Introduction

Pneumocystis jiroveci, a fungus, is the causative agent of Pneumocystis pneumonia (PCP). The diagnosis of PCP should be considered in solid organ transplant (SOT) recipients with new onset of fever, pulmonary symptoms, and hypoxemia. The contemporary incidence of this infection among SOT recipients is on average <1 %; however, the average incidence of PCP in SOT recipients prior to the implementation of prophylaxis was 10–12 % but varied depending on the on the organ type and immunosuppressive regimen [1, 2]. The introduction of PCP prophylaxis for SOT recipients has dramatically modified the epidemiological landscape of this infection. Currently, we are experiencing the era of “late PCP,” where the majority of infections are diagnosed greater than 12-months post-transplant in patients who are no longer receiving prophylaxis.

Because there is no available culture system for Pneumocystis jiroveci, the mainstay of the diagnosis of PCP has been microscopic examination of lower respiratory tract samples. Traditionally, the diagnostic yield of lower respiratory tract microscopy with staining in non-HIV patients has historically been considered lower when compared to those who are HIV-infected [2]. Newer diagnostic methods are being explored and will hopefully increase the validity of diagnosis in SOT patients.

Herein, we review the epidemiology of PCP in solid organ transplant recipients, with a focus on changing epidemiology, transmission, and diagnosis.

Epidemiology

The average incidence of PCP in SOT recipients prior to the implementation of prophylaxis was 10–12 % but varied depending on the on the organ type and immunosuppressive regimen [1, 2]. During this era, the risk of PCP in kidney transplant recipients ranged from 0.6 to 14 %; similar frequency of infection was described in liver transplant recipients (3–11 %) [3–5]. The likelihood of developing disease in heart transplant recipients was in the order of 2–41 %; whereas lung and combined heart-lung recipients had the highest risk ranging from 6.5 to 43 % [6–9]. Timing of PCP infection was highest between the second- and sixth-month post-transplantation [2]. Proposed risks factors for early-onset PCP included immunosuppressive therapies (anti-thymocyte globulin, alemtuzumab, calcineurin inhibitors), CMV infection, and allograft rejection [2, 8, 10–15].

The American Society of Transplantation (AST) Infectious Diseases Community of Practice recommends PCP prophylaxis for all solid organ transplant recipients for at least 6–12 months [16]. At completion of initial prophylaxis, the need for additional prophylaxis should be evaluated. Transplant recipients who require increased immunosuppression in the face of graft rejection may have a longer at-risk period. In addition, the need to re-initiate PCP prophylaxis should be considered in those patients who develop CMV infection or acquire other risk factors for PCP. Many transplant professionals advocate for life-long PCP prophylaxis in higher-risk organ groups such as heart, lung, and heart-lung transplant recipients [13, 16, 17].

Trimethoprim-sulfamethoxazole (TMP-SMX) is the prophylactic agent of choice for PCP, while dapsone, atovaquone, clindamycin plus pyrimethamine, and pentamidine are considered alternative or second-line agents [13, 16]. The advantages of TMP-SMX over other prophylactic agents include higher efficacy, lower cost, and breadth of coverage for other opportunistic infections. Multiple studies have demonstrated the efficacy of TMP-SMX as a prophylactic agent for PCP in SOT recipients [18–20]; breakthrough infections are rare [21]. Breakthrough infections are more common in patients taking a non-TMP-SMX agent; these breakthrough episodes tend to have atypical presentations, and microscopy on respiratory samples is often negative [13]. Due to the benefits of TMP-SMX over other agents, it is worthwhile to consider re-introduction of TMP-SMX in those SOT recipients on other prophylactic regimens in the following scenarios: no clear contraindication to TMP-SMX; an alternative explanation for the adverse effect initially attributed to TMP-SMX is identified; or other drugs that could have lead to additive toxicity have been discontinued or dose reduced.

The introduction of prophylaxis has modified the landscape of the epidemiology of PCP in SOT recipients. In order to understand recent changes, it is necessary to evaluate studies describing the epidemiology of this disease among patients transplanted after the year 2000. The Transplant Associated Infection Surveillance Network (TRANSNET), consisting of 23 US transplant centers, performed prospective surveillance for fungal infections from 2001 to 2006 [22]. The surveillance cohort, defined as prevalent and incident transplant patients from the centers who developed infection, revealed that pneumocystosis represents 0–3 % of all invasive fungal infections in SOT recipients. The incidence cohort estimated a cumulative incidence of PCP at 12 months of <0.1 %; 75, 50, and 20 % of the cases occurred after 6, 12, and 36 months, respectively [22]. A single center retrospective study from Slovenia also conducted between 2001 and 2006 described 13 (2.2 %) cases among 601 renal transplant recipients [23]. All patients received induction with basiliximab, and the protocol for PCP prophylaxis consisted of TMP-SMX (80/400 mg daily) for 12 months. The median time from transplantation to diagnosis was 17 months (range 3–148 months). Ten of the 13 cases occurred more than 12 months after transplantation; none of the 3 patients who developed PCP within the first 12 months was receiving prophylaxis. Six of the 13 cases had recent or concomitant CMV infection. The mortality among patients with PCP was 23 %, owing to bacterial, viral, or fungal super-infection.

A single center retrospective study from Korea performed between 2008 and 2009 described a 12-month incidence of PCP of 0.9 % among liver transplant recipients [24]. The protocol for PCP prophylaxis at this high-volume center (performing more than 300 liver transplants per year) was TMP/SMX 160/800 mg every other day for 6 months. The median time from transplantation to PCP was 9.5 months (range 1–67 months); half of the cases experienced a recent episode of rejection. The crude mortality in this study for patients who developed PCP was 50 %. Two other studies from Canada and Spain have also described similar findings [25, 26]. Collectively, these data suggests that the incidence of PCP among SOT recipients receiving prophylaxis is low. Most cases of PCP currently occur more than 12 months after transplant and are associated with treatment for rejection and CMV infection. These findings emphasize the need to consider re-introduction of prophylaxis for PCP among those at risk.

Transmission

During the last two decades, there has been a dramatic increase in the number of studies reporting outbreaks and clusters of infections in transplant recipients [10, 18, 27–36]. To understand the results and implications of these studies, an overview of the pathogenesis and mode of transmission of this infection is warranted. When symptomatic PCP infection occurs, more than 50 % of patients have more than one strain of the organism [37–39]. Animal and human studies favor an airborne transmission route [40]. With regards to the mode of acquisition, multiple facts point toward human-to-human transmission in cases of Pneumocystis pneumonia: (1) Pneumocystis organisms are host specific; rat, mouse, ferret, or horse species infect only their specific host and are not sources of human infection; and (2) there is no identifiable reservoir for P. jiroveci other than individuals infected with this organism. Serological data from infants indicate that most humans are infected within the first 2 to 4 years of life [41]. Given the ubiquitous nature of this early exposure, it is unlikely that the organism is transmitted only by humans with symptomatic infection. In the past, it was thought that most cases of PCP were due to reactivation. Recent data support the notion that disease can be the result of a recently acquired infection.

A systematic review of outbreaks and clusters of PCP in renal transplant recipients from 1980 to 2010 was recently published [33]. This study included a total of 16 outbreaks in renal transplant recipients, described in 15 articles. Interestingly, none of the studies reported a simultaneous increase in the incidence of pneumocystosis in immunocompromised hosts, including non-renal transplant recipients. The median number of cases was 12 per outbreak. Most outbreaks were marked by cases occurring in patients who were not receiving PCP prophylaxis. A period of low incidence of PCP (<2 %) preceding the outbreak was described in 75 % of the outbreaks and quoted as the rationale for the lack of chemoprophylaxis for this organism. Of note, 11 of the 15 studies describe a temporal association between the cessation of the outbreak and the introduction of prophylaxis with TMP-SMX. Treatment for rejection and CMV infection were risk factors for developing PCP during the outbreaks. A transmission map was frequently a part of the analysis and suggested either a common environmental source or patient-to-patient transmission. The lack of evidence for an environmental source and molecular analysis points toward patient-to-patient transmission. The use of more advanced molecular techniques, namely restriction fragment length polymorphisms (RFLP), reveals that no two cases of sporadic PCP are caused by the same stain [42]. Using RFLP, investigators were able to demonstrate that a single strain was responsible for outbreaks in two geographically distinct regions of Europe [36]. In an additional study [34], colonized individuals were purported as potential infectious sources for PCP cases. In summary, human-to-human transmission is likely to be more common than reactivation of prior infection. Although controversial, these findings provide support to the recommendations made by some experts to isolate patients with PCP from other susceptible hosts [16, 43, 44]. Without definitive data, formal recommendations regarding infection control cannot be made.

Clinical Presentation of PCP

The diagnosis of PCP should be considered in SOT recipients with new onset of fever, pulmonary symptoms, and hypoxemia. Groups at increased risk include those not receiving prophylaxis, those prescribed prophylactic regimes other than trimethoprim-sulfamethoxazole, or in those whose compliance and/or tolerance to prophylaxis has been poor [45, 46].

It is important to recognize that the clinical presentation of SOT recipients with PCP differs from that of HIV-infected patients with PCP [47–51]. Studies from the early HIV epidemic suggested a shorter duration of symptoms in HIV-negative patients with PCP when compared with those who had AIDS [47]. More contemporary studies that included and compared transplant recipients confirm these findings [48–50]. A retrospective single center study from the USA conducted from 1996 to 2008 included a total of 97 cases of confirmed PCP: 65 in HIV-positive and 32 in HIV-negative patients (19 SOT recipients) [49]. HIV-negative patients reported a shorter duration of symptoms, fewer days of dyspnea (10 ± 2 days vs. 17 ± 2 days; P = 0.02) and of fever (8 ± 2 days vs. 15 ± 4 days; P = 0.02). Otherwise, there were no statistical differences in the frequency of tachycardia, tachypnea, hypoxemia, and laboratory abnormalities including lactate dehydrogenase (354.9 ± 29.2 vs. 496 ± 50.5 U/L; P = 0.10). A prospective multicenter study from France included a total of 544 cases of confirmed PCP from 2007 to 2010, 223 HIV-positive and 321 HIV-negative [50]. The HIV-negative sub-group included 99 SOT recipients (80 kidney, 8 liver, 8 heart, 3 lung). The median time from onset of respiratory symptoms to diagnosis was significantly shorter for non-HIV patients (5 vs. 21; P < 0.0001). However, hypoxemia, intensive care admission, need for invasive or noninvasive mechanical ventilation, and shock were more common among non-HIV patients. Viral, bacterial, and fungal infections were common in the HIV-negative group, with 30.5 % having ≥1 microbial co-infection and 6 % having ≥2 microbial co-infections. These data emphasize the importance of ruling out concomitant infections in SOT recipients with PCP.

Radiography

There is a paucity of contemporary data regarding radiographic findings in SOT recipients with PCP. The most common findings on chest radiograph are diffuse bilateral alveolar or interstitial pulmonary infiltrates [49, 50]. Normal roentgenograms, consolidation, solitary, or multiple nodules with or without cavitations have also been described [52, 53]. High resolution computed tomography (HRCT) is considered more sensitive in the early phases of the infection [54]. Patchy or diffuse ground glass opacities sparing the lung periphery are the characteristic findings in HRCT; consolidations, nodules, cysts, pneumoceles, pneumothoraces, and other patterns have also been described [53]. Use of aerosolized pentamidine as prophylaxis has been associated with upper lobe infiltrates [55, 56]. In a study from France, all patients diagnosed with PCP had a chest radiograph while computed tomography scans were performed when deemed necessary by the clinicians. For the 231 non-HIV patients, the percentage of chest radiographic findings was typical (bilateral interstitial or alveolointerstitial opacities) in 77 %; atypical findings (focal consolidation, pleural effusion, subpleural nodules, and cavitations) in 15 %; pneumothorax in 2.2 %; and a normal chest radiograph in 8 %. Diffuse ground glass opacities were considered typical findings in CT scans; septal lines and centrilobular nodules were also interpreted as supportive of a PCP diagnosis.

Diagnosis

Because there is no available culture system for this organism, the mainstay of the diagnosis of PCP has been microscopic examination of lower respiratory tract samples (induced sputum, bronchoalveolar lavage, transbronchial or open lung tissue biopsies). In general, more invasive procedures produce a higher diagnostic yield [2]. Noninvasive testing of a lower respiratory sample microscopic examination should be attempted as an initial step to confirm the diagnosis. Although Pneumocystis is rarely identified in expectorated sputum, the organism can be frequently detected in sputum induced by inhalation of aerosolized hypertonic saline [16, 57]. If PCP is not identified using this diagnostic modality, then a bronchoscopy with bronchoalveolar lavage (BAL) should be performed with consideration of a transbronchial biopsy. Video-assisted thoracoscopic biopsies are often reserved for those patients in whom other diagnostic approaches have been unrevealing [58, 59].

Numerous staining methods have been utilized for microscopy of lower respiratory tract samples [2]. Direct fluorescent antibody staining using a fluorescein-conjugated monoclonal antibody has the highest sensitivity as it can identify both trophic forms and cysts and has become a common technique used [60]. The diagnostic yield of lower respiratory tract microscopy with staining in non-HIV patients has historically been considered lower when compared to those who are HIV-infected [2, 45]. However, contemporary studies challenge this notion [49, 50]. A single center retrospective study from the USA suggested similar diagnostic yields in these two groups, while the study from France showed a higher diagnostic yield when using bronchoscopy in non-HIV patients (87 vs. 97 % P = 0.0003). Despite the advances in diagnostic techniques, the exact performance in SOT recipients remains unknown.

The notion that the sensitivity of microscopy is lower in non-HIV patients, perhaps as a result of lower organism burden, has fueled the interest in other diagnostic modalities: polymerase chain reaction (PCR) and serology with serum ß-D-glucan. The studies with these diagnostic tools have the following limitations: scarcity of data specific to SOT recipients and lack specificity [16]. Using different techniques and target genes, PCR assays have been developed for the detection of PCP in induced sputum, BAL, and oral-wash samples [61–67]. A single center prospective study conducted from 2002 to 2005 enrolled 448 HIV-negative patients (21 SOT recipients) [65]. Using conventional PCR targeting the large subunit of mitochondrial RNA, the sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) were 87.2, 92.2, 51.5, and 98.7 %, respectively. The lack of specificity and low PPV result from inability of the assay to distinguish disease from colonization. A more recent study published by the same group suggests a better performance using real-time PCR; unfortunately, only 6 % (14/238) of the patients included in the study were SOT recipients [66]. Data assessing the performance of the ß-D-glucan test has similar limitations [68–71]. A study in HIV-infected patients, using a dichotomized value of 80 pg/ml showed a sensitivity, specificity, PPV, and NPV of 92, 65, 85, and 80 %, respectively (27). A retrospective study from Japan that included 295 consecutive patients (HIV-positive and HIV-negative) from 1998 to 2005 revealed a PPV and NPV of 61 and 89 % when using a cut-off level of 31.1 pg/ml [68]. A meta-analysis that included both HIV-positive and HIV-negative patients disclosed a sensitivity and specificity of 94.8 and 86.3 %, respectively; unfortunately, only 5 of the 14 studies included enrolled SOT recipients [71]. In clinical practice, the results of these two assays should be interpreted in the context of the clinical characteristics and pre-test probability. To optimize the performance of these diagnostic tests, clinicians should avoid using them indiscriminately and reserve them for patients with a compatible clinical presentation when other diagnostic tests are either not feasible or available.

Table 1 summarizes the AST recommendations for the diagnosis of PCP in SOT recipients [16]. Clinicians should consider PCP in their differential diagnosis in patients at risk as prompt diagnosis and initiation of therapy is essential to optimize outcomes [49, 50].

Table 1 Recommendations for the diagnosis of PCP in SOT recipients [16]
Full size table

Conclusions

Since the first description of Pneumocystis pneumonia more than 100 years ago, there has been remarkable advancement in our understanding of the epidemiology, clinical manifestations, and diagnostic armamentarium. PCP remains an important opportunistic infection among transplant recipients, and the introduction of prophylaxis has dramatically modified the epidemiological landscape of this infection; we are currently experiencing the era of late PCP. Current studies emphasize the need to consider prolonging the prophylaxis course and re-introducing prophylaxis among patients who remain at risk for PCP. Although molecular diagnostic methods have improved our understanding of the epidemiology of PCP, there is a dearth of diagnostic testing data in SOT recipients; traditionally, sensitivity of testing is lower when compared to HIV-infected patients. Although there is a low incidence of PCP in the SOT population, associated morbidity and mortality may be high, necessitating additional studies aimed at diagnosis.