This Article Abstract Full Text (PDF) All Versions of this Article: 200309-1322SOv1 170/1/22    most recent Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kotloff, R. M. Articles by Crawford, S. W. Search for Related Content PubMed PubMed Citation Articles by Kotloff, R. M. Articles by Crawford, S. W.

Pulmonary Complications of Solid Organ and Hematopoietic Stem Cell Transplantation

Section of Advanced Lung Disease and Lung Transplantation, Pulmonary, Allergy, and Critical Care Division, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania; and Pulmonary and Critical Care Division, Naval Medical Center, San Diego, California

Correspondence and requests for reprints should be addressed to Robert M. Kotloff, M.D., 838 West Gates, University of Pennsylvania Medical Center, 3400 Spruce Street, Philadelphia, PA 19027. E-mail: kotloff{at}mail.med.upenn.edu

	ABSTRACT

TOP ABSTRACT CONTENTS SOLID ORGAN TRANSPLANTATION HEMATOPOIETIC STEM CELL... CONCLUSIONS REFERENCES

 
The ability to successfully transplant solid organs and hematopoietic stem cells represents one of the landmark medical achievements of the twentieth century. Solid organ transplantation has emerged as the standard of care for select patients with severe vital organ dysfunction and hematopoietic stem cell transplantation has become an important treatment option for patients with a wide spectrum of nonmalignant and malignant hematologic disorders, genetic disorders, and solid tumors. Although advances in surgical techniques, immunosuppressive management, and prophylaxis and treatment of infectious diseases have made long-term survival an achievable goal, transplant recipients remain at high risk for developing a myriad of serious and often life-threatening complications. Paramount among these are pulmonary complications, which arise as a consequence of the immunosuppressed status of the recipient as well as from such factors as the initial surgical insult of organ transplantation, the chemotherapy and radiation conditioning regimens that precede hematopoietic stem cell transplantation, and alloimmune mechanisms mediating host-versus-graft and graft-versus-host responses. As the population of transplant recipients continues to grow and as their care progressively shifts from the university hospital to the community setting, knowledge of the pulmonary complications of transplantation is increasingly germane to the contemporary practice of pulmonary medicine.

Key Words: hematopoietic stem cell transplantation • organ transplantation • pneumonia

	CONTENTS

TOP ABSTRACT CONTENTS SOLID ORGAN TRANSPLANTATION HEMATOPOIETIC STEM CELL... CONCLUSIONS REFERENCES

 
Solid Organ Transplantation

Infectious Pulmonary Complications

Noninfectious Pulmonary Complications

Hematopoietic Stem Cell Transplantation

Overview

General Risk Factors for Pulmonary Complications

Infectious Pulmonary Complications

Noninfectious Pulmonary Complications

Conclusions

Solid organ transplantation and hematopoietic stem cell (HSC) transplantation emerged in the 1960s as novel therapeutic approaches to human disease. Propelled by advances in basic immunobiology and clinical care, solid organ transplantation is now widely applied in the treatment of vital organ dysfunction, as is HSC transplantation in the treatment of malignant, hematologic, autoimmune, and genetic diseases. Although offering extended survival to many patients with otherwise lethal conditions, these techniques remain fraught with risk. Pulmonary complications, both infectious and noninfectious, are among the most commonly encountered hazards and they contribute significantly to morbidity and mortality. Factors enhancing the risk of pulmonary complications include the immunosuppressed status of the transplant recipient, the surgical techniques used to carry out organ transplantation, and the chemoradiation conditioning regimens employed in HSC transplantation.

The topic of pulmonary complications of transplantation was previously the subject of a State of the Art review in 1991 (1–3). Advances in the diagnosis, treatment, and prevention of these complications, as well as a greater appreciation for their prevalence and natural history, have prompted this update. The past decade has witnessed a continued proliferation of transplant centers, a steady increase in the number of procedures performed, and a progressive shift in the care of recipients from the university hospitals to the community setting. Once esoteric, knowledge of the pulmonary complications of transplantation has become increasingly relevant to the contemporary practice of pulmonary medicine.

The pulmonary complications encountered among solid organ and HSC transplant populations are sufficiently distinct in spectrum, presentation, and clinical course to warrant that they be addressed in separate sections of this review. The section on solid organ transplantation focuses on experience with the four most commonly performed procedures—heart, kidney, lung (including heart–lung), and liver transplantation. Complications limited to specific organ recipient populations are noted in the subheadings.

	SOLID ORGAN TRANSPLANTATION

TOP ABSTRACT CONTENTS SOLID ORGAN TRANSPLANTATION HEMATOPOIETIC STEM CELL... CONCLUSIONS REFERENCES

 
Infectious Pulmonary Complications
Although the incidence of infectious complications after organ transplantation has declined with the introduction of more effective prophylactic strategies and with refinements in immunosuppressive regimens, infection remains a common life-threatening complication faced by these patients. The lungs are particularly vulnerable, representing the leading infectious site in lung and heart transplant recipients (4–6) and the second most common site (after intraabdominal infection) in liver transplant recipients (7). The incidence of pulmonary infection is lowest in kidney transplant recipients, reflecting the less rigorous surgical procedure required to implant the allograft and the decreased level of immunosuppression required to maintain it.

The spectrum of microorganisms responsible for posttransplantation infections is similar among the various solid organ transplantation populations and is discussed in detail in the sections to follow. The sequence in which these different organisms appear in the posttransplantation course is fairly characteristic (Figure 1) (8). The first month is influenced predominantly by infectious risks posed by surgery and intensive care, and to a lesser extent by the initiation of immunosuppressive agents. Nosocomial bacterial infections predominate, similar to that of the general surgical population. The second stage extends from 1 to 6 months, a period of maximum sustained immunosuppression that is characterized by the emergence of opportunistic pathogens. Beyond 6 months, allograft function in the majority of patients is sufficiently stable to permit reduction in the level of immunosuppression. Consequently, infections are largely due to common community-acquired pathogens. Opportunistic infections occur less frequently in this later period but remain especially prevalent among the subset of patients requiring augmentation of immunosuppression for treatment of chronic rejection or recurrent episodes of acute rejection.

View larger version (24K): [in this window] [in a new window]   Figure 1. Timeline of infectious complications (pulmonary and nonpulmonary) after solid organ transplantation. CMV = cytomegalovirus; EBV = Epstein–Barr virus; HSV = herpes simplex virus; PTLD = posttransplant lymphoproliferative disorder; RSV = respiratory syncytial virus; VZV = varicella zoster virus. Reprinted by permission from Fishman and Rubin (N Engl J Med 1998;338:1741–1751) (8).

Community-acquired bacterial pneumonia occurs later in the posttransplantation period. Haemophilus influenzae, Streptococcus pneumoniae, and Legionella species are among the commonly identified organisms (5, 10, 15). Response to therapy is generally excellent, with reported mortality rates of 0–33% (5, 9, 15). Lower respiratory tract infections occurring in the later phases posttransplantation are particularly prevalent among lung transplant recipients who have developed bronchiolitis obliterans syndrome (BOS) (6). These patients typically present with recurrent episodes of purulent bronchitis and pneumonia. Bronchiectasis can be documented on high-resolution computed tomography (CT) in up to one-third of these patients (16, 17). Pseudomonas aeruginosa is identified as the etiologic agent in the majority of cases (6).

In the early era of organ transplantation, Nocardia infections were relatively common, with a prevalence of 2–13% documented in several large series (6, 18, 19). More recent case series have suggested a lower frequency of infection, on the order of 0.2–2.1% (5, 20, 21). This trend has been attributed to the introduction of cyclosporine-based immunosuppressive regimens that have permitted the use of reduced doses of corticosteroids and, more recently, to the widespread use of sulfonamides for Pneumocystis carinii pneumonia (PCP) prophylaxis (5, 11). Nonetheless, clinicians must remain particularly vigilant for this infection in patients in whom trimethoprim–sulfamethoxazole has either not been administered because of allergy or has been discontinued after the first year. Infection due to this aerobic, gram-positive filamentous rod is most common beyond the first month after transplantation. Patients may be asymptomatic or may present in a subacute fashion with fever, nonproductive cough, pleuritic chest pain, dyspnea, hemoptysis, and weight loss. Dissemination to brain, skin, and soft tissue occurs in up to one-third of infected patients. Chest radiographs and CT scans typically demonstrate one or several nodules that may be cavitary (22, 23). Sulfonamides are the treatment of choice; minocycline, amikacin, imipenem, and ceftriaxone are alternative agents for sulfa-allergic patients. Treatment for at least 3 months is recommended for pulmonary infection, and for up to 12 months for disseminated disease. Mortality directly attributable to Nocardia infection ranges from 0 to 30% among the various solid organ transplant populations (11, 20, 21, 24, 25).

Tuberculosis.
Tuberculosis has been reported in about 0.5–2% of organ transplant recipients in the United States and Europe (26–28), but in up to 15% of recipients in endemic areas such as India (29). Although a relatively uncommon posttransplantation infection in developed countries, the annualized rate of infection is 30- to 100-fold higher than that of the general population (26–28).

Reactivation of latent infection is believed to be the predominant mechanism for development of active tuberculosis after transplantation, but this assumption is difficult to verify on the basis of available data. In published series, information about pretransplantation PPD status is available for fewer than half of organ transplant patients who developed active tuberculosis after transplantation. Among this subgroup, only 22% demonstrate a positive response, although one must assume that there is a high false negative skin test rate due to multiple factors present in patients with chronic organ dysfunction (26). In addition, only 5% of transplant recipients with tuberculosis have a pretransplantation history of infection and only 12% demonstrate chest radiographic evidence of prior infection (26). Less common modes of acquisition include nosocomial outbreaks and donor transmission through infected kidney, lung, and liver allografts (26).

An analysis by Singh and Paterson of the published literature, encompassing a total of 511 cases of tuberculosis between 1967 and 1997, provides important insight into the clinical presentation of this infection in solid organ transplant recipients (26). Onset of infection occurred a median of 9 months after transplantation and nearly two-thirds of cases developed within the first year. Fifty-one percent of patients had tuberculosis restricted to the lungs, 16% had focal extrapulmonary infection, and 33% had disseminated disease. Counting both patients with pulmonary tuberculosis and those with lung involvement as part of disseminated disease, the overall rate of lung involvement was 71%. Fever was the most common presenting symptom, occurring in more than 90% of patients with disseminated disease but in only 66% of those with pulmonary tuberculosis. Chest radiographic abnormalities included focal infiltrates in 40%, a miliary pattern in 22%, pleural effusions in 13%, diffuse interstitial infiltrates in 5%, and cavitary lung disease in only 4%.

Treatment of active tuberculosis in organ transplant recipients involves the use of combination therapy as per standard guidelines for the general population. Administration of this regimen to transplant recipients can be challenging, however. The risk of isoniazid-induced hepatotoxicity is markedly enhanced in liver transplant recipients, necessitating discontinuation of this agent in 41–83% of patients (26, 30). Other recipient populations tolerate this agent better, with reported discontinuation rates of less than 5% (26). Administration of rifampin, a potent inducer of the hepatic P-450 microsomal enzyme system, dramatically increases the clearance of cyclosporine and tacrolimus, consequently lowering blood levels of these drugs and enhancing the risk of rejection (31–33). Concurrent use of rifampin mandates frequent monitoring of immunosuppressive drug levels and appropriate adjustment in dosing to maintain therapeutic levels. Because of difficulties in maintaining therapeutic drug levels, some clinicians have advocated avoidance of rifampin in favor of alternative agents (31).

Even in contemporary series, mortality among transplant recipients with tuberculosis remains in the range of 25–40% (26, 31). This high mortality rate reflects not only the direct consequences of the infection but also the impact of enhanced rejection and graft loss among treated but suboptimally immunosuppressed patients and the adverse impact of comorbid conditions that may have contributed to development of tuberculosis. For patients who complete a full course of treatment, response is highly favorable and mortality rates are low (31).

Given the significant morbidity and mortality associated with active tuberculosis in the transplant population, routine skin testing and preemptive treatment of latent infection are recommended (34). Ideally, skin testing should be performed before transplantation, when there is a greater likelihood that the latently infected patient will mount a delayed hypersensitivity response to antigenic challenge. Because waiting times for transplantation often exceed 1 year in all but the most urgent circumstances, the recommended 9-month course of isoniazid instituted at the time of listing can often be completed before transplantation. In this way, interactions with immunosuppressive drugs that could lead to an enhanced risk of hepatotoxicity can be avoided. The approach to liver transplant candidates and liver transplant recipients with latent infection is more problematic. There is an understandable reluctance to use isoniazid before transplantation because of the presence of severe liver disease. However, the feasibility of administering isoniazid to liver transplant candidates was demonstrated in a series of 18 patients, all of whom tolerated a complete course without a significant change in hepatic function (35). The use of isoniazid after liver transplantation is associated with an increased risk of hepatotoxicity, but the risk may be lower when the drug is administered as a single agent rather than as part of a multidrug antituberculous regimen (36). If treatment of latent infection with isoniazid is attempted after liver transplantation, liver biopsy should be considered in the setting of elevated liver enzymes, because etiologies other than drug toxicity (e.g., allograft rejection) may be responsible in up to half of these instances (26). A 4-month course of rifampin is an acceptable alternative to isoniazid and is ideally initiated before transplantation to avoid interactions with the calcineurin inhibitors. The other approved regimen for treatment of latent infection—a 2-month course of rifampin and pyrazinamide—has been associated with frequent and severe hepatotoxicity (37). For this reason, this regimen is absolutely contraindicated in patients with chronic liver disease awaiting transplantation and must be used with extreme caution in liver transplant recipients.

Nontuberculous mycobacterial infection.
Among lung transplant recipients, the nontuberculous mycobacteria may be more common than Mycobacterium tuberculosis as a cause of pulmonary infections. In the largest published series encompassing 261 patients, pulmonary infection due to nontuberculous mycobacteria was documented in 16 patients (6.1%) compared with only 2 cases of pulmonary tuberculosis (0.8%) (38). Thirteen of the 16 cases were due to Mycobacterium avium complex; M. kansasii, M. abscessus, and M. asiaticum each accounted for one case. Pulmonary infection tended to occur late in the posttransplantation course and was associated with preexistent chronic rejection in more than half the cases. Treatment resulted in clinical improvement in approximately half of treated patients and there were no deaths directly attributable to these infections.

Pulmonary infection due to nontuberculous mycobacterial species is considerably less common in other solid organ transplant populations. Published series identify pulmonary involvement in only 8 of 502 heart transplant recipients (1.6%) (39) and 5 of 3,763 renal transplant recipients (0.1%) (40, 41). Only single case reports document this complication after liver transplantation (42). The prevailing pathogens responsible for pulmonary infection in these populations are M. kansasii and M. avium complex (39, 40, 43).

Cytomegalovirus.
Cytomegalovirus (CMV) is the most common viral pathogen encountered in all solid organ recipient populations. Infection can occur by transfer of virus with the allograft or by reactivation of latent virus remotely acquired by the recipient. Seronegative recipients who acquire organs from seropositive donors are at greatest risk for developing infection, and these primary infections tend to be the most severe. The use of antilymphocyte antibody therapy for immunosuppression also enhances the likelihood and severity of infection in susceptible recipients.

CMV infection typically emerges 1 to 3 months after transplantation, although onset may be delayed in patients receiving prophylaxis. Infection is often subclinical, manifested as asymptomatic viremia or shedding of virus in the respiratory tract or urine. Clinical disease can assume a number of forms including a mononucleosis-like "CMV syndrome" with fever, malaise, and leukopenia; and organ-specific involvement of the lungs, gastrointestinal tract, liver, myocardium, and central nervous system. In addition to the direct impact of tissue invasion in producing symptoms and dysfunction of involved organs, CMV infection appears to enhance the overall level of host immunosuppression, possibly accounting for the frequent emergence of other opportunistic infections in its wake. More insidiously, CMV infection has also been linked to the subsequent development of chronic allograft dysfunction and graft loss (44, 45).

Contemporary series employing a variety of antiviral prophylactic strategies document an incidence of CMV pneumonitis of 0–9.2% among liver transplant recipients (46–49), 0.8–6.6% among heart transplant recipients (5, 9, 50, 51), and less than 1% among renal transplant recipients (52, 53). In contrast, an incidence of 15–55% has been reported after lung transplantation (54–58). The higher frequency of CMV pneumonitis in the lung transplant population is consistent with the notion that the lung is a major site of CMV latency and, therefore, that large quantities of CMV can be transmitted in the allograft (59). The high incidence likely also reflects the widespread performance of surveillance bronchoscopy in this population, facilitating detection of subclinical disease. Indeed, surveillance bronchoscopy has demonstrated that about 10–15% of cases of CMV pneumonitis are asymptomatic (56, 60). For the majority of transplant recipients, a prodrome of fever, malaise, and myalgias frequently precedes the onset of pneumonitis, which is heralded by nonproductive cough and dyspnea. Associated laboratory findings of leukopenia, thrombocytopenia, and elevated liver transaminases provide important clues to the presence of CMV. Radiographically, CMV pneumonitis is associated with a number of nonspecific findings including ground glass opacities, airspace consolidation, and nodules (61, 62).

A diagnosis of CMV pneumonitis is unequivocally established only by demonstration of characteristic viral inclusions on histologic or cytologic specimens. Unfortunately, the yield of transbronchial lung biopsy and bronchoalveolar lavage in demonstrating these changes is relatively low and surgical lung biopsy, while definitive, is invasive. This has prompted the development of alternative diagnostic techniques but the clinical limitations of these newer tools must be appreciated. Use of the rapid shell vial culture of bronchoalveolar lavage fluid is extremely efficient in detecting virus in the lung. However, because shedding of virus into the respiratory tract can occur in the absence of tissue invasion, a positive culture in the appropriate clinical setting at best provides only a presumptive diagnosis. Precise quantification of viral load in peripheral blood can now be achieved by means of the pp65 antigenemia assay and by polymerase chain reaction (PCR) techniques. Data on the role of these assays as surrogate diagnostic markers of invasive disease are limited (46, 54). In one preliminary study, Sanchez and colleagues demonstrated that the blood viral load measured by PCR was significantly higher in lung transplant recipients with biopsy-proven CMV pneumonitis compared with those without (54). Nonetheless, using a threshold optimized by receiver operating curve analysis, the positive predictive value of this technique in diagnosing CMV pneumonitis was only 40%, although the negative predictive value was 89%. There is a larger body of literature on the use of peripheral blood viral load quantification not as a diagnostic tool per se, but as a means of identifying patients most likely to progress from subclinical infection to clinical disease (46, 48, 50, 63). These studies suggest that a high or rapidly increasing peripheral blood viral load is a sensitive but nonspecific marker of imminent progression to symptomatic disease, with a positive predictive value of only 49–64% and a negative predictive value in excess of 95%. In the final analysis, viral load quantification is most appropriately employed in targeting patients for preemptive prophylaxis (see below), but its role in diagnosing established or incipient CMV disease (including pneumonitis) is limited.

The efficacy of ganciclovir in the treatment of CMV disease in various solid organ transplant populations has been suggested in uncontrolled clinical series and has become the standard of care (64). In renal transplant recipients, for example, ganciclovir reduced the overall mortality associated with CMV pneumonitis from 50 to 20% and mortality in the subset requiring mechanical ventilation from more than 90 to 60% (1). Standard treatment consists of a 2- to 3-week course of intravenous ganciclovir at a dose of 5 mg/kg twice daily, adjusted for renal insufficiency. Some advocate the addition of CMV hyperimmune globulin in treatment of severe disease, but evidence supporting this practice is scant (65). Although treatment is effective, relapse rates of up to 60% in primary infection and 20% in previously exposed recipients have been reported (66). Administration of a 3-month course of oral ganciclovir after definitive intravenous therapy modestly reduces the risk of relapse (67).

In an attempt to minimize the adverse impact of CMV infection on the posttransplantation course, emphasis has shifted to preventive strategies. Numerous prospective, randomized trials, summarized in several reviews and in one metaanalysis, have documented the efficacy of antiviral prophylaxis in diminishing the risk of CMV infection and invasive disease (45, 68, 69). Ganciclovir, administered orally or intravenously, has been the drug most commonly employed for the various solid organ recipient populations; valacyclovir is an alternative agent whose benefits have been demonstrated principally after kidney transplantation (53). In current practice, oral valganciclovir is emerging as an attractive option due to its excellent bioavailability after oral administration and once-daily prophylactic dosing requirement. Retrospective data as well as a recent prospective, randomized trial suggest that this agent is at least as effective as oral ganciclovir in the prevention of CMV disease in solid organ recipients (52, 70).

Universal prophylaxis of all donor-seropositive/recipient-seronegative patients is recommended as the risk of CMV disease is high (66, 71). Because the risk of disease is significantly lower in seropositive recipients (independent of donor status), it has been argued that universal prophylaxis of this group leads to overtreatment, increasing costs, and unduly exposing patients to the risk of drug toxicity. In this population, preemptive strategies targeting antiviral therapy exclusively to patients demonstrating a rising viral load in peripheral blood are being developed (66, 72).

Emergence of ganciclovir-resistant strains of CMV threatens to unravel the therapeutic gains that have been achieved in the treatment of this infection. Analysis of CMV isolates from various solid organ transplant populations at Loyola University Medical Center (Maywood, IL) documented ganciclovir resistance in 15.2% of lung, 5.3% of heart, 5.6% of liver, and 2.2% of kidney recipients (73). Across all populations, donor-positive/recipient-negative patients are at greatest risk of developing resistance, likely because of the high viral load associated with primary infection (74, 75). Other risk factors that have been identified include the use of potent immunosuppressive agents such as antilymphocyte antibodies and daclizumab, and prolonged exposure to ganciclovir (74, 75). There is particular concern that the low serum levels achieved with use of oral ganciclovir may predispose to development of resistance (8). Foscarnet is the agent of choice for treatment of ganciclovir-resistant disease, but its use may be limited by associated nephrotoxicity. The presence of ganciclovir-resistant disease is associated with decreased survival in lung transplant recipients (75, 76).

Community respiratory viruses.
Infections due to the community respiratory viruses—influenza, parainfluenza, adenovirus, and respiratory syncytial virus (RSV)—are common in the general population, typically presenting as mild, self-limited upper respiratory tract illnesses. It is unclear whether the risk of acquiring these viral infections is increased among solid organ transplant recipients, but there is a greater propensity for these pathogens to involve the lower respiratory tract and, therefore, to result in a more severe spectrum of illness (77, 78). Of the various organ transplant populations, the highest rates of infection have been reported in lung transplant recipients, up to 21% of whom develop respiratory viral infections (77–80). The ostensibly higher rate of infection documented after lung transplantation must be interpreted with caution as it could reflect enhanced detection due to the frequent performance of bronchoscopy rather than a true increase in susceptibility.

RSV and influenza virus infections occur seasonally, with epidemics in the winter and spring months, whereas adenovirus and parainfluenza infections are seen throughout the year. Patients with lower respiratory tract involvement, in the form of bronchiolitis or pneumonitis, typically present with fever, dyspnea, cough, and wheezing (78, 79). Chest radiographs may be normal or may show only subtle interstitial changes (77). CT of the chest is more sensitive and findings of ground glass, airspace consolidation, nodules, and tree-in-bud opacities are seen (Figure 2) (81). The diagnosis rests on demonstration of virus in respiratory secretions obtained by nasopharyngeal swabbing, nasal wash, or bronchoalveolar lavage. Viral culture represents the "gold standard" for diagnosis but typically entails a period of 3–14 days before results are available. Rapid diagnostic tests utilizing enzyme-linked immunosorbent assays or immunofluorescence techniques to identify viral antigens are now available, but results should be corroborated by standard viral culture.

View larger version (126K): [in this window] [in a new window]   Figure 2. Respiratory syncytial virus (RSV) infection. (A) High-resolution computed tomography (CT) demonstrating diffuse centrilobular nodules, peripheral tree-in-bud opacities, and focal bronchial wall thickening and dilatation consistent with bronchiolitis. The patient presented with wheezing and dyspnea several months after heart transplantation and had RSV recovered from bronchoalveolar lavage specimens. (B) Standard CT image of another transplant recipient with RSV pneumonia, demonstrating diffuse ground glass opacities and mild reticulation.

Treatment options are limited and largely supportive. The efficacy of antiviral therapy has been convincingly demonstrated only in the treatment of influenza in immunocompetent hosts. In this context, initiation of amantadine or rimantadine within 48 hours of symptom onset shortens the severity and duration of illness due to influenza A (87). Similar results are achieved with the early administration of the neuraminidase inhibitors zanamivir and oseltamivir, which have the added advantage of efficacy against both influenza A and B (88–90). Studies confirming benefit of these agents in organ transplant recipients are notably lacking. There is no established treatment for infections due to the other respiratory viruses although aerosolized ribavirin has been advocated as treatment for RSV and parainfluenza virus infections on the basis of anecdotal reports of success (82, 84, 91).

The limited treatment options have led to an increased focus on prevention. Respiratory viruses are highly contagious and infection control measures that emphasize minimizing contact with infected individuals and frequent hand washing may reduce risk. The inactivated influenza vaccine should be administered to all transplant recipients and close contacts. Although the antibody response appears to be attenuated in the solid organ transplant recipient, many are still able to mount protective antibody responses (92, 93). Case reports suggesting that vaccination may augment the alloimmune response and increase the risk of allograft rejection have not been confirmed in studies and should not prevent routine vaccination (94, 95). Chemoprophylaxis of transplant recipients with amantadine, ramantadine, or one of the neuraminidase inhibitors should be considered in the setting of a major influenza outbreak or after exposure to a sick contact with presumed influenza (96, 97).

Aspergillosis.
Although a number of opportunistic and endemic fungi have been reported to cause pulmonary infections in organ transplant recipients, Aspergillus species are by far the most frequent and lethal fungal pathogens encountered. The incidence of invasive aspergillosis approximates 5% among the liver, heart, and lung transplant populations, but occurs considerably less frequently after kidney transplantation (5, 9, 15, 98, 99). Invasive disease is most commonly diagnosed within the first 6 months and nearly always involves the lung. Dissemination to distant sites, particularly the brain, occurs in a considerable minority of patients (98, 99). Symptoms of invasive aspergillosis are nonspecific and include fever, cough, pleuritic chest pain, and hemoptysis. Radiographically, pulmonary aspergillosis may appear as single or multiple nodular opacities, cavities, or alveolar consolidation. The halo sign, considered a highly characteristic radiographic feature of invasive aspergillosis in the HSC transplant population (see below), is infrequently encountered and considerably less specific in the solid organ transplant populations (61, 100, 101).

Diagnosis of invasive aspergillosis can be problematic. Aspergillus is cultured from sputum in only 8–34% and from bronchoalveolar lavage fluid in 45–62% of patients with invasive disease (98). Conversely, the false positive rate associated with recovery of Aspergillus from respiratory tract cultures ranges from 28 to 55% in organ transplant recipients, with the highest rates of airway colonization seen after lung transplantation (99, 102). In the context of compatible clinical and radiographic features and/or demonstration of Aspergillus in respiratory secretions by culture or cytology, the clinician must exercise judgment in deciding whether to initiate an empiric trial of antifungal therapy or pursue more definitive proof by means of transthoracic needle biopsy or surgical lung biopsy.

Amphotericin B has traditionally been the mainstay of therapy for invasive aspergillosis, but its many side effects and need for intravenous administration have made it a rather unappealing choice. More recently, liposomal amphotericin preparations have been introduced that are less nephrotoxic than the parent compound, an important feature for a drug used concurrently with calcineurin inhibitors (cyclosporine and tacrolimus). The triazoles itraconazole and voriconazole offer the advantages of availability in both oral and intravenous formulations and absence of nephrotoxicity. Voriconazole absorption from the gastrointestinal tract is much more reliable than that of itraconazole, making it the preferred oral agent. Voriconazole was shown to have superior efficacy and less toxicity than amphotericin B in the treatment of invasive aspergillosis, and it may soon emerge as first-line therapy (103). Both of the triazoles are potent inhibitors of the P-450 hepatic enzyme system and can lead to dangerously high blood levels of cyclosporine and tacrolimus if appropriate adjustments in the dosing of these calcineurin inhibitors are not made. Because of particularly profound interactions, the concurrent use of voriconazole and sirolimus is contraindicated. Caspofungin, the first of a new class of echinocandins, has been approved in the United States for treatment of invasive aspergillosis in patients who fail or are intolerant of first-line therapy. Despite the availability of antifungal therapy, mortality in published series is in the range of 30–90%, with the highest mortality rates associated with disseminated disease (11, 98, 99, 103). The therapeutic role of surgical resection remains uncertain, but surgery has been advocated in cases of localized pulmonary infection refractory to medical therapy (104–106).

Endobronchial aspergillosis is uniquely encountered in the lung transplant population, with an observed frequency of about 5% (99). In most cases, infection is localized to the bronchial anastomosis, where devitalized cartilage and foreign suture material create a nurturing environment. Less commonly, infection may present as a more diffuse ulcerative bronchitis with formation of pseudomembranes, typically following in the wake of a severe ischemic injury to the bronchial mucosa. Clustered within the first 6 months posttransplantation, these airway infections are usually asymptomatic and detected only by surveillance bronchoscopy. Although usually responsive to oral itraconazole or to inhaled or intravenous amphotericin, airway infections have rarely progressed to invasive pneumonia or have resulted in fatal erosion into the adjacent pulmonary artery (99, 107). An increased risk of subsequent bronchial stenosis or bronchomalacia has also been reported, but it is unclear whether this is a consequence of the infection or of an underlying ischemic injury to the bronchus that predisposed to infection (108, 109).

Pneumocystis carinii pneumonia.
Before the widespread introduction of chemoprophylaxis, Pneumocystis carinii pneumonia (PCP) was observed to be a common opportunistic infection among solid organ transplant recipients. As documented in older series, organ-specific prevalence rates for at-risk patients (i.e., not receiving prophylaxis) were 4% for kidney and heart transplant recipients, 11% for liver transplant recipients, and up to 33% for heart–lung recipients (110). With the administration of low-dose trimethoprim–sulfamethoxazole or an alternative prophylactic agent, PCP can be effectively prevented. In a large contemporary series from the Cleveland Clinic (Cleveland, OH), PCP was documented in only 25 of 1,299 solid organ transplant recipients and it occurred exclusively in patients who were not receiving prophylaxis (110). The greatest risk of PCP falls between the second and sixth posttransplantation months. The risk declines significantly beyond the first year for all groups except lung transplant recipients (110), prompting many nonlung programs to discontinue prophylaxis beyond this point. Likely because of the need for augmented immunosuppression, patients with refractory acute allograft rejection or with chronic allograft rejection appear to be at increased risk for the late development of PCP and continuation or resumption of PCP prophylaxis may be warranted under these circumstances (110). Indefinite prophylaxis is advocated for lung transplant recipients.

PCP in solid organ transplant recipients typically presents in a subacute fashion, with a mean duration of symptoms before diagnosis of 14 days (110). Dyspnea, fever, and cough are the most common presenting symptoms (110). Radiographic abnormalities are typically bilateral and may appear as interstitial, alveolar, or ground glass opacities, the latter best demonstrated on CT scan. The diagnosis can be established by bronchoalveolar lavage alone in about 90% of cases; performance of transbronchial biopsies modestly enhances the yield (110). High-dose trimethoprim-sulfamethoxazole represents first-line therapy, with intravenous pentamidine typically reserved for patients who fail or cannot tolerate trimethoprim-sulfamethoxazole. Mortality directly attributable to PCP was 18% (5 of 28) in the Cleveland Clinic series (110).

Miscellaneous fungal infections.
Although Aspergillus species account for the majority of invasive fungal infections of the lung in solid organ transplant recipients, a myriad of other fungal organisms can cause pulmonary disease. These include Cryptococcus neoformans, the zygomycetes (especially Mucor species), and the geographically restricted endemic fungi (Histoplasma capsulatum, Coccidioides immitus, and Blastomyces dermatitidis) (111–114). In general, the presenting features of these fungal infections are neither clinically nor radiographically distinct and all have the potential to cause disseminated disease in addition to pneumonia. Candida species are responsible for a number of serious posttransplant infections, including sepsis, intraabdominal abscesses, and urinary tract infections, but involvement of the lungs is conspicuously rare. The one notable exception to this is infection of the bronchial anastomosis that occasionally occurs after lung transplantation (108, 115). An emerging pulmonary pathogen is Scedosporium apiospermum (previously known as Pseudallescheria boydii). This organism has been documented to cause infection in all of the major solid organ transplant populations (116). Invasive pulmonary disease is a feature in about 50% of cases; central nervous system and endovascular involvement as well as widespread dissemination are also common (116). In distinction to the other fungal pathogens discussed above, Scedosporium is frequently resistant to amphotericin B; voriconazole appears to be the preferred agent.

Noninfectious Pulmonary Complications
Perioperative complications: liver transplantation.
Liver transplantation entails the performance of extensive upper abdominal surgery on a population of patients considered by conventional standards to be poor surgical candidates due to malnutrition, debility, and critical illness. Not surprisingly, prolonged ventilatory support is a relatively common feature of the early postoperative period. In a survey of 546 liver transplant patients, Glanemann and colleagues found that 11% of patients required mechanical ventilatory support beyond 24 hours (117). Risk factors for prolonged ventilator support included acute liver failure before transplantation, severe postoperative graft dysfunction, and retransplantation. All patients were eventually extubated but 36% of those who had been ventilated for more than 24 hours and 12% of those ventilated for less than 24 hours required reintubation. The most common indications for reintubation were pneumonia, encephalopathy, and surgical bleeding (118). The need for reintubation was associated with significantly poorer survival.

Acute respiratory distress syndrome is a highly lethal cause of postoperative respiratory failure after liver transplantation. The reported incidence ranges from 4.5 to 15.7%, with a mortality rate approximating 80% (119). Sepsis is the most common risk factor, but other potential risk factors include massive blood transfusions, transfusion-related acute lung injury, aspiration, and the use of OKT3 antilymphocyte therapy.

Perioperative pleural effusions are present in the majority of liver transplant recipients (120). Effusions are transudative and are either right-sided or bilateral, but are never exclusively on the left. Disruption of diaphragmatic lymphatics during hepatectomy is postulated to be the principal mechanism of fluid accumulation (120). Effusions may enlarge over the first postoperative week but typically resolve by the third week. Approximately 10% of patients require therapeutic drainage because of respiratory compromise (120). Effusions that continue to enlarge beyond the first week, and isolated left-sided effusions, should be sampled to rule out other causes.

Right-sided diaphragmatic dysfunction is an underappreciated perioperative complication of liver transplantation that is presumed to result from crush injury to the right phrenic nerve by the suprahepatic vena caval clamp placed during surgery (121). A prospective study of 48 liver transplant recipients detected evidence of delayed or absent right-sided phrenic nerve conduction by transcutaneous electrophysiologic testing in 79% of patients whereas the left phrenic was normal in all cases. In 38% of patients, there was associated right diaphragmatic paralysis documented by ultrasound. Phrenic nerve injury was not associated with increased duration of mechanical ventilatory support or hospital stay. In a subset of patients monitored by serial testing, abnormalities in phrenic nerve conduction and diaphragmatic excursion normalized by 9 months after surgery.

Perioperative complications: lung transplantation.
Mild, transient pulmonary edema ("reimplantation response," "reperfusion pulmonary edema") is a nearly universal feature of the freshly transplanted lung allograft. It is presumed to be a consequence of ischemia–reperfusion injury and the attendant increase in microvascular permeability, but surgical trauma and lymphatic disruption may be contributing factors. In about 12–22% of cases, injury is sufficiently severe to cause a form of acute respiratory distress syndrome termed "primary graft failure" (122–124). Risk factors for development of primary graft failure that have been preliminarily identified include donor female sex, donor African-American ethnicity, donor age (less than 21 or more than 45 years), and a recipient diagnosis of primary pulmonary hypertension (122). An association between prolonged graft ischemic time and primary graft failure has been observed in some studies (125) but not in others (124, 126). The diagnosis of primary graft failure rests on the development of widespread radiographic infiltrates and markedly impaired oxygenation within the initial 72 hours of transplantation, and the exclusion of other causes of early graft dysfunction such as volume overload, pneumonia, rejection, atelectasis, and pulmonary venous outflow obstruction. Histologic examination of lung tissue from patients with primary graft failure reveals a prevailing pattern of diffuse alveolar damage (126). Treatment is supportive, relying on conventional mechanical ventilation utilizing low tidal volume strategies, as well as on such adjunct measures as independent lung ventilation and extracorporeal life support for select patients who otherwise cannot be stabilized (127, 128). The use of nitric oxide in patients with established graft injury has been associated with sustained reduction in pulmonary artery pressures and improvement in oxygenation (129, 130). However, the prophylactic administration of nitric oxide to all recipients at the time of reperfusion does not reduce the incidence of primary graft failure (123). With an associated in-hospital mortality rate of 42–60% (124, 126), primary graft failure is the leading cause of perioperative deaths among transplant recipients (131). Recovery among survivors is often protracted and incomplete, although attainment of normal lung function and exercise tolerance is possible. Results of emergent retransplantation in this setting have been poor (132).

Major dehiscence of the bronchial anastomosis was once a leading cause of perioperative mortality, but refinements in surgical technique, tissue preservation, and immunosuppression have reduced the risk of life-threatening dehiscence to a negligible level. Focal dehiscence, incidentally detected on bronchoscopy or heralded by the appearance of a spontaneous pneumothorax, is still encountered in 1–6% of patients (133–135). Tube thoracostomy may be required for evacuation of pneumothoraces, but focal dehiscence usually heals without surgical intervention. Currently, the most common form of airway complication is anastomotic narrowing, with a reported frequency of 12–24% in contemporary series (133, 136). Some centers have reported a higher rate of anastomotic stenosis associated with the use of the telescoped anastomosis compared with the end-to-end technique (137, 138). Narrowing of the anastomosis can result from ischemia-induced stricture or bronchomalacia, or from the formation of excessive granulation tissue. Independent of the mechanism, anastomotic narrowing typically develops within several weeks of transplantation. Clues to its presence include focal wheezing on the involved side, recurrent bouts of pneumonia or purulent bronchitis, and suboptimal pulmonary function studies demonstrating airflow obstruction. Diagnosis is based on direct bronchoscopic visualization. Techniques employed to address anastomotic narrowing include balloon dilatation, laser debridement, and stent placement, all readily performed through a flexible bronchoscope (133). Endobronchial brachytherapy has been successfully employed as an adjunct in managing recurrent episodes of stenosis due to hyperplastic granulation tissue (139).

Acute hyperinflation of the native lung leading to respiratory and hemodynamic compromise in the immediate postoperative period has been reported in 15–30% of patients with emphysema undergoing single-lung transplantation (140, 141). Although risk factors remain poorly defined, the combination of positive pressure ventilation and significant allograft edema serves to magnify the compliance differential between the two lungs and may predispose to this complication. Acute hyperinflation can be rapidly addressed by initiation of independent lung ventilation, ventilating the native lung with a low respiratory rate and prolonged expiratory time to facilitate complete emptying. Beyond the perioperative period, some single-lung transplant recipients with underlying emphysema demonstrate exaggerated or progressive native lung hyperinflation that more insidiously compromises the function of the allograft. In this setting, surgical volume reduction of the native lung can result in significant functional improvement (142).

Phrenic nerve injury after lung transplantation can result from intraoperative traction or from division of the nerve in the setting of extensive fibrous adhesions and difficult hilar dissection, In two prospective studies, phrenic nerve injury was documented in 7.4 and 29.6% of patients (143, 144). The significantly higher frequency found in the latter study likely relates to the more sensitive screening technique utilized, involving phrenic nerve electrophysiologic testing in addition to radiographic visualization of diaphragmatic motion. The presence of phrenic nerve injury may prolong the stay in the intensive care unit but does not typically compromise long-term outcomes (144, 145).

Perioperative complications: heart transplantation.
Heart transplant recipients are subject to the same generic perioperative pulmonary complications encountered in the general cardiac surgical population. These include atelectasis, pulmonary edema, pleural effusions, and mediastinitis. Diaphragmatic dysfunction was documented in 12% of heart transplant recipients in one small series and was predominantly right-sided (143).

Perioperative complications: kidney transplantation.
Kidney transplantation is performed with relatively few perioperative pulmonary complications, reflecting the use of a lower abdominal incision and the comparative good health of the recipients (1, 146). The vast majority of patients are extubated in the operating room. The most common noninfectious pulmonary complication is pulmonary edema due to impaired salt and water excretion in the setting of early allograft dysfunction or rejection. There appears to be an increased incidence of thromboembolic events, possibly related in part to surgical manipulation of the pelvic veins (1).

Neoplastic disorders.
"Posttransplant lymphoproliferative disorder" (PTLD) is a term applied to a spectrum of abnormal B cell proliferative responses ranging from benign polyclonal hyperplasia to more commonly encountered malignant lymphomas. Epstein–Barr virus (EBV) has been identified as the stimulus for B cell proliferation, which proceeds in an unchecked fashion because of the muted cytotoxic T cell response in the immunosuppressed host. EBV-naive recipients who acquire primary infection at the time of organ transplantation are at greatest risk of developing PTLD (147, 148). A higher intensity of immunosuppression and, in particular, the use of antilymphocyte antibody preparations have also been implicated as risk factors (149). Likely reflecting differences in the magnitude of immunosuppression employed, the incidence of PTLD is only 1–2% among kidney and liver transplant recipients, but is in the range of 5–7% among recipients of hearts and lungs (150, 151). The incidence is greatest within the first year posttransplantation (150, 151). Lung transplant recipients and heart transplant recipients are the most likely to present with intrathoracic involvement, which typically assumes the form of one or multiple pulmonary nodules occasionally accompanied by regional adenopathy or pleural effusions (Figure 3) .

View larger version (126K): [in this window] [in a new window]   Figure 3. Posttransplantation lymphoproliferative disorder. Standard CT image of a lung transplant recipient showing multifocal nodular opacities. Surgical lung biopsy confirmed a diagnosis of posttransplant lymphoproliferative disorder.

The published literature is somewhat ambiguous with respect to prognosis. In a study that encompassed several different organ transplant populations, 45% of patients died during a median follow-up period of approximately 3 years, but it was not stated whether all deaths were a direct consequence of PTLD (152). In multivariate analysis, increased age, elevated lactate dehydrogenase levels, severe organ dysfunction, presence of B symptoms (fever, night sweats, and weight loss), and multiorgan involvement were independent markers of poor prognosis, but the type of transplant was not. A series restricted to lung transplant recipients documented a 37% mortality rate directly attributable to PTLD over a median observation period of 3 years (151). Patients with extrapulmonary involvement had the poorest prognosis. Neither of these two studies included a comparison group of patients without PTLD. In contrast, a study from Stanford composed chiefly of heart transplant recipients demonstrated no difference in survival between patients with and without PTLD (150). To date, there are insufficient data to ascertain the impact of the various treatment regimens on survival.

Other malignancies occasionally present in the lungs after transplantation. Bronchogenic carcinoma has been reported in 1.6–4.1% of heart transplant recipients (155). It may also arise in the native lung of single-lung transplant recipients with underlying chronic obstructive pulmonary disease or pulmonary fibrosis, the vast majority of whom are also smokers. The reported incidence of lung cancer after lung transplantation is 2.0–3.7% for patients with chronic obstructive lung disease and 3.4–4.0% for the pulmonary fibrosis population (156, 157). It is unclear whether the reported frequency of lung cancer in the heart and lung populations represents an increased risk or simply reflects the expected occurrence rate in a population with a high prevalence of cigarette smoking. Among liver transplant recipients with a pretransplantation history of hepatocellular carcinoma, the lung is the most common site of recurrence (22). Recurrence usually occurs within 2 years of transplantation and appears radiographically as single or multiple lung nodules. An elevated -fetoprotein level provides an important clue to the possibility of recurrent disease (22).

Pulmonary metastatic calcification: kidney and liver transplantation.
Deposition of calcium in the lung parenchyma and other organs is a well documented complication of chronic renal failure and is postulated to relate to alterations in calcium and phosphate balance and parathyroid hormone secretion (158). Although renal transplantation should have a mitigating effect, pulmonary metastatic calcification rarely may progress despite successful transplantation (159, 160) and may accelerate in association with graft failure (161, 162). Pulmonary metastatic calcification has also been described in liver transplant recipients, with a reported incidence in two series of 5.2 and 47% (163, 164). Renal insufficiency was a common but not universal feature of these patients. Secondary hyperparathyroidism, due to transient hypocalcemia induced by large-volume infusion of citrate-containing blood products and, when present, to renal insufficiency, has been offered as a possible mechanism (164).

Metastatic pulmonary calcification is most often clinically silent but rarely may lead to restrictive lung disease or fulminant respiratory failure (161, 162, 165). The major import of this disorder lies in its ability to radiographically mimic more ominous processes such as infection or malignancy. Single or multiple nodular opacities or areas of alveolar consolidation are seen on plain chest radiographs (Figure 4A) , but calcification of these lesions may not be apparent. The demonstration of high attenuation (greater than 100 HU) parenchymal opacities by CT scan (Figure 4B) or of increased uptake of tracer in the lung by technetium bone scintigraphy is helpful in establishing a diagnosis (158). In instances of diagnostic uncertainty, transbronchial or surgical lung biopsy may be necessary. There is no established treatment for pulmonary metastatic calcification but, given the overall favorable prognosis, this is rarely a consideration.

View larger version (110K): [in this window] [in a new window]   Figure 4. Metastatic calcification. (A) Chest radiograph demonstrating multifocal nodular opacities in an asymptomatic liver transplant recipient. (B) Standard CT image of the same patient demonstrating three nodules, each containing areas of high attenuation consistent with calcification.

As a consequence of its inhibitory effects on fibroblast proliferation, sirolimus has the potential to impair wound healing. Fueling concerns about this is a report of an unusually high incidence of serious bronchial anastomotic dehiscence in a small series of lung transplant recipients who received sirolimus immediately postoperatively (171). Pending further study, sirolimus should be used with caution before complete healing of the airway has occurred.

Hepatopulmonary syndrome and portopulmonary hypertension: liver transplantation.
Hepatopulmonary syndrome and portopulmonary hypertension represent two unusual pulmonary complications of advanced liver disease. Although their onset precedes liver transplantation, these disorders do not immediately or invariably correct after this intervention and they can therefore contribute significantly to posttransplantation morbidity and mortality.

Hepatopulmonary syndrome is defined as the triad of liver disease, arterial hypoxemia, and abnormal intrapulmonary vascular dilatation. The vast majority of cases involve diffuse dilatation of the pulmonary microvasculature at the precapillary and capillary level (Type 1 pattern). Rarely, discrete macroscopic arteriovenous communications are present (Type 2 pattern). Proof of the presence of intrapulmonary vascular abnormalities is provided by demonstration of systemic uptake of technetium-labeled macroaggregated albumin on standard radionuclide perfusion imaging or the delayed appearance of bubbles in the left atrium by contrast echocardiography. Diffuse microvascular dilatation is postulated to cause hypoxemia by increasing the distance through which oxygen must diffuse and thereby creating a central stream of inadequately oxygenated red cells. Unlike a true shunt, this process can be partially corrected with administration of 100% oxygen, which serves to increase the pressure gradient favoring transfer of oxygen from the alveolus to the bloodstream. An unusual feature of hepatopulmonary syndrome is the tendency for oxygenation to worsen in the erect as opposed to supine position (orthodeoxia), presumably due to basilar predominance of the vascular abnormalities.

Although hepatopulmonary syndrome was once considered an absolute contraindication to liver transplantation, subsequent demonstration of its resolution after transplantation has led to the reversal of this stance. Nonetheless, because hypoxemia often does not correct immediately and may in fact dramatically worsen in the early postoperative period, excessive morbidity and mortality have been observed among these patients. In a review of 13 published reports encompassing 81 patients with hepatopulmonary syndrome who underwent liver transplantation, refractory hypoxemia contributed significantly to the observed 16% perioperative mortality, and 21% of survivors required prolonged mechanical ventilatory support (172). A prospective study of 24 patients with hepatopulmonary syndrome undergoing liver transplantation at two experienced centers found a 1-year survival of 71%, considerably below the overall 1-year survival of 90% achieved by these centers. Six of the seven patients with hepatopulmonary syndrome who died had persistent postoperative hypoxemia necessitating mechanical ventilation. The vast majority of patients who survive to hospital discharge ultimately achieve normal room air oxygen levels, but the time to resolution of hypoxemia is often measured in months. Late recurrence of hepatopulmonary syndrome coincident with deteriorating allograft function has been reported (173).

Attempts to identify factors predictive of posttransplantation outcome in patients with hepatopulmonary syndrome have yielded conflicting results. Arguedas and colleagues found that a preoperative arterial oxygen tension on room air of less than or equal to 50 mm Hg and an extrapulmonary shunt fraction of greater than or equal to 20% (calculated on the basis of uptake of tracer in the brain on macroaggregated albumin scintigraphy) were both strong predictors of mortality, with positive predictive values of 67 and 64% and negative predictive values of 93 and 100%, respectively (174). In contrast, Taille and colleagues were unable to discern any significant relationship between these parameters and postoperative mortality, although they did find a correlation between the magnitude of preoperative hypoxemia and the length of time to achieve a normal oxygen level posttransplantation (175).

Postoperative care of the patient with persistent hypoxemia due to hepatopulmonary syndrome centers on the administration of supplemental oxygen while awaiting resolution of the disorder. On the basis of the positional nature of hypoxemia discussed above, one group reported improvement in oxygenation in a patient with use of the Trendelenburg position (176). With the demonstration that nitric oxide is an important mediator of the abnormal vascular dilatation that characterizes hepatopulmonary syndrome, attention has turned to the potential therapeutic role of inhibitors of nitric oxide in treating hypoxemia (177). Schenk and colleagues described significant improvement in oxygenation and shunt fraction for up to 10 hours after administration of a single dose of methylene blue to seven patients with advanced cirrhosis and hepatopulmonary syndrome (178). Brussino and coworkers administered nebulized N^G-nitro-L-arginine methyl ester—an inhibitor of nitric oxide synthesis—to a single patient with hepatopulmonary syndrome and demonstrated a marked rise in arterial oxygen tension and a dramatic reduction in the magnitude of bubbles appearing in the left atrium on contrast echocardiography (179). Although these preliminary observations are promising, additional corroborating studies are necessary before the widespread clinical use of these agents can be recommended. In the rare instances in which discrete arteriovenous malformations are present (Type 2 pattern), coil embolization can be an effective treatment strategy (180).

Portopulmonary hypertension describes the development of pulmonary hypertension in patients with advanced liver disease and portal hypertension. The diagnosis rests on demonstration of a mean pulmonary artery pressure exceeding 25 mm Hg and a normal pulmonary capillary wedge pressure. Some authors also include the criterion of a pulmonary vascular resistance exceeding 120 dyn/second per cm^–5 to distinguish this syndrome from the more benign and common finding of elevated pulmonary pressures due solely to increased cardiac output. Histologically, the abnormalities in the pulmonary vascular bed are identical to those seen in primary pulmonary hypertension: medial hypertrophy, intimal fibrosis, and plexiform lesions. Although a mechanistic link has yet to be defined, the observation that the prevalence of pulmonary hypertension in patients with liver disease exceeds that of the general population suggests that there is a connection. Portopulmonary hypertension is encountered in 1–2% of patients with chronic liver disease and in up to 12.5% of patients referred for liver transplant evaluation (181).

Although mild pulmonary hypertension does not appear to adversely impact liver transplantation, more significant elevations in pressure are associated with excessive posttransplantation mortality. Data derived from a large retrospective review and from a multicenter liver transplant database document mortality rates of 60–100% for liver transplant recipients with a preoperative mean pulmonary artery pressure of or exceeding 50 mm Hg, 35–40% for those with mean pressures between 35 and 50 mm Hg, and 0–17% for those whose mean pulmonary artery pressure is below 35 mm Hg. Most deaths occur intraoperatively or in the immediate postoperative period and are largely attributable to progressive right heart failure and hemodynamic collapse. Although the experience with severe portopulmonary hypertension has generally been poor, there are rare reports of improvement in pulmonary hemodynamics after transplantation (181, 182).

The demonstration that intravenous prostacyclin (epoprostenol) significantly improves pulmonary hemodynamics in patients with portopulmonary hypertension provides a strategy for optimizing the condition of some patients who would otherwise be excluded from consideration for transplantation (183). Several case reports document successful transplantation of patients with severe portopulmonary hypertension who responded favorably to the preoperative initiation of prostacyclin (184, 185). In addition, this therapy has been successfully employed in the posttransplantation treatment of progressive or recurrent pulmonary hypertension (182, 186).

Pulmonary allograft rejection: lung transplantation.
Among the solid organ transplant populations, lung transplant recipients are uniquely predisposed to pulmonary injury mediated by alloreactive T lymphocytes in the form of acute allograft rejection. Approximately 55–75% of transplant recipients experience at least one episode of acute rejection within the first year (56, 187). Beyond this period, the frequency of acute rejection declines to a low level (55, 56). Factors that influence the likelihood of developing acute rejection remain poorly defined. The degree of HLA discordance between donor and recipient has been inconsistently identified as a risk factor (188, 189). One study suggests that polymorphisms in toll-like receptor 4 that downregulate recipient innate immune responsiveness may be associated with a lower incidence of acute rejection (188).

Symptoms of acute rejection are nonspecific and include malaise, low-grade fever, dyspnea, and cough. Radiographic infiltrates, a decline in arterial oxygenation at rest or with exercise, and an abrupt fall of greater than 10% in spirometric values are important clues to the possible presence of rejection, but similar findings accompany infectious episodes. Published series employing surveillance lung biopsies have demonstrated that histologically significant acute rejection (i.e., Grade A2 or higher) may be clinically silent in 14–40% of patients (55, 56, 60, 190).

Transbronchial lung biopsy has emerged as the gold standard for diagnosis of acute rejection. The procedure is safe, can be performed in serial fashion over time, and has a sensitivity of 61–94% and specificity in excess of 90% for the diagnosis of acute rejection (191). The histologic hallmark of acute rejection is the presence of perivascular lymphocytic infiltrates that, in more severe cases, spill over into the adjacent interstitium and alveolar airspaces. Lymphocytic bronchitis or bronchiolitis may accompany the parenchymal involvement or may be an independent feature. A histologic classification system has been universally adopted to categorize and grade the severity of acute cellular rejection (192).

Conventional treatment of acute rejection consists of a 3-day pulse of high-dose intravenous methylprednisolone. In most cases, this results in rapid improvement in symptoms, pulmonary function, and radiographic abnormalities, but follow-up biopsies show histologic evidence of persistent rejection in 30% of patients with prior mild acute rejection and 44% of patients with prior moderate acute rejection (190). A transient increase in oral prednisone and subsequent tapering over several weeks is advocated by some clinicians after the intravenous steroid pulse or as primary treatment for histologically mild episodes.

Chronic rejection, in the form of bronchiolitis obliterans, develops in up to two-thirds of lung transplant recipients and represents the major impediment to long-term graft and patient survival (193–195). Bronchiolitis obliterans is a fibroproliferative process characterized by submucosal inflammation and fibrosis of the bronchiolar walls, ultimately leading to complete obliteration of the airway lumen. The functional consequence of this process is progressive and irreversible airflow obstruction. Because the characteristic histology is difficult to demonstrate by transbronchial lung biopsy, a surrogate diagnostic schema based on the magnitude of decline in FEV₁ has evolved, termed "bronchiolitis obliterans syndrome" (BOS) (196).

Acute rejection, particularly when recurrent or severe, and lymphocytic bronchiolitis have been consistently identified as major risk factors for development of BOS, supporting the view that BOS is a consequence of alloimmune injury (193–195, 197–199). Nonimmune factors may also be important in initiating or perpetuating injury, but have been more variably substantiated. These factors include CMV and other respiratory viral infections, a synergistic interaction between older donor age and prolonged ischemic times, and gastroesophageal reflux with occult aspiration (196, 199).

The incidence of BOS is greatest within the first 2 years, but the risk remains considerable and steady beyond this time point (197). Onset of disease is typically insidious but may be abrupt in more aggressive cases. Dyspnea, cough, and recurrent bouts of purulent tracheobronchitis, with recovery of Pseudomonas aeruginosa from sputum cultures, are highly characteristic features. Although chest radiographs are usually unremarkable, high-resolution computed tomography reveals evidence of air trapping on expiratory images in the majority of patients and evidence of bronchiectasis in some (16, 17, 200). Progressive airflow obstruction is the rule, although the pace of decline is highly variable and the course may be interrupted by periods of functional stability. The prognosis is generally poor, with a 40% mortality rate within 2 years of onset (201). Patients with early onset of BOS (i.e., within the first 3 years) appear to experience more rapid decline in lung function and higher mortality (202).

A myriad of immunosuppressive modalities have been employed in the treatment of BOS, including pharmacologic agents, anti-lymphocyte antibodies, photopheresis, and total lymphoid irradiation, but consensus is lacking on the optimal approach (201, 203–207). At best, immunosuppressive measures appear to slow the rate of decline rather than to fully arrest or reverse the process. A potential role for macrolide therapy is suggested by a pilot study documenting improvement in the FEV₁ in five of six patients with BOS treated with azithromycin (208). Macrolides possess antiinflammatory and antipseudomonal properties, either or both of which could account for the observed beneficial effect. The only definitive treatment is retransplantation, but this strategy remains highly controversial in the context of a scarce donor organ pool (209).

The development of strategies to prevent BOS is an area of intense interest but, to date, little substantive progress. In recognition of the established link between acute rejection and BOS, many transplant centers routinely perform surveillance lung biopsies to detect and treat clinically silent acute rejection, but the impact of this strategy on risk of BOS remains uncertain (210, 211). Data from an uncontrolled, retrospective cohort study suggest that the use of statins may be associated with a reduction in the number and severity of acute rejection episodes and a lower incidence of BOS (212).

	HEMATOPOIETIC STEM CELL TRANSPLANTATION

TOP ABSTRACT CONTENTS SOLID ORGAN TRANSPLANTATION HEMATOPOIETIC STEM CELL... CONCLUSIONS REFERENCES

 
Overview
The term "hematopoietic stem cell transplantation" has supplanted the previously employed term "bone marrow transplantation" to reflect the broader range of donor stem cell sources that are now available: bone marrow, fetal cord blood, and growth factor–stimulated peripheral blood. HSC transplantation is further defined by whether the source of stem cells is from the patient himself (autologous), an identical twin (syngeneic), or a nonidentical sibling or unrelated individual (allogeneic). After allogeneic HSC transplantation, the disparities in match between the donor graft and the recipient for the human leukocyte antigens (HLA) mediate graft-versus-host disease (GVHD) and graft rejection (host-versus-graft reactions). Although these reactions are generally detrimental to the recipient, a graft-versus-malignancy effect is thought to improve antileukemic responses (213–215).

Before stem cell infusion, high doses of chemotherapy with or without total body irradiation are typically administered to ablate the bone marrow, maximize tumor cell kill, and, in the case of allogeneic transplantation, induce immunosuppression to prevent rejection of the donor stem cells. This conditioning regimen is similar for both autologous and allogeneic transplantation and is likely responsible for many of the pulmonary complications seen after transplantation. In the allogeneic setting, immunologic injury mediated by host-reactive donor immune cells (i.e., GVHD) may lead to additional pulmonary complications. Attempts to limit the toxicity of the conditioning regimen have led investigators to develop reduced intensity nonmyeloablative regimens for allogeneic transplantation. This strategy depends on the induction of GVHD and the associated graft-versus-malignancy effect, rather than the tumoricidal effects of conventional high-dose chemotherapy and radiation, to eliminate residual tumor. Nonmyeloablative HSC transplantation may have its greatest application among older patients and those with multiple medical problems who may not tolerate more intensive conditioning regimens (213–217).

Traditionally, the main indication for HSC transplantation has been in the treatment of hematologic malignancies and solid tumors, for which the technique serves as a "rescue" therapy to restore marrow function after lethal, marrow-ablating doses of radiation and chemotherapy employed to eradicate malignant cells. More recently, the technique has been used to replace dysfunctional hematopoietic or lymphoreticular precursors involved in nonmalignant disorders such as aplastic anemia, thalassemia, and congenital immune deficiency syndromes. Currently, HSC transplantation is also undergoing study as a treatment for immune-mediated diseases such as systemic sclerosis, systemic lupus erythematosus, and multiple sclerosis, for which immune reconstitution may be beneficial (218, 219).

General Risk Factors for Pulmonary Complications
Pulmonary complications, both infectious and noninfectious, occur in up to 60% of HSC transplant recipients and nearly one-third of recipients require intensive care after transplantation (220, 221). In one respect, the HSC transplantation recipient may present a less confusing clinical diagnostic picture than that of other immunocompromised patients because specific complications tend to occur within well defined time periods (Figure 5) (220). The timing and intensity of cytoreductive therapies, the pattern of immune reconstitution that follows, and the use of prophylactic strategies for infectious agents influence the duration of these intervals (222).

View larger version (23K): [in this window] [in a new window]   Figure 5. Timeline of pulmonary complications occurring after hematopoietic stem cell transplantation. ARDS = acute respiratory distress syndrome; BMT = bone marrow transplantation (i.e., hematopoietic stem cell transplantation); BO =