical presentations of PCP and GGO

85 patients with anti-PCP treatments

105 patients without anti-PCP treatments

BD+ and PCR+, n=53

BD- and PCR+, n=21

BD+ and PCR+, n=7

BD- and PCR+, n=26

36 undetermined, 13 definite PCP, 3 probable aspergillosis 1 candidaemia

16 undetermined, 2 IP, 2 probable aspergillosis, 1 RP

3 probable aspergillosis 2 BP, 1 ARDS due to

7 IP, 6 BP, 2 ARDS due to bacteraemia, 2 probable aspergillosis, 1 chronic necrotizing aspergillosis, 1 COP, 1 tuberculosis, 1 pulmonary embolism, 1 pulmonary proteinosis, 1 RP, 1 DIP, 2 undetermined

BD+ and PCR-, n=4 1 graft-versus-host disease, 1 IP

candidaemia,

1 undetermined

BD- and PCR-, n=7 2 IP, 2 DIP, 1 BP, 1 ARDS due to bacteraemia, 1 undetermined

BD+ and PCR-, n=5

BD- and PCR-, n=67

1 ARDS due to cryptococcaemia, 1 tuberculosis, 1 lymphangitis carcinomatosa, 1 PH, 1 BP

21 IP, 16 BP, 6 PH, 3 ARDS due to candidaemia,

3 DIP, 3 viral pneumonia, 2 ARDS due to bacteraemia, 2 tuberculosis, 1 congestive heart failure, 1 acute leukemia, 1 lymphangitis cartinomatosa, 1 COP, 1 RP, 6 undetermined

Fig. 1. Clinical diagnosis in 190 suspected Pneumocystis pneumonia patients. IP, interstitial pneumonia; BP, bacterial pneumonia; ARDS, acute respiratory distress syndrome due to sepsis; DIP, drug-induced pneumonia; RP, radiation pneumonitis; COP, cryptogenic organizing pneumonia; PH, pulmonary hemorrhage; PCRþ, polymerase chain reaction (PCR)positive; PCR, PCR-negative, BDþ, b-D-glucan-positive; BD, b-D-glucan-negative.

clinical symptoms, into the suspected PCP cohort. However, this CT finding has been observed in other lung diseases such as viral infection, interstitial pneumonia, and acute respiratory distress syndrome (ARDS). In our cohort, interstitial pneumonia and ARDS were determined as final diagnoses, especially in patients who did not receive anti-PCP treatment. The b-D-glucan assay was routinely performed at our institution on the same day that the plasma was obtained. As previous reports indicated that the b-D-glucan was elevated in microscopically definite PCP, we conducted a preemptive intervention of anti-PCP treatments based on positive b-D-glucan results in patients suspected of PCP. As a result, more patients who received anti-PCP treatment had positive b-D-glucan and Pneumocystis PCR results than those without anti-PCP treatments. The b-D-glucan assay is not specific for Pneumocystis and could be elevated due to other fungal infections. In previous studies of neutropenic episodes, 70e80% of patients with positive b-D-glucan results had an invasive fungal infection (aspergillosis or candidiasis), but not PCP [24,25]. In contrast, in our suspected PCP patients with positive b-D-glucan results, invasive fungal infections were found only in 4 of 57 patients (7.0%) receiving anti-PCP treatments. Among 12 patients without anti-PCP treatment, 5 patients (41.7%) had invasive fungal infections. A small number of patients with invasive fungal infections were included, probably because we enrolled patients having GGO on a chest CT. It remains controversial whether the b-D-glucan assay can be used for the assessment of severity of illness, treatment response, or mortality. However, recent studies reported no relationship between b-D-glucan levels and hypoxia [15,26,27], treatment response [15,28], or mortality [12,14,15,28,29]. Our data also cannot find relationship between severity of illness and mortality. The 30-day mortality rates of the suspected PCP patients in this study (21.6%) were similar to those in previous reports among microscopically positive PCP patients without HIV infection (20%

and 40%) [2,17]. Although a previous study did not show that b-Dglucan levels could predict mortality in definite PCP [15], all of the definite PCP patients in our study were positive for both b-D-glucan and PCR and survived under this b-D-glucan-based treatment strategy. Furthermore, b-D-glucan- or PCR-positive patients who were suspected of PCP and received anti-PCP treatments had significantly better survival rates than those with negative b-Dglucan or PCR. The results remained significant even when definite PCP patients were excluded. The univariate analyses revealed that positive b-D-glucan and positive PCR results were associated with survival in microscopically negative patients receiving anti-PCP treatments (Table 3), whereas the multivariate analysis showed that only positive PCR results independently associated with survival. This discordance may be associated with multicollinearity of PCR and b-D-glucan assays. In fact, 40 of 72 microscopically negative patients (56%) were positive

Table 2 Thirty-day mortality rates among patients who were suspected of having Pneumocystis pneumonia. Characteristics

30-day mortality rate (number of deaths/total patients) Total

All patients

PCP treatment þ

P-value

PCP treatment 

21.6% (41/190) 20.0% (17/85) 22.9% (24/105) 0.72 1.00 0.15 15.0% (16/107) 12.2% (9/74) 21.2% (7/33) 0.25 0.0% (0/13) 0.0% (0/13) e (0/0) e 17.0% (16/94) 14.8% (9/61) 21.2% (7/33) 0.57

b-D-glucan-positive 10.1% (7/69) 10.5% (6/57) 8.3% (1/12) b-D-glucan-negative 28.1% (34/121) 39.3% (11/28) 24.7% (23/93) Positive PCR Definite PCP Other than definite PCP Negative PCR

30.1% (25/83)

72.7% (8/11)

All definite PCP patients were positive by PCR. PCP, Pneumocystis pneumonia.

23.6% (17/72)

0.002

Y. Matsumura et al. / J Infect Chemother 20 (2014) 109e114

113

Table 3 Factors associated with 30-day mortality in patients receiving anti-Pneumocystis pneumonia treatments without microscopic detection of P. jirovecii by univariate analysis. Characteristics

Age, years Male sex Underlying disease Collagen vascular disease Solid malignancy Hematological malignancy Organ transplantation HIV infection Corticosteroids Immunosuppressive agents Other than corticosteroids Anti-tumor chemotherapy PCP prophylaxis Hypoxia Mechanical ventilation Pneumothorax SOFA score Laboratory findings Neutrophils/mm3 Lymphocytes/mm3 Thrombocytes/mm3 C-reactive protein, mg/dL Albumin, mg/dL Creatinine, mg/dL Blood urea nitrogen, mg/dL Total bilirubin, mg/dL Lactate dehydrogenase, IU/L b-D-glucan, pg/ml Positive b-D-glucan Positive Pneumocystis PCR Positive respiratory culture Invasive fungal infection PCP treatment by pentamidine Antibiotic treatment Antifungal treatment Adjunctive steroid therapy

Non-survivors

Survivors

(N ¼ 17)

(N ¼ 55)

OR (95% CI)

P-value

69 9

(64e73) (53%)

64 33

(55e73) (60%)

0.75 (0.25e2.24)

0.20 0.78

9 3 2 2 0 13 8

(53%) (18%) (12%) (12%) (0%) (76%) (47%)

34 13 10 5 2 41 27

(62%) (24%) (18%) (9%) (4%) (75%) (49%)

0.69 0.69 0.60 1.33 0.61 1.11 0.92

(0.23e2.08) (0.17e2.79) (0.12e3.05) (0.23e7.58) (0.03e13.36) (0.31e3.97) (0.31e2.74)

0.58 0.75 0.72 0.67 1.00 1.00 1.00

3 4 17 9 3 4

(18%) (24%) (100%) (53%) (18%) (4e8)

14 5 37 9 3 2

(25%) (9%) (67%) (16%) (5%) (1e3)

0.63 3.08 17.27 5.75 3.71

(0.16e2.51) (0.72e13.11) (0.98e303.24) (1.75e18.91) (0.67e20.45)

0.75 0.20 <0.001 0.008 0.14 <0.001

5800 500 9.0 6.6 2.6 0.8 27 0.8 419 4.3 6 9 7 1 2 15 11 15

(2800e7600) (200e1000) (3.5e19.7) (3.0e19.2) (2.4e2.9) (0.7e1.2) (18e32) (0.5e2.4) (394e486) (4.3e34.4) (35%) (53%) (41%) (6%) (12%) (88%) (65%) (88%)

6300 700 19.3 5.9 2.9 0.7 18 0.6 335 27.4 38 52 11 5 2 30 18 38

(4100e9900) (400e1100) (13.2e27.2) (2.2e9.7) (2.5e3.3) (0.5e1.2) (13e25) (0.4e0.8) (266e486) (7.5e85.6) (69%) (95%) (20%) (9%) (4%) (55%) (33%) (69%)

0.24 0.06 2.80 0.63 3.53 6.25 3.77 3.36

(0.08e0.77) (0.01e0.29) (0.87e9.02) (0.07e5.75) (0.46e27.23) (1.30e29.98) (1.20e11.82) (0.69e16.33)

0.50 0.19 0.01 0.26 0.05 0.17 0.03 0.01 0.07 0.03 0.02 <0.001 0.11 1.00 0.24 0.02 0.03 0.21

Data are presented as the numbers (%) or medians (interquartile range). Some patients had two or more underlying diseases or conditions. b-D-glucan values of more than 11 pg/ml (commercial cut-off value) were considered positive. Among 44 positive b-D-glucan patients, b-D-glucan levels were not associated with mortality (median 83.5 pg/ ml [interquartile range, 27.2e111.4] in non-survivors and 59.3 [23.1e115.2] in survivors; P ¼ 0.63), hypoxia (P ¼ 0.92), mechanical ventilation (P ¼ 0.64), pneumothorax (P ¼ 0.13), and SOFA score more than 4 points (P ¼ 0.26). PCP, Pneumocystis pneumonia; OR, odds ratio; CI, confidence interval; HIV, human immunodeficiency virus; SOFA, Sequential Organ Failure Assessment; PCR, polymerase chain reaction.

for both PCP and b-D-glucan (Fig. 1). Positive b-D-glucan did not reach statistical significance in the multivariate analysis (P ¼ 0.05, Table 4), however, positive b-D-glucan was independently associated with survival in the model without PCR results (P ¼ 0.001). As previously reported [30e33], hypoxemia, mechanical ventilation, albumin and total bilirubin were associated with mortality in this study. Even after adjustments for these confounders, positive PCR and positive b-D-glucan remained predictors of survival. The results support high diagnostic performance of these assays for microscopically negative PCP. Among 65 patients receiving antiPCP treatments except for definite PCP patients, 52 PCR- or b-Dglucan-positive patients (80%) had undetermined diagnosis (Fig. 1). Although microscopically negative or non-definite PCP has not been clearly defined [4,6], these patients might have PCP, responded to anti-PCP treatments, and subsequently survived. The PCR- or b-D-glucan-negative, non-PCP patients might have higher mortality than the microscopically negative PCP patients. We consider that the difference in mortality made PCR and b-D-glucan assays predictors of survival. To date, the diagnostic performances of PCR and b-D-glucan assays for microscopically negative PCP were studied with only patients survived with anti-PCP treatments [4,5]. This study indicates that microscopically negative PCP patients who are diagnosed by positive PCR or b-D-glucan would survive more frequently with anti-PCP treatments and these assays can detect

patients who should be considered for anti-PCP treatment with high accuracy. As PCR has a high negative predictive value for diagnosing PCP in non-HIV patients suspected of PCP and PCRnegative patients had a high mortality rate even after anti-PCP treatment in this study (Table 1), PCR-negative patients were unlikely to have PCP. This study had some limitations. First, although all of the patients in this study were suspected to have PCP, the administration of antiPCP treatments, as well as antibiotics or anti-fungal drugs, was dependent on the clinical decision, which biased the outcomes.

Table 4 Factors associated with 30-day mortality in patients receiving anti-Pneumocystis pneumonia treatments without microscopic detection of P. jirovecii by stepwise multivariate analysis. Characteristics

OR (95% CI)

P-value

SOFA score, per 1 point Pneumocystis PCR Positive b-D-glucan

1.42 (1.08e1.88) 0.14 (0.02e0.74) 0.25 (0.06e1.02)

0.01 0.02 0.05

All variables with a P value <0.10 in the univariate analyses were subjected to selection using a forward stepwise logistic procedure. The goodness of fit of the last model was evaluated by the HosmereLemeshow test (P ¼ 0.57). PCP, Pneumocystis pneumonia; OR, odds ratio; CI, confidence interval; SOFA, Sequential Organ Failure Assessment; PCR, polymerase chain reaction.

114

Y. Matsumura et al. / J Infect Chemother 20 (2014) 109e114

However, the randomized design for the administration of anti-PCP treatments seems quite difficult considering the immunocomprised status of patients at risks for PCP. With regard to treatments other than PCP, positive b-D-glucan or PCR results were good indicators for survival independent of antibiotic or antifungal use, and all of the patients with invasive fungal diseases received adequate anti-fungal therapy. Second, there was a small number of definite PCP patients in this cohort, likely because BAL was obtained by only a quarter of the patients. In conclusion, after adjustments of various prognostic factors, this study suggested that positive PCR and positive b-D-glucan results were independently associated with a reduced 30-day mortality rate among immunocompromised patients suspected of PCP without microscopic diagnosis. Despite the limitations of nonrandomized design of this study, Pneumocystis colonization in PCRpositive patients, and non-specificity of the b-D-glucan assay for Pneumocystis exists, anti-PCP treatments should be considered for patients suspected of PCP with positive PCR or b-D-glucan results to promote the survival of the patients. Conflict of interest None. Acknowledgments This study was supported by internal funding. References [1] Morris A, Lundgren JD, Masur H, Walzer PD, Hanson DL, Frederick T, et al. Current epidemiology of Pneumocystis pneumonia. Emerg Infect Dis 2004;10: 1713e20. [2] Mansharamani N, Garland R, Delaney D, Koziel H. Management and outcome patterns for adult Pneumocystis carinii pneumonia, 1985 to 1995: comparison of HIV-associated cases to other immunocompromised states. Chest 2000;118:704e11. [3] Limper A, Offord K, Smith T, Wn Martin. Pneumocystis carinii pneumonia. differences in lung parasite number and inflammation in patients with and without AIDS. Am Rev Respir Dis 1989;140:1204e9. [4] Azoulay E, Bergeron A, Chevret S, Bele N, Schlemmer B, Menotti J. Polymerase chain reaction for diagnosing Pneumocystis pneumonia in non-HIV immunocompromised patients with pulmonary infiltrates. Chest 2009;135:655e61. [5] Matsumura Y, Ito Y, Iinuma Y, Yasuma K, Yamamoto M, Matsushima A, et al. Quantitative real-time PCR and the (1/3)-b-D-glucan assay for differentiation between Pneumocystis jirovecii pneumonia and colonization. Clin Microbiol Infect 2012;18:591e7. [6] Krajicek BJ, Thomas CF, Limper AH. Pneumocystis pneumonia: current concepts in pathogenesis, diagnosis, and treatment. Clin Chest Med 2009;30: 265e78. [7] Caliendo AM, Hewitt PL, Allega JM, Keen A, Ruoff KL, Ferraro MJ. Performance of a PCR assay for detection of Pneumocystis carinii from respiratory specimens. J Clin Microbiol 1998;36:979e82. [8] Alvarez-Martínez MJ, Miró JM, Valls ME, Moreno A, Rivas PV, Solé M, et al. Sensitivity and specificity of nested and real-time PCR for the detection of Pneumocystis jiroveci in clinical specimens. Diagn Microbiol Infect Dis 2006;56:153e60. [9] Ribes JA, Limper AH, Espy MJ, Smith TF. PCR detection of Pneumocystis carinii in bronchoalveolar lavage specimens: analysis of sensitivity and specificity. J Clin Microbiol 1997;35:830e5. [10] Morris A, Wei K, Afshar K, Huang L. Epidemiology and clinical significance of Pneumocystis colonization. J Infect Dis 2008;197:10e7. [11] Alanio A, Desoubeaux G, Sarfati C, Hamane S, Bergeron A, Azoulay E, et al. Real-time PCR assay-based strategy for differentiation between active Pneumocystis jirovecii pneumonia and colonization in immunocompromised patients. Clin Microbiol Infect 2011;17:1531e7.

[12] de Boer MG, Gelinck LB, van Zelst BD, van de Sande WW, Willems LN, van Dissel JT, et al. b-D-Glucan and S-adenosylmethionine serum levels for the diagnosis of Pneumocystis pneumonia in HIV-negative patients: a prospective study. J Infect 2011;62:93e100. [13] Tasaka S, Hasegawa N, Kobayashi S, Yamada W, Nishimura T, Takeuchi T, et al. Serum indicators for the diagnosis of Pneumocystis pneumonia. Chest 2007;131:1173e80. [14] Sax PE, Komarow L, Finkelman MA, Grant PM, Andersen J, Scully E, et al. Blood (1/3)-beta-D-glucan as a diagnostic test for HIV-related Pneumocystis jirovecii pneumonia. Clin Infect Dis 2011;53:197e202. [15] Held J, Koch M, Reischl U, Danner T, Serr A. Serum (1/3)-beta-D-glucan measurement as early indicator for Pneumocystis jirovecii pneumonia and evaluation of its prognostic value. Clin Microbiol Infect 2011;17:595e602. [16] Karageorgopoulos DE, Vouloumanou EK, Ntziora F, Michalopoulos A, Rafailidis PI, Falagas ME. b-D-glucan assay for the diagnosis of invasive fungal infections: a meta-analysis. Clin Infect Dis 2011;52:750e70. [17] Bollée G, Sarfati C, Thiéry G, Bergeron A, de Miranda S, Menotti J, et al. Clinical picture of Pneumocystis jiroveci pneumonia in cancer patients. Chest 2007;132:1305e10. [18] Kelley C, Checkley W, Mannino D, Franco-Paredes C, Del Rio C, Holguin F. Trends in hospitalizations for AIDS-associated Pneumocystis jirovecii pneumonia in the United States (1986 to 2005). Chest 2009;136:190e7. [19] Radhi S, Alexander T, Ukwu M, Saleh S, Morris A. Outcome of HIV-associated Pneumocystis pneumonia in hospitalized patients from 2000 through 2003. BMC Infect Dis 2008;8:118. [20] Vincent J, de Mendonça A, Cantraine F, Moreno R, Takala J, Suter P, et al. Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: results of a multicenter, prospective study. Working group on “sepsis-related problems” of the European Society of Intensive Care Medicine. Crit Care Med 1998;26:1793e800. [21] De Pauw B, Walsh T, Donnelly J, Stevens D, Edwards J, Calandra T, et al. Revised definitions of invasive fungal disease from the European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) Consensus Group. Clin Infect Dis 2008;46:1813e21. [22] Wakefield A, Pixley F, Banerji S, Sinclair K, Miller R, Moxon E, et al. Detection of Pneumocystis carinii with DNA amplification. Lancet 1990;336:451e3. [23] Vogel MN, Vatlach M, Weissgerber P, Goeppert B, Claussen CD, Hetzel J, et al. HRCT-features of Pneumocystis jiroveci pneumonia and their evolution before and after treatment in non-HIV immunocompromised patients. Eur J Radiol 2012;81:1315e20. [24] Senn L, Robinson JO, Schmidt S, Knaup M, Asahi N, Satomura S, et al. 1,3-BetaD-glucan antigenemia for early diagnosis of invasive fungal infections in neutropenic patients with acute leukemia. Clin Infect Dis 2008;46:878e85. [25] Ellis M, Al-Ramadi B, Finkelman M, Hedstrom U, Kristensen J, Ali-Zadeh H, et al. Assessment of the clinical utility of serial beta-D-glucan concentrations in patients with persistent neutropenic fever. J Med Microbiol 2008;57:287e95. [26] Watanabe T, Yasuoka A, Tanuma J, Yazaki H, Honda H, Tsukada K, et al. Serum (1/3) beta-D-glucan as a noninvasive adjunct marker for the diagnosis of Pneumocystis pneumonia in patients with AIDS. Clin Infect Dis 2009;49:1128e31. [27] Koga M, Koibuchi T, Kikuchi T, Nakamura H, Miura T, Iwamoto A, et al. Kinetics of serum b-D-glucan after Pneumocystis pneumonia treatment in patients with AIDS. Intern Med 2011;50:1397e401. [28] Koo S, Baden LR, Marty FM. Post-diagnostic kinetics of the (1/3)-b-D-glucan assay in invasive aspergillosis, invasive candidiasis and Pneumocystis jirovecii pneumonia. Clin Microbiol Infect 2012;18:E122e7. [29] Nakamura H, Tateyama M, Tasato D, Haranaga S, Yara S, Higa F, et al. Clinical utility of serum beta-D-glucan and KL-6 levels in Pneumocystis jirovecii pneumonia. Intern Med 2009;48:195e202. [30] Matsumura Y, Shindo Y, Iinuma Y, Yamamoto M, Shirano M, Matsushima A, et al. Clinical characteristics of Pneumocystis pneumonia in non-HIV patients and prognostic factors including microbiological genotypes. BMC Infect Dis 2011;11:76. [31] Walzer PD, Evans HE, Copas AJ, Edwards SG, Grant AD, Miller RF. Early predictors of mortality from Pneumocystis jirovecii pneumonia in HIV-infected patients: 1985e2006. Clin Infect Dis 2008;46:625e33. [32] Fernandez P, Torres A, Miro JM, Vieigas C, Mallolas J, Zamora L, et al. Prognostic factors influencing the outcome in Pneumocystis carinii pneumonia in patients with AIDS. Thorax 1995;50:668e71. [33] Fei MW, Kim EJ, Sant CA, Jarlsberg LG, Davis JL, Swartzman A, et al. Predicting mortality from HIV-associated Pneumocystis pneumonia at illness presentation: an observational cohort study. Thorax 2009;64:1070e6.