The Role of C4d Deposition in the Diagnosis of Antibody-Mediated Rejection after Lung Transplantation

Abstract

Antibody-mediated rejection (AMR) is an increasingly recognized form of lung rejection. C4d deposition has been an inconsistent finding in previous reports and its role in the diagnosis has been controversial. We conducted a retrospective single-center study to characterize cases of C4d-negative probable AMR and to compare these to cases of definite (C4d-positive) AMR. We identified 73 cases of AMR: 28 (38%) were C4d-positive and 45 (62%) were C4d-negative. The two groups had a similar clinical presentation, and although more patients in the C4d-positive group had neutrophilic capillaritis (54% vs. 29%, p = 0.035), there was no significant difference in the presence of other histologic findings. In spite of aggressive antibody-depleting therapy, 19 of 73 (26%) patients in the overall cohort died within 30 days, but there was no significant difference in freedom from chronic lung allograft dysfunction (CLAD) or survival between the two groups. We conclude that AMR may cause allograft failure, but the diagnosis requires a multidisciplinary approach and a high index of suspicion. C4d deposition does not appear to be a necessary criterion for the diagnosis, and although some cases may initially respond to therapy, there is a high incidence of CLAD and poor survival after AMR.

INTRODUCTION

Lung transplantation is the ultimate treatment for patients with end-stage lung disease, but long-term outcomes remain disappointing. According to the latest International Society for Heart and Lung Transplantation (ISHLT) Registry Report, the median survival after transplantation is approximately 6 years, and the leading cause of death beyond the first year after transplantation is chronic lung allograft dysfunction (CLAD) (1). Antibody-mediated rejection (AMR) is an increasingly recognized form of lung allograft rejection that often results in CLAD development and allograft failure (2–5). The ISHLT recently developed a consensus report to establish diagnostic criteria and a working definition of AMR after lung transplantation (6). These were based on early experience with AMR after kidney and heart transplantation and the conclusions of the national conference to assess AMR in solid organ transplantation (7–10). In the ISHLT consensus report on pulmonary AMR, the number of present criteria increases diagnostic certainty, and the diagnosis of definite AMR is based on the presence of allograft dysfunction, histologic evidence suggestive of AMR, C4d deposition, circulating donor-specific antibodies (DSA), and the reasonable exclusion of other causes (6). However, the sensitivity of C4d deposition was questioned, and the consensus report recognized that emerging evidence suggests that pulmonary AMR can be diagnosed in the absence of C4d deposition (6). Indeed, C4d deposition was notably absent in the majority of patients diagnosed with AMR in 2 recent studies (2, 4). C4d staining has been difficult to interpret in lung biopsies because of poor reproducibility, high background staining, and poor specificity for AMR (11–13). Moreover, advances in kidney transplantation have demonstrated that C4d deposition has limited sensitivity in AMR, and C4d-negative AMR is now a widely recognized phenotype (14–16). Importantly, this has led to the recognition of a unique AMR pathogenesis independent of complement activation, mediated primarily by NK cell interaction with DSA bound to endothelial cells (17, 18).

The purpose of this study was to determine the incidence of C4d-negative probable AMR after lung transplantation and compare the clinical presentation and outcomes to C4d-positive definite AMR.

METHODS

Study design and patients

We conducted a retrospective single-center cohort study. Between 7/1/2005 and 12/31/2015, 620 adults underwent 641 lung transplant procedures at Barnes-Jewish Hospital; 21 underwent re-transplantation. Six recipients were treated with a desensitization regimen before transplantation and were excluded from this study; an additional patient had a positive virtual crossmatch at the time of transplantation and was also excluded (Figure 1). Study follow-up was complete through 12/31/2016. To identify potential cases of AMR, we reviewed electronic medical records and selected cases of acute allograft dysfunction without an obvious clinical cause for evaluation for AMR. We defined acute allograft dysfunction as the development of any of the following signs: new radiographic infiltrates, hypoxemia (SpO₂ < 90% breathing room air), respiratory failure requiring invasive or non-invasive mechanical ventilation, or a ≥ 10% decline in forced vital capacity or forced expiratory volume in 1 second. During the study period, we identified 149 recipients who developed acute allograft dysfunction (Figure 1). Among these, 76 were excluded for the following reasons: 51 did not have DSA at the time of allograft dysfunction, 10 were not tested for DSA, 6 had DSA but other causes of allograft dysfunction could not be excluded, 5 had DSA but other causes of allograft dysfunction were more likely than AMR, and 4 had DSA but did not undergo a lung biopsy (Figure 1). The remaining 73 were diagnosed with definite or probable AMR according to the ISHLT definition (6): 28 fulfilled all criteria for definite AMR (C4d-positive group) and 45 fulfilled criteria for probable AMR (C4d-negative group) (Figure 1). We restricted the evaluation of probable AMR solely to cases that were C4d-negative and did not include other cases of probable AMR; we did not include cases of probable AMR where DSA was not identified, or there was no abnormal lung pathology, or where other causes of allograft dysfunction could not be excluded (6). We included cases were bacterial cultures from bronchoscopic specimens were positive if patients had a history of bacterial airway colonization and the clinical picture at the time of presentation was not consistent with bacterial pneumonia. Specifically, we excluded patients who had positive bacterial cultures if they had fever or leukocytosis at presentation. In addition, we excluded cases where a community-acquired respiratory virus, CMV, a mycobacterium, or a fungal pathogen was identified. Our Institutional Review Board approved the study protocol (IRB ID# 201701113).

Figure 1.

Flow diagram of patient inclusion into the study cohort.

Clinical management

At our center, patients and donors undergo low resolution HLA-A, B, C, DRB1, DRB3, DRB4, DRB5, and DQB typing using DNA-based methods before transplantation. Patients are screened for pre-formed human leukocyte antigen (HLA) antibodies using the LABScreen single-antigen assay every three months before transplantation, and donor lungs are accepted if a virtual crossmatch with all previously identified antibodies is compatible; sera are pre-treated with ethylenediaminetetraacetic acid (EDTA). If antibody testing detects DQA, DP, or allele-specific antibodies, we conduct additional HLA-DQA, DP, or allele level typing for the donor and recipient. Patients who had any HLA antibodies detected before transplantation are considered allosensitized. A direct complement-dependent cytotoxicity (CDC) crossmatch is performed at the time of transplantation. After transplantation, recipients are screened for DSA using the LABScreen single antigen assay at the following time points: 10 days, 1, 2, 3, 6, and 12 months then, every three months until 36 months. In addition, recipients are tested for DSA if they develop signs or symptoms of allograft dysfunction. Our center’s HLA laboratory defines DSA positivity with a threshold mean fluorescence intensity (MFI) ≥ 2,000. C1q-binding is assessed using the C1qScreen assay and considered positive if the MFI ≥ 500. Patients undergo surveillance bronchoscopy with bronchoalveolar lavage (BAL) at the following time points: 1, 2, 3, 6, and 12 months. We previously described our program’s general approach to the evaluation of acute allograft dysfunction (3). In cases where AMR is suspected based on the clinical or histological findings, C4d staining is performed using immunohistochemistry, and this is considered positive only when staining in a linear circumferential capillary sub-endothelial pattern is present (Figure 2C) (19). In contrast, irregular granular C4d deposits are considered negative. Additional examples of C4d-positive staining are shown in the online data supplement (Figure S1). A negative C4d stain is illustrated in Figure 2D. A single pathologist (DC) reviewed all cases of definite and probable AMR included in this cohort.

Figure 2.

A. Neutrophilic capillaritis – neutrophils and karyorrhectic debris in the alveolar septa. B. Diffuse alveolar damage – acute pneumonitis with hyaline membrane deposition. C. C4d deposition in the capillary endothelium by immunohistochemistry. D. Absence of C4d deposition in the capillary endothelium; staining represents C4d trapping by elastic tissues.

During the study period, recipients were treated with induction immunosuppression consisting of either basiliximab or equine anti-thymocyte globulin (ATG). Seven recipients who were included in this cohort were enrolled in a multicenter placebo-controlled trial evaluating the efficacy of Fresenius ATG (20). After transplantation, all recipients were started on maintenance immunosuppression consisting of tacrolimus, prednisone, and either mycophenolate mofetil or azathioprine. Subsequently, the maintenance regimen was changed for patient-specific indications.

Statistical analysis

We used descriptive statistics to characterize patients’ demographics and compared continuous variables between groups using t-tests (or Wilcoxon-Rank sum tests if the data were not normally distributed). We compared categorical variables between groups using chi-squared tests. Means are expressed ± standard deviation. We used the Kaplan-Meier method to estimate freedom from CLAD, survival, and CLAD-free survival after the onset of AMR, and compared groups using the log rank test. We conducted statistical analysis using SPSS 24.0 (SPSS Inc., Chicago IL) and Prism 7 (GraphPad Software Inc., La Jolla CA) and considered p < 0.05 statistically significant.

RESULTS

During the study period, 28 recipients (4%) developed C4d-positive definite AMR, and 45 (7%) developed C4d-negative probable AMR; the two groups had similar baseline characteristics (Table 1). In all C4d-positive cases, C4d deposition was focal involving < 50% of capillaries. Demographics for all patients transplanted during the study period are presented in the online data supplement (Table S1). Mean follow-up after the diagnosis of AMR was 2.0 ± 2.8 years. All recipients had a negative virtual crossmatch and a negative direct CDC crossmatch at the time of transplantation. There was no significant difference in the proportion of patients treated with the different induction immunosuppression agents or the different maintenance immunosuppression agents at the time of AMR diagnosis between the two groups (Table 1). In the overall cohort, recipients developed AMR a mean 314 ± 464 (median 186) days after transplantation. There was no significant difference in the time of onset of AMR after transplantation between the two groups (Table 2). Patients presented with non-specific signs and symptoms of acute allograft dysfunction. Fifty-five (75%) recipients were hospitalized, and 28 (38%) required invasive mechanical ventilation. There were no significant differences in the clinical presentation of AMR between the two groups (Table 2).

Table 1.

Recipient demographics and immunosuppression at the time of diagnosis of antibody-mediated rejection.

Variable	C4d-positive AMR, (n = 28)	C4d-negative AMR, (n = 45)	p value
Female gender	12 (43%)	19 (42%)	0.957
Age (mean ± SD)	45.2 ± 16.5	47.5 ± 15.1	0.610
Diagnosis			0.815
COPDa	8 (29%)	16 (36%)
Pulmonary Fibrosis	8 (29%)	13 (29%)
Pulmonary Hypertension	0 (0%)	2 (4%)
Cystic Fibrosis	8 (29%)	10 (22%)
Re-transplant	1 (4%)	1 (2%)
Other	3 (11%)	3 (7%)
Transplant operation			0.323
Bilateral lung	26 (93%)	42 (93%)
Single lung	2 (7%)	1 (2%)
Heart-lung	0 (0%)	2 (4%)
Pre-transplant allosensitization	8 (29%)	15 (33%)	0.670
CMVb serologic status			0.651
Recipient negative/Donor negative	4 (14%)	8 (18%)
Recipient negative/Donor positive	11 (39%)	13 (29%)
Recipient positive	13 (46%)	24 (53%)
Induction immunosuppression			0.597
Basiliximab	17 (61%)	30 (67%)
Equine ATGc	9 (32%)	10 (22%)
Fresenius ATG study drug	2 (7%)	5 (11%)
Calcineurin inhibitor			0.303
Tacrolimus	26 (93%)	44 (98%)
Cyclosporine A	2 (7%)	1 (2%)
Cell cycle inhibitor			0.345
Mycophenolate mofetil	21 (75%)	29 (64%)
Azathioprine	7 (25%)	16 (36%)

^aCOPD: chronic obstructive pulmonary disease

^bCMV: cytomegalovirus

^cATG: anti-thymocyte globulin

Table 2.

Clinical presentation of antibody-mediated rejection.

	C4d-positive AMR, (n = 28)	C4d-negative AMR, (n = 45)	p value
Time from transplantation to AMR
Mean ± SD, days	285 ± 357	332 ± 522	0.675
Median, days	193	161	0.928
Hospitalized	23 (82%)	32 (73%)	0.359
Radiographic infiltrates	23 (82%)	36 (80%)	0.821
Invasive mechanical ventilation	12 (43%)	16 (36%)	0.533
DSA class			0.504
HLA class I only	4 (14%)	3 (7%)
HLA class II only	20 (71%)	33 (73%)
HLA class I and class II	4 (14%)	9 (20%)
DSA to HLA DQ	20 (71%)	36 (80%)	0.399
DSA MFI
Mean ± SD	8764 ± 4141	6839 ± 3993	0.130
Median	10352	6619	0.129
Interquartile range	7644	4531
Range	3103 – 15428	2453 – 19867
C1q-positive DSAa	12/12 (100%)	12/18 (67%)	0.025
Histologic findings
Acute cellular rejection	8 (29%)	11 (24%)	0.696
A1	3 (11%)	6 (13%)
A2	4 (14%)	3 (7%)
A3	1 (4%)	2 (4%)
A4	0 (0%)	0 (0%)
Acute bronchitis	10 (36%)	18 (40%)	0.714
Neutrophilic capillaritis	15 (54%)	13 (29%)	0.035
Acute pneumonitis	22 (79%)	33 (73%)	0.614
Diffuse alveolar damage	8 (29%)	17 (38%)	0.420

^aSamples were available for C1q testing in 30 patients (C4d-positive AMR, n = 12; C4d-negative AMR, n = 18).

At the time of AMR diagnosis, bacterial cultures from bronchoscopy specimens were positive in 12 patients. There was no significant difference in the proportion of patients who had a positive bacterial culture between the C4d-positive group and the C4d-negative group; 5 (18%) patients in the C4d-positive group and 7 (16%) in the C4d-negative group had a positive bacterial culture from bronchoscopy specimens (p = 0.796). Among these, cultures were positive from bronchial washings but negative from BAL fluid in 9 and positive from both bronchial washings and BAL fluid in 3 patients. Nine of the 12 (75%) patients had CF, and Pseudomonas aeruginosa was the most commonly isolated organism (n = 9); Staphylococcus aureus, Escherichia coli, and Stenotrophomonas maltophilia were isolated in one case each. Based on the clinical presentation, radiographic and bronchoscopic findings, bacterial pneumonia was not thought to be the cause of allograft dysfunction in these cases. We excluded all cases where a community-acquired respiratory virus, CMV, a mycobacterium, or a fungal pathogen was identified.

All patients had circulating DSA at the time of AMR diagnosis. Fifty-four (74%) patients developed new DSA at the time of AMR diagnosis while 19 (26%) had DSA a median of 58 days (mean ± SD = 78 ± 66; IQR = 87) before the diagnosis of AMR. There was no significant difference in the proportion of patients with concurrent or pre-existing DSA between those who had C4d-positive and those who had C4d-negative AMR (p = 0.875). In the overall cohort, 7 (10%) had DSA only to class I HLA, 53 (73%) had DSA only to class II HLA, and 13 (18%) had DSA to class I and class II HLA. Notably, 56 (77%) had DSA to HLA-DQ. Overall, the mean MFI of the immunodominant DSA was 7690 ± 4125, and the median MFI was 7094. There was no significant difference in the distribution of DSA class or the proportion of recipients with DSA to HLA-DQ between the C4d-positive and the C4d-negative groups (Table 2). Samples obtained at the time of AMR diagnosis were available for retrospective C1q testing in 30 patients; 12 of 12 (100%) in the C4d-positive group had C1q-positive DSA, and 12 of 18 (67%) in the C4d-negative group had C1q-positive DSA (Table 2; p = 0.025). Among those who had C1q-positive DSA, there was no significant difference in the DSA C1q MFI between the two groups (p = 0.488). In the overall cohort, acute pneumonitis was the most common histologic finding, occurring in 55 (75%) recipients. In addition, concomitant acute cellular rejection (ACR) was present in 19 (26%) recipients: 9 had A1, 7 had A2, 3 had A3, and none had A4 (Table 2). Acute bronchitis and diffuse alveolar damage (DAD) (Figure 2B) were identified in 28 (38%) and 25 (34%) recipients, respectively. Lastly, neutrophilic capillaritis (Figure 2A) was present in 28 (38%) cases: 15 (54%) recipients in the C4d-positive group and 13 (29%) in the C4d-negative group had capillaritis (Table 2; p = 0.035). With the exception of capillaritis, there was no significant difference in the distribution of histologic findings between the two groups (Table 2). In addition, there was no significant difference in the distribution of histologic findings between those with C1q-positive DSA and those with C1q-negative DSA (Table S2).

Our approach to treatment evolved over time and regimens were individualized based on patients’ severity of allograft dysfunction, clinical course, and response to initial therapy. All patients were treated with intravenous immune globulin (IVIG). Dosing varied between 500 mg/kg and 2 g/kg, and treatment was continued monthly after the initial dose. The different regimens are listed in Table 3. The combination of Rituximab and IVIG was the most commonly used regimen (n = 35) (Table 3). All patients who received Rituximab were treated with a single dose of 375 mg/m². Five to 10 plasma exchange treatments were used depending on the clinical course and follow-up DSA testing. Bortezomib was given in various combinations with other treatments and dosed at 1.3 mg/m² intravenously for 4–8 doses every 3 days (Table 3). All patients who had concomitant ACR were also treated with methylprednisolone 500 mg daily for 3 days. There was no significant difference in regimen used between the C4d-positive and the C4d-negative groups (Table 3). During the study period, 24 (33%) patients in the overall cohort cleared all DSA; there was no significant difference in DSA clearance between the C4d-positive and the C4d-negative groups (p = 0.258). There was no significant association between treatment regimen and DSA clearance (Table S3; p = 0.115). Of the 28 patients who required invasive mechanical ventilation, 10 (36%) were liberated from ventilatory support; there was no significant difference in liberation from ventilatory support between the two groups (p = 0.172). In the overall cohort, 19 (26%) patients died within 30 days, but there was no statistically significant difference in 30-day mortality between the two groups; 5 (18%) recipients in the C4d-positive group and 14 (31%) in the C4d-negative group died within 30 days of the diagnosis (p = 0.210).

Table 3.

Treatment regimens.

	C4d-positive AMR, (n = 28)	C4d-negative AMR, (n = 45)	p value
Treatment regimen			0.825
Rituximab + IVIGa	16 (57%)	19 (42%)
IVIG alone	3 (11%)	8 (18%)
Rituximab + IVIG + ATGb	1 (4%)	3 (7%)
Rituximab + IVIG + PLEXc	3 (11%)	5 (11%)
Rituximab + Bortezomib + IVIG + PLEX	2 (7%)	2 (4%)
Rituximab + IVIG + ATG + PLEX	1 (4%)	1 (2%)
Rituximab + Bortezomib + IVIG	1 (4%)	1 (2%)
Rituximab + Bortezomib + ATG + IVIG	0 (0%)	1 (2%)
Rituximab + Bortezomib + ATG + IVIG + PLEX	0 (0%)	1 (2%)
ATG + IVIG + PLEX	0 (0%)	1 (2%)
ATG + IVIG	0 (0%)	2 (4%)
Bortezomib + IVIG	0 (0%)	1 (2%)
Eculizumab + Bortezomib + IVIG + PLEX	1 (4%)	0 (0%)

^aIVIG: intravenous immune globulin

^bATG: anti-thymocyte globulin

^cPLEX: plasma exchange

Six patients developed CLAD before the diagnosis of AMR and were excluded from the analysis of freedom from CLAD. During the study period, 36 patients in the overall cohort (15 in the C4d-positive group and 21 in the C4d-negative group) developed CLAD after the diagnosis of AMR. There was no significant difference in freedom from CLAD between the C4d-positive and the C4d-negative groups (Figure 3A). Among the 36 who developed CLAD, 23 (64%) developed bronchiolitis obliterans syndrome (BOS) and 13 (36%) developed restrictive allograft syndrome (RAS). There was no significant difference in the distribution of CLAD phenotype between the C4d-positive and the C4d-negative groups. In the C4d-positive group, 10 patients developed BOS and 5 developed RAS; similarly, in the C4d-negative group, 13 patients developed BOS and 8 developed RAS (p = 0.769). Sixty-one (84%) patients in the overall cohort died (n = 59) or required re-transplantation (n = 2); the median allograft survival time after the diagnosis of AMR was 246 days. However, there was no significant difference in allograft survival between the two groups (Figure 3B). CLAD was the leading cause of death accounting for 30 deaths followed by refractory AMR accounting for 21 deaths, but there was no significant difference in causes of death between the two groups (Table S5; p = 0.950). Similarly, there was no significant difference in CLAD-free allograft survival between the two groups (Figure 3C). Although there was no significant difference in freedom from CLAD between those who cleared all DSA and those who had persistent DSA (Figure 4A), those who cleared all DSA had significantly better allograft survival (Figure 4B) and CLAD-free allograft survival (Figure 4C). This difference in allograft survival was primarily due to death from refractory AMR; 21 patients who had persistent DSA died because of AMR compared to none of those who cleared the DSA.

Figure 3.

A. There was no significant difference in freedom from chronic lung allograft dysfunction (CLAD) between the C4d-positive and the C4d-negative groups (log rank p = 0.846). B. There was no significant difference in allograft survival between the C4d-positive and the C4d-negative groups (log rank p = 0.581). C. There was no significant difference in CLAD free allograft survival between the C4d-positive and the C4d-negative groups (log rank p = 0.771).

Figure 4.

A. There was no significant difference in freedom from CLAD between patients who cleared the donor-specific antibody (DSA) and those who had persistent DSA (log rank p = 0.640). B. Patients who cleared the DSA had significantly better allograft survival than those who had persistent DSA (log rank p < 0.0005). C. Patients who cleared the DSA had significantly better CLAD-free allograft survival than those who had persistent DSA (log rank p = 0.004).

DISCUSSION

In this study, the incidence of C4d-positive AMR was 4% over a 10-year period. An additional 7% of recipients developed C4d-negative probable AMR for a combined incidence of 11%. Although the ISHLT consensus report classified cases lacking C4d deposition as probable AMR, the working group recognized that “there is building evidence that AMR can be diagnosed confidently in the absence of C4d staining” (6). Indeed, our findings demonstrate that C4d-negative cases have similar clinical presentations and outcomes as C4d-positive cases. This suggests that C4d deposition is not a necessary criterion for the diagnosis of AMR. Interpretation of C4d staining in lung biopsies has been fraught with complications, and C4d deposition has not been a sensitive or specific marker for AMR (2, 4, 11–13). In addition, staining of biopsies taken at later stages of AMR after extensive endothelial injury may be patchy or granular and would be interpreted as negative (21). It is possible that C4d-negativity may not be due solely to methodological problems with staining or interpretation. Indeed, 6 of 18 (33%) C4d-negative cases in this cohort were due to non-complement binding DSA. Antibodies bound to HLA on endothelial cells can engage Fc receptors on NK cells or macrophages and elicit antibody-dependent cellular cytotoxicity (22). However, in contrast to our findings, C4d-positive AMR after kidney transplantation is associated with worse allograft survival than C4d-negative AMR (23, 24). This raises the suspicion that the lack of C4d deposition in some cases in our cohort is a false negative due to methodological problems with the assay rather than complement-independent pathways. In fact, 12 of 18 (67%) C4d-negative cases were due to complement-binding DSA. Nonetheless, we propose that cases of acute allograft dysfunction with histologic findings of lung injury and circulating DSA where other causes of allograft dysfunction are excluded be considered definite AMR even in the absence of C4d deposition. This would facilitate the diagnosis and increase awareness of AMR. Nevertheless, C4d deposition may have therapeutic implications if complement inhibitors are being considered (25–27). Thus, AMR cases may be phenotyped according to the presence or absence of C4d deposition.

Our finding of better allograft survival after AMR in patients who clear the DSA is consistent with previous reports and corroborates the paradigm that antibodies are directly pathogenic in AMR (3, 5). However, it is noteworthy that only a minority of patients cleared all DSA in spite of aggressive antibody-depleting therapy. Furthermore, the prevalence of DSA to class II HLA, and to HLA-DQ in particular, in our cohort is consistent with previous studies in pulmonary, renal, and cardiac AMR (2, 3, 5, 28–33). Emerging data suggest that HLA-DQ may be uniquely immunogenic (34). Furthermore, it is recognized that inflammatory cytokines can induce the expression of class II HLA on endothelial cells (34–36). This is consistent with endothelial cells being the focal point of initial alloimmune injury. Neutrophilic capillaritis is an obvious consequence of endothelial cell injury, but this was identified in a minority of cases in the overall cohort. Importantly, this was seen more frequently in C4d-positive cases than in C4d-negative cases. Nonetheless, capillaritis is a difficult histologic diagnosis on transbronchial lung biopsies, and it is possible that this was obscured by the findings of acute lung injury or acute pneumonitis in some cases. As previously reported, there is a high incidence of CLAD and mortality after AMR even among those who have an initial response to therapy (2–5).

This study has multiple limitations. First, since this was a retrospective study, it is possible that we did not identify some cases of AMR. There is currently no widely accepted and standardized definition of acute allograft dysfunction after lung transplantation, and this may have affected our case identification. In addition, our understanding and awareness of AMR evolved during the study period. Missing cases would affect our reported incidence of AMR. In addition, we excluded cases where the diagnosis was ambiguous because of a concomitant infection. Although this scenario is not uncommon in clinical practice, we sought to identify a cohort of AMR in a stringent manner to avoid misclassifying cases. We did not include other definitions of probable AMR proposed by the ISHLT consensus document in this study because we believe this is beyond the scope of this study and requires a focused analysis. Another limitation is that we exclusively used immunohistochemistry for the interpretation of C4d deposition. Although immunofluorescence has a better inter-observer agreement (11), our pathology lab is more experienced with staining and interpreting C4d by immunohistochemistry, and this is the clinically used assay at our center. In addition, immunohistochemistry allows C4d staining on specimens submitted in formalin whereas immunofluorescence requires submission in Michel’s solution, which requires making a decision to perform C4d staining at the time of obtaining biopsies rather than having the flexibility to review the morphology first. However, it is possible that discordance between immunohistochemistry and immunofluorescence may be present, and this would result in misclassification of some cases. In addition, we were not able to further characterize DSA by IgG subclass or C1q binding in many cases because of sample availability. Finally, our treatment approach was highly variable and depended on the severity of allograft dysfunction and clinical course. This makes comparing the efficacy of different regimens difficult particularly in light of the small sample size.

We conclude that C4d deposition is not a necessary criterion for the confident diagnosis of AMR in lung transplantation. Furthermore, the most common histologic findings are non-specific markers of lung injury, and although neutrophilic capillaritis may raise the suspicion of AMR, there are no diagnostic histologic findings. Thus, the diagnosis of pulmonary AMR requires a high index of suspicion and a multidisciplinary approach. Moreover, outcomes after AMR remain disappointing in spite of aggressive immunosuppressive and antibody-directed therapy. Finally, multi-center studies are necessary to validate our findings and identify better therapeutic interventions.

Supplementary Material

Supp FigS1

Figure S1: In these 4 examples (A–D), there is linear circumferential sub-endothelial C4d deposition (black arrows).

Supp FigS2

Figure S2: In these analyses, patients who had concomitant acute cellular rejection were excluded. (A) There was no significant difference in freedom from CLAD between patients who cleared the DSA and those who had persistent DSA (log rank p = 0.561). (B) Patients who cleared the DSA had significantly better allograft survival than those who had persistent DSA (log rank p < 0.0005). (C) Patients who cleared the DSA had significantly better CLAD-free allograft survival than those who had persistent DSA (log rank p = 0.005).

Supp TableS1-5

Table S1: Recipient demographics.

Table S2: Histologic findings of cases of antibody-mediated rejection stratified by C1q-binding donor-specific antibodies.

Table S3: Treatment regimens for antibody-mediated rejection (AMR) and donor-specific antibody (DSA) clearance.

Table S4: Treatment regimens for chronic lung allograft dysfunction (CLAD) and donor-specific antibody (DSA) clearance.

Table S5: Causes of death.

ABBREVIATIONS

ACR

Acute cellular rejection

AMR

Antibody-mediated rejection

ATG

Anti-thymocyte globulin

BAL

Bronchoalveolar lavage

BOS

Bronchiolitis obliterans syndrome

C4d

Complement component 4d

C1q

Complement component 1q

CDC

Complement-dependent cytotoxicity

CLAD

Chronic lung allograft dysfunction

CMV

Cytomegalovirus

DAD

Diffuse alveolar damage

DSA

Donor-specific antibodies

EDTA

Ethylenediaminetetraacetic acid

HLA

Human leukocyte antigen

ISHLT

International Society for Heart and Lung Transplantation

IVIG

Intravenous immune globulin

MFI

Mean fluorescence intensity

RAS

Restrictive allograft syndrome