INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In recipients of lung transplants who experience clinical deterioration, diagnostic uncertainty often remains despite extensive investigations including bronchoscopy, transbronchial biopsy, and open lung biopsy. The differentiation of acute rejection (AR) from infection is often difficult and patients may be unnecessarily treated for both conditions. In addition, up to 50% of lung transplant recipients develop bronchiolitis obliterans (BO), which accounts for more than 50% of deaths after 1 yr posttransplantation (1). The cause of BO is unknown, but it may represent a form of chronic rejection, resulting from repeated episodes of AR (2). There is therefore a need for new ways of monitoring the lung allograft for immunological complications.

Nitric oxide (NO) is an important mediator in the lungs and pulmonary circulation (3, 4), subserving diverse functions. NO is synthesized from L-arginine by two types of synthases. The constitutive NO synthases (cNOS) are present in many cell types (e.g., endothelial and neuronal cells), and release small amounts of NO to mediate physiological responses (e.g., vasodilatation). The inducible forms of NO synthase (iNOS) produce larger amounts of NO and are increasingly expressed in inflammatory states. Inflammatory cells such as macrophages, lymphocytes, and neutrophils can synthesize NO and also secrete cytokines such as interleukin 1 (IL-1), tumor necrosis factor  (TNF-), and interferon , which can induce iNOS expression in respiratory epithelial and endothelial cells (5).

There is increasing evidence that NO is of great relevance to organ transplantation. In animal and human studies, NO synthesis is augmented in the alloimmune response (6). In experimental graft rejection, NO synthesis is increased in activated macrophages and graft-infiltrating cells (10). In human liver (11) and cardiac transplantation (12), high levels of plasma NO metabolites (nitrite and nitrate) have been shown to correlate with episodes of AR.

NO is excreted in exhaled gas, and there is increasing interest in exhaled NO (ENO) as a noninvasive marker of airway inflammation (13). Elevated ENO has been described in bronchial asthma (13, 14), bronchiectasis (15), and viral respiratory infections in normal subjects (16). In asthma, iNOS is induced in the respiratory epithelium (5) and ENO falls after treatment with inhaled corticosteroids (17).

Published values of ENO vary greatly in asthma (13, 14). We measured ENO using a previously described validated method (18) to exclude nasal NO, which is present in high concentrations relative to the lower respiratory tract (19). In addition, the measurements were made at a fixed low expiratory flow rate (45 ml/s), as ENO is highly flow dependent (18). At this flow rate, ENO in normal subjects is up to 50 ppb, with levels up to 400 ppb in untreated bronchial asthma.

We postulated that ENO would be increased in AR owing to the intense inflammatory response, and that this increase could be detected in increased levels in exhaled gas. In addition, we also proposed that ENO might rise in bronchiolitis obliterans syndrome (BOS). If so, the measurement of ENO would provide a rapid noninvasive test to aid in the differential diagnosis of complications.

We found that ENO was indeed significantly higher in episodes of clinical AR but not elevated in bronchiolitis obliterans syndrome (BOS) or infection. ENO in stable patients was not different from that in healthy controls. Our findings suggest that ENO could help distinguish AR from infection. We conclude that ENO shows great promise as a marker of acute rejection in lung transplantation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Study of ENO

Exhaled NO was measured in ambulatory recipients of lung transplants over a 12-mo period, during visits to the weekly follow-up clinic of the Toronto Lung Transplant Program (Toronto, ON, Canada). The research protocol was approved by the Human Ethics Committee of The Toronto Hospital in accordance with the principles of the 1983 Helsinki declaration. ENO was measured only when informed consent was obtained from all participants. No exclusion criteria were applied.

Lung transplant protocols. The standard follow-up protocol involves visits once every 3 mo during the first year after transplant, once every 6 mo in the second year, and henceforth yearly, with unplanned visits in the case of complications. Patients were asked to perform an ENO measurement at each visit to the clinic.

Immunosuppression. The standard immunosuppression maintenance regimen includes prednisone (20-30 mg/day at 1-3 mo, reducing to 15 mg/d every second day at 12 mo posttransplant), cyclosporine (5 mg/kg twice daily with target blood levels by radioimmune assay between 250 and 350 ng/ml at 1-3 mo, reducing to 150-200 ng/ml at 12 mo posttransplant), and azathioprine according to leukocyte count (1- 1.5 mg/kg/d at 1-3 mo, reducing to 0.5-1.5 mg/kg/d at 12 mo posttransplant).

ENO measurement. NO was measured with a rapid linear-response chemiluminescent analyzer (270B; Sievers, Boulder, CO) with a lower limit of detection of approximately 2 ppb. Each day, a two-point calibration was performed to an NO concentration of 360 ppb, using serial dilutions of a standard NO gas (2 ppm; Matheson Gas Products, Whitby, ON, Canada). The single-breath ENO profile was displayed on a chart recorder (7046A; Hewlett-Packard, Palo Alto, CA). We measured ENO with a previously described technique (18). The patient inhaled to total lung capacity (TLC) from a clean air source (containing ~ 1 ppb NO) and exhaled at a constant low exhalation rate (45 ml/ s) via a high expiratory resistance. The patient targeted a constant positive mouth pressure of 20 mm Hg, which closed the vellum to prevent nasal NO contamination. ENO was sampled just distal to the mouthpiece. The single-breath NO profile showed a rapid rise as the dead space was washed out and reached a plateau. The ENO level was derived from a steady plateau of at least 5-s duration. Exhalations were repeated until three values of ENO that varied by < 10% were recorded. The physician measuring ENO was blinded to the clinical diagnosis at the time of the ENO measurement.

Diagnostic categories. The patients were allocated retrospectively by a physician blinded to the ENO results, according to the following clinical categories:

Stable patients.  Patients without any acute or chronic pulmonary complication at the time of any ENO measurement during the study period

Acute rejection.  Presumed acute rejection on clinical/physiological/ pathological grounds within 20 d before or after the ENO measurement and treated according to the standardized protocol (intravenous methylprednisolone, 1 g/d, for 3 d with augmented oral prednisone for 3-4 wk). We included posttreatment ENO as some patients present in AR with no recent ENO, are treated, and then attend for ENO measurement. Rejection histology was graded according to the system proposed by Yousem and coworkers (20), and all biopsies were interpreted by the same pathologist

Bronchiolitis obliterans syndrome.  Patients with a diagnosis of BOS established according to the staging system of Cooper and colleagues (21) during the study period

Unstable bronchiolitis obliterans syndrome.  Clinical and/or physiological deterioration in a patient with BOS not attributed to other causes

Infection.  A diagnosis (based on clinical [symptoms] and/or laboratory grounds [radiology, culture]) at the time of ENO measurement of acute pulmonary infection (bronchitis/bronchiolitis/pneumonia) due to any microorganism; infections were treated with antimicrobial therapy

Other complications.  Patients with other pulmonary or extrapulmonary complications or multiple complications

Each patient was classified in only one group; e.g., if a patient had AR, subsequent ENO measurements when this had resolved were not used in the analysis of stable ENO values. All group values were compared with previously established ENO values obtained in 21 healthy volunteers (12 males, 9 females; age range, 18-65 yr), using the identical ENO measurement technique.

ENO and Bronchoscopy Findings

In a separate arm of this study, we examined the correlation of ENO with findings obtained from a concurrent bronchoscopy (± 3 d). ENO was not performed on the bronchoscopy day itself. Surveillance bronchoscopies are performed once every 3 mo during the first year after transplant, once every 6 mo in the second year, and henceforth yearly, with urgent bronchoscopy in case of complications. Biopsies were classified as normal, AR grades I-III, bronchiolitis obliterans, or nonspecific inflammatory change. Bronchoalveolar lavage (BAL) culture was defined as positive or negative (any microorganism), and differential lymphocyte and neutrophil counts were recorded.

Statistical Analysis

All results are reported as mean values ± SEM. Chi-square test and ANOVA followed by the Student-Newman-Keuls multiple-comparison test were used for comparing baseline characteristics among groups, with respect to categorical and continuous variables, respectively. On post hoc testing, the distribution of ENO was not normal even after log transformation, and thus nonparametric tests were used. ANOVA on ranks followed by the Dunn's method multiple-comparison test was used for comparing ENO levels among the clinical diagnosis groups as well as among the biopsy groups. The Mann- Whitney rank sum test was used for comparing between BAL-positive and -negative culture groups. The association between ENO and differential cell counts was assessed by Spearman's rank correlation. All analyses were two-sided and a value of p  0.05 was considered significant.

The stable patient group was analyzed in order to establish normal values and the intraclass correlation coefficient. For all groups the average of all ENO values recorded during the period that a particular group definition was valid was taken as the basis for comparison.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Clinical Study of ENO

From November 1995 to December 1996, ENO was measured in 108 patients (94 double-lung and 14 single-lung transplants) with a mean (SEM) time after transplantation of 1,083.6 (± 82.6) d. In the 10 yr since the program started, more than 250 lung transplants have been performed and 125 are under active follow-up. Pretransplant diagnoses were as follows: cystic fibrosis (29), chronic obstructive pulmonary disease (COPD) with or without ₁-antitrypsin deficiency (41), idiopathic pulmonary fibrosis (15), pulmonary hypertension (10), and other diverse diseases (13). The median number of ENO measurements was 2.53/patient (range, 1-9). Six patients, four of them with AR, refused to consent for ENO measurements and were excluded.

Demographic and clinical data for the different groups are presented in Table 1. The dose of prednisone was significantly higher in the AR group, possibly because of the shorter period after transplantation and the inclusion of ENO values taken after the start of augmented prednisone therapy. The BOS patients were on a significantly lower dose of cyclosporine, with increased use of FK 506 and methotrexate.

The ENO results for the separate patient groups are presented in Figure 1. Groups with insubstantial numbers were not analyzed statistically and are described in Table 4.

View larger version (35K): [in this window] [in a new window]   Figure 1.   Exhaled nitric oxide (ENO) levels in healthy controls (n = 21), stable patients (n = 53), and patients with acute rejection (n = 8), infection (n = 14), or stable BOS (n = 25). The ENO mean ± SEM of each diagnostic category is shown under each box. *p < 0.001 compared with all other groups. Unstable BOS (n = 2) and other or multiple complications (n = 6) appear in Table 4.

The stable patient group (n = 53) showed reproducible ENO values from time to time with an intraclass correlation coefficient r1 of 0.65 for a subgroup of 34 patients with two measurements and an r1 of 0.59 for 27 of these 34 subjects with three measurements. No correlation was found between ENO and time posttransplant in this group (R² = 0.01, p = 0.5026). The ENO values in this group did not differ from ENO values previously established in our laboratory for healthy nonsmoking controls, using the described measurement technique.

The AR group showed a significantly increased ENO when compared with all other groups and with healthy controls. The ENO data and clinical and laboratory features of patients with AR are shown in Table 2. Patient 5 had a surveillance biopsy showing AR grade II in the absence of clinical symptoms or functional deterioration and was treated on this basis. The sensitivity and specificity of ENO for the diagnosis of AR was 88 and 87% (ENO > 30 ppb), 75 and 96% (ENO > 40 ppb) and 63 and 98% (ENO > 60 ppb), respectively.

All groups, other than the AR group, were indistinguishable on the basis of ENO. In the infection group, one patient had a high ENO value (Table 3). The patients with BOS had reproducible ENO levels over time with an intraclass correlation coefficient rho of 0.79 for a subgroup of 9 of the 25 BOS patients with two ENO measurements and 0.83 for 7 of these 9 patients who had three ENO measurements.

Patients with unstable BOS, other complications, or multiple complications were not statistically analyzed owing to small numbers, and are presented for information in Table 4.

Other sporadic complications associated with an elevated ENO value included two patients with an asthma-like syndrome (symptomatic bronchospasm with a marked response to bronchodilator) and one with a lymphoproliferative disease (see Table 4). No ENO (mean ± SD, ppb) difference was found between stable patients with single-lung (16.9 ± 6.5, n = 5) and double-lung transplantation (19.8 ± 8.1, n = 48, p = 0.443).

ENO and Bronchoscopy Findings

There were 99 bronchoscopic examinations with a concurrent ENO test performed within 3 d and on the day before bronchoscopy in > 95% of cases. Biopsies were taken in 80 bronchoscopies. There was no significant difference in ENO value (mean ± SEM; ppb) between patients yielding normal biopsies (n = 48; 21.2 ± 3.1), AR grade I biopsies (n = 17; 24.7 ± 5.0), and biopsies revealing nonspecific inflammatory change (n = 13; 18.8 ± 1.9). We found no cases of a biopsy showing AR grade II or III among patients performing an ENO test within 3 d of the bronchoscopy, and only two cases of BO were found that were not analyzed statistically. BAL culture results were available from 98 bronchoscopies. ENO values (mean ± SEM; ppb) associated with culture-positive samples (n = 34; 16.2 ± 1.9) were not significantly different from those associated with culture-negative samples (n = 64; 18.6 ± 1.6). Differential lymphocyte and neutrophil counts from 92 bronchoscopies [median ± SD (range)] were 7.5 ± 10.6% (0-44%) and 2.0 ± 8.5% (0-45%), respectively. There was no correlation between ENO and differential lymphocyte or neutrophil counts (Spearman rank order correlations: 0.0993, p = 0.3457; 0.0589, p = 0.5762, respectively).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We evaluated ENO as a noninvasive marker of airway inflammation in recipients of lung transplantation, for whom there is a need to develop new techniques of monitoring for complications and improving diagnostic accuracy. Acute rejection was associated with a raised ENO value, in contrast to infection and BOS. Thus ENO may aid in the distinction between infection and rejection.

ENO is increasingly recognized as a noninvasive marker of lung inflammation (13, 22) in bronchial asthma and in pulmonary infection (15, 16). However, the precise correlation of ENO with objective indices of inflammation remains to be established. Compliance with the ENO measurements was achieved in the vast majority of patients, who found the test comfortable. We used a previously validated measurement technique that excluded ENO contamination by the relatively high NO concentrations of the nasal cavity and was standardized for expiratory flow rate (18).

Stable recipients of lung transplants produced ENO values that were consistent from visit to visit as assessed by the intraclass correlation coefficient, and that did not differ from those of healthy controls (among whom ENO is felt to be a constitutive NOS product) (23). It is probable that the immunosuppressive regimen followed by these stable patients suppressed iNOS expression (24) and that ENO was a product of cNOS, which is not steroid sensitive (22). Thus this study has established ENO values (specific to our measurement technique) for stable recipients of lung transplants against which patients with complications can be referenced.

Exhaled NO was high in those AR patients with symptoms, physiological deterioration, and new radiological findings (Table 2). The unexpectedly low incidence of AR in this study may be related to the relatively long period of time elapsed between transplantation and time of study (mean ± SEM, 1,083.6 ± 82.6 d). Interestingly, ENO was in the normal range in one subject included in the AR group (Table 2, subject 5), who was treated for AR on the basis of a surveillance biopsy that showed grade II rejection, but was clinically stable. The sensitivity and specificity of ENO for AR were good. However, in view of the small number of AR cases, our results need to be interpreted with caution. The predictive value of ENO in the early diagnosis of AR remains to be established, but the time course of ENO in some patients (Table 2) suggests that ENO may rise in the latent phase before clinical AR manifests itself.

In vitro and in vivo studies of allograft rejection have shown an increase in NO production from immune cells such as macrophages and lymphocytes (6). Accordingly, the cellular sources of ENO in lung rejection may include infiltrating inflammatory cells (e.g., macrophages and lymphocytes), endothelial cells, or bronchial epithelial cells stimulated by cytokines. Worral and coworkers (25), in a study of rat lung transplantation, have demonstrated that iNOS was upregulated in mononuclear cells infiltrating the peribronchial, perivascular, and interstitial areas but not in the respiratory epithelium or endothelium. This is in contrast to bronchial asthma, a condition in which iNOS is upregulated in the respiratory epithelium (5). Lung rejection may involve the vessels, interstitium, and airways and, as NO is freely diffusible, we reasoned that the increased NO production known to accompany AR would augment ENO.

In accordance with our findings in AR are three analogous studies in which plasma NO metabolites (nitrite, nitrate) were examined in the setting of human organ transplantation. Devlin and colleagues (11) reported that plasma nitrite concentrations were the most predictive parameter of acute liver rejection, and that the levels correlated with the histological grade. Furthermore, other complications including chronic rejection showed no such rise. Similarly, Ioannidis and co-workers (9) reported that serum nitrate rose transiently immediately after liver transplantation, but fell within 3 d in patients with an uncomplicated postoperative course. This rise in nitrate was, however, augmented and prolonged in those with rejection or infection. The recurrence of inflammatory complications was associated with a recurrent nitrate peak. In addition, Benvenuti and colleagues (12) reported that serum nitrate was a good predictor of AR in human cardiac allograft rejection and that rejection grade correlated closely with the serum nitrate concentration.

In contrast to AR, infection and BOS were not associated with a rise in ENO despite the possibility of airway inflammation. In AR, by definition, the host immune response has overcome the immunosuppression. However, in patients with infection or BOS, the immunosuppressive treatment may have inhibited the inflammatory response and iNOS upregulation. In BOS, the normal ENO is probably related to the type, activity, and chronicity of the inflammation, which is distinct from AR (20). Another factor may be the more frequent use in the treatment of BOS of FK 506, which is a more powerful suppressor of NO production than corticosteroids (11). Although the two patients with unstable BOS did not have high ENO, it remains that ENO in new-onset or rapidly progressive BO could be high. Similar to our findings in BOS was the lack of any rise in plasma nitrite in chronic liver rejection reported by Devlin and co-workers (11).

In infection, the lack of ENO rise may also be due to the early institution of antimicrobial therapy, as the level of vigilance regarding infection is high in these immunosuppressed patients. It is possible that more advanced infection would be associated with an ENO rise. Despite the one infected patient with a high ENO value (Table 3), our results suggest an important utility of ENO in distinguishing between infection and rejection.

A high ENO value was observed in two patients with an asthma-like syndrome, in keeping with the high ENO reported in asthma (13, 14). In addition, ENO was raised in one patient with lymphoproliferative disease, possibly owing to a reduction in cyclosporine dose, and in another with multiple complications. Thus, although ENO was highly sensitive for acute rejection, it is not 100% specific.

The retrospective analysis of ENO and bronchoscopic findings showed that there was no variation in ENO according to histology (normal, AR grade I, nonspecific inflammation). This apparently contradicts our clinical ENO study, in which AR was associated with an ENO rise. This may be explained on the basis that grade I AR on surveillance biopsy, in the absence of any decline in symptoms or FEV₁, does not necessarily represent a significant complication as it is not associated with iNOS upregulation. Indeed, it is our usual practice not to treat such patients with pulsed corticosteroids. In contrast, two of the AR group had a raised ENO but normal biopsies. In our program, there is a scarcity of biopsies with higher grades of AR and this leaves open the possibility that ENO would rise in more advanced histological grades. In agreement with these findings, Devlin and colleagues (11) reported that in human liver transplantation, plasma nitrite was reported to be normal in grade I AR, but elevated in more advanced histological grades.

The lack of difference in ENO levels between positive and negative BAL culture groups is in accordance with the lack of ENO rise in clinical infection. Finally, the lack of correlation of ENO with differential cell counts shows that the precise link between ENO and objective indices of airway inflammation remains to be established. It is possible that the analysis of BAL cytokines would help elucidate the mechanisms involved in the induction of iNOS.

Nitric oxide may promote graft damage due to cytotoxic metabolites such as peroxynitrite (26). Thus inhibition of NO production increased graft survival in rat cardiac and lung transplantation (27). There are, however, beneficial effects of NO such as the neutralization of superoxide radicals and bacterial killing (4). In one mouse model of spontaneous hyporesponsiveness (graft tolerance), NO was produced by graft cells perhaps attenuating immune cell activation, as inhibition of NO synthesis caused AR to appear (30).

More than 90% of patients survive the lung transplant operation. However, acute complications such as rejection and infection occur in the early postoperative course and are often difficult to distinguish. Episodes of AR may be related to the development of BOS, which results in significant late morbidity and mortality. We have shown that ENO is raised in symptomatic AR as compared with all other groups, but not in BOS and infection. In addition, a biopsy finding of AR grade I in the absence of clinical findings was not associated with ENO rise. This initial study supports the utility of ENO measurements in the early detection and monitoring of immunological complications in lung transplantation and to aid in the differential diagnosis between acute rejection and infection. Further studies are required to determine the relationship between ENO and histological rejection grade, ENO in new-onset BOS, and the role of ENO in assessing the response to therapy. In addition, the precise cellular sources of ENO remain to be determined. Exhaled NO measurement may prove to be a useful tool in managing recipients of lung transplants.