SEARCH

SEARCH BY CITATION

Keywords:

  • Age;
  • graft failure;
  • lung transplantation;
  • primary graft dysfunction;
  • prognosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Lungs from older adult organ donors are often unused because of concerns for increased mortality. We examined associations between donor age and transplant outcomes among 8860 adult lung transplant recipients using Organ Procurement and Transplantation Network and Lung Transplant Outcomes Group data. We used stratified Cox proportional hazard models and generalized linear mixed models to examine associations between donor age and both 1-year graft failure and primary graft dysfunction (PGD). The rate of 1-year graft failure was similar among recipients of lungs from donors age 18–64 years, but severely ill recipients (Lung Allocation Score [LAS] >47.7 or use of mechanical ventilation) of lungs from donors age 56–64 years had increased rates of 1-year graft failure (p-values for interaction = 0.04 and 0.02, respectively). Recipients of lungs from donors <18 and ≥65 years had increased rates of 1-year graft failure (adjusted hazard ratio [HR] 1.23, 95% CI 1.01–1.50 and adjusted HR 2.15, 95% CI 1.47–3.15, respectively). Donor age was not associated with the risk of PGD. In summary, the use of lungs from donors age 56 to 64 years may be safe for adult candidates without a high LAS and the use of lungs from pediatric donors is associated with a small increase in early graft failure.


Abbreviations
BMI

body mass index

CF

cystic fibrosis

COPD

chronic obstructive pulmonary disease

FEV1

forced expiratory volume in 1 s

ILD

interstitial lung disease

LAS

Lung Allocation Score

LTOG

Lung Transplant Outcomes Group

OPTN

Organ Procurement and Transplantation Network

PAF

population attributable fraction

PaO2/FiO2

ratio of partial pressure of oxygen in the arterial blood to fraction of inspired oxygen

PGD

primary graft dysfunction

UNOS

United Network for Organ Sharing

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

The scarcity of suitable donor lungs limits the number of adults and children who can undergo lung transplantation each year. In 2012, only 1779 adults underwent lung transplantation in the United States, while 2329 adults were added to the waiting list that same year. At the end of 2012, 1621 adults were still actively listed for lung transplantation, 224 adult waiting list candidates had died and 189 had been removed from the list because they were “too sick,” having missed their opportunity for transplantation [1].

Potential solutions to increase the number of lung transplants performed each year include improvements in organ allocation strategies (such as the Lung Allocation Score [LAS] system), sharing of organs across organ procurement organization boundaries for high-priority candidates, growth in donation rates and organ recovery rates [2, 3], ex vivo reconditioning of unsuitable donor lungs [4], the use of lungs procured from donors after circulatory determination of death [5, 6] and the use of lungs from extended criteria donors [7]. Traditionally, the ideal organ donor is young (age <55 years), has smoked less than 20 cigarette pack-years, has no acute or chronic lung disease and has a clear chest X-ray, preserved gas exchange and an absence of lower respiratory tract infection [8].

A number of previous studies have reported an association between older donor age and higher mortality in unadjusted analyses [9, 10] or in combination with longer ischemic times [11, 12]. Older donor age has also been associated with poor gas exchange early after transplantation [13], higher rates of hospitalization for rejection [11] and higher rates of late bronchiolitis obliterans syndrome [9]. In light of the organ donor shortage, we wondered if the risk associated with older donor age might be acceptable, particularly in the current era of assigning priority to waiting list candidates at a high risk of death without transplantation. We hypothesized that older donor age would be associated with an increased rate of graft failure during the first posttransplant year and a higher risk of primary graft dysfunction (PGD) independent of donor, recipient and procedure-related factors.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Study design, data sources and participants

We performed a retrospective cohort study of adults who underwent lung transplantation in the United States between May 4, 2005 (initiation of the LAS system) and June 14, 2011 with follow-up through December 14, 2012 using data provided by the Organ Procurement and Transplantation Network (OPTN). Data obtained from the OPTN have been previously described [14]. We excluded retransplantation events for subjects who had more than one lung transplantation during the study period, pediatric (age <18) recipients, living donor transplant recipients, lobar transplant recipients, multiple organ transplant recipients, recipients who received a transplant from a donor after circulatory determination of death and those with a missing diagnosis.

We also performed a retrospective cohort study of adult lung transplant recipients enrolled in the Lung Transplant Outcomes Group (LTOG) study. LTOG is an NHLBI-funded multicenter prospective cohort study with the primary aim of examining risk factors for PGD (a list of investigators and participating centers is included in the Acknowledgment section) [15-20]. Inclusion and exclusion criteria were similar to those of the primary analysis; however, we expanded the study window to include recipients from March 30, 2002 to December 29, 2010, and we excluded recipients of lungs from donors ≥65 years (n = 14; see Supplement Methods and Figure S1). The Columbia University Medical Center Institutional Review Board approved this study and waived informed consent.

Variables and analytic approach for survival analyses

The exposure of interest was the age of the donor as ascertained by organ procurement organization personnel at the time of organ donation. The primary outcome of interest was 1-year graft failure, measured as the number of days from transplantation until death or retransplantation censored upon 365 days. A secondary outcome of interest was 5-year graft failure among 1-year graft survivors, measured as the number of days from transplantation until death or retransplantation censored upon 1825 days. We estimated hazard ratios [HRs] for graft failure using stratified Cox proportional hazards models with strata for transplant center, previous lung transplantation prior to entering the study window (May 4, 2005) and LAS diagnosis since the baseline hazard of death after lung transplantation is known to vary across levels of these variables. Mechanical ventilation at the time of transplantation and donor diabetes were later included as strata in order to satisfy the assumption of proportional hazards. We evaluated all continuous covariates (recipient age, body mass index [BMI], height and LAS; donor height and BMI and ischemic time) for nonlinear associations with the rate of 1-year graft failure as fractional polynomials in Model 3 using the “mfp” function in Stata 12.0 (Stata Corp LP, College Station, TX). The “mfp” function uses an established algorithm of likelihood ratio tests to determine whether a continuous variable has a nonlinear relationship with the outcome, and if so, determines the optimal degrees of freedom in the fractional polynomial to explain the nonlinear variation [21]. Only recipient BMI and height varied nonlinearly with the rate of 1-year graft failure, and were included in the final models as fractional polynomials each with 2 degrees of freedom. We purposefully selected multiplicative interaction terms (donor–recipient height interaction and donor–recipient gender interaction) and covariates based on the plausibility of each being associated with graft failure, donor age and/or clinical selection of older donors based on recipient factors. We employed a hierarchical modeling approach in order to allow for description of outcomes as a sum of effects from related covariates. Our hierarchical models were as follows: Model 1: donor characteristics; Model 2: donor and recipient characteristics and Model 3: donor, recipient and procedure characteristics. Covariates are described in the footnote to Table 2 and in the Supplement. Statistical significance was not a criterion for inclusion of covariates in our models. All models used robust variance estimation to account for within-center correlation with graft outcome [22].

Donor age was examined as both a nonlinear continuous variable and a categorical variable with the following levels: <18, 18–30, 30–54, 55–64 and ≥65 years. The boundaries of these categories were determined by the form of the donor age–graft failure relationship, the distribution of donor age and clinical judgment. We performed sensitivity analyses varying donor age category thresholds (deciles and decades of donor age) to examine the impact of various thresholds on our findings.

We examined the following a priori stratified analyses: quartiles of recipient age, LAS diagnosis, quartiles of LAS score and ischemic time. We had planned to examine an arbitrarily defined donor lung quality score as an ordinal variable of 0, 1, 2 or 3 with one point assigned for each of the following donor factors: smoked >20 cigarette pack-years, pulmonary infection at the time of explantation and a ratio of partial pressure of oxygen in the arterial blood to fraction of inspired oxygen (PaO2/FiO2) ratio less than 300. However, these findings were uninformative, and therefore, we examined models stratified on each of the three donor quality variables described above. We performed additional post hoc stratified analyses as described in the Supplement.

We used multiple imputation with a Markov Chain Monte Carlo method to account for missing covariate values in our multivariable analysis, as we have done previously [14]. The proportionality assumption was examined by regressing Schoenfeld residuals against time to test for independence between residuals and time. Predicted survival curves were generated from models using the missing indicator method since these analyses cannot be performed on multiply imputed data sets. The fractional polynomial plot of continuous associations of donor age and the risk of graft failure could only be generated from participants with complete data.

The population attributable fraction (PAF) is the proportion outcomes related to an exposure of interest (and unmeasured and poorly measured confounders) and represents the greatest possible proportional reduction in the number of outcomes if the exposure of interest was eliminated from the population [23]. We estimated the PAF for all recipients at 1 year using the fully adjusted multivariable Cox proportional hazard model (Model 3). Statistical significance was defined as a two-tailed p-value <0.05. Analyses were performed with SAS 9.1 (SAS Institute, Cary, NC) and Stata 12.0 (Stata Corp LP).

Variables and analytic approach for primary graft dysfunction analyses

We examined the LTOG cohort in order to determine whether donor age is independently associated with PGD, since PGD is a main contributor to early mortality after lung transplantation [15, 17]. The primary outcome was Grade 3 (PaO2/FiO2 < 200) PGD on postoperative Day 2 or 3 after transplantation [18, 24]. We estimated the risk ratios for PGD using generalized linear mixed models with Poisson regression and robust variance estimation. We performed stratified analyses with conditional logistic regression with donor age dichotomized at age 50 years, because there were too few subjects to allow further categorization of donor age or use of generalized linear mixed regression models. We used a modeling approach similar to that which is described above (see Supplement for details).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

During the study period, 9548 adults underwent lung transplantation in the United States. We excluded 688 who did not meet our criteria, leaving 8860 who formed the United Network for Organ Sharing cohort (Figure 1). The median recipient age was 58 years (interquartile range [IQR] 48–63; range 18–81 years), 41% were women, 65% underwent bilateral lung transplantation, 50% had interstitial lung disease (ILD), 33% had chronic obstructive pulmonary disease, 13% had cystic fibrosis and 4% had pulmonary vascular disease. For the entire cohort, the median follow-up time was 2.7 years (IQR: 1.1–4.0 years) and the 1-year cumulative incidence of graft failure was 19% (95% CI: 18–20%).

image

Figure 1. Flow chart of study participants.

Download figure to PowerPoint

The median donor age was 31 years (IQR 21–46, range 6–76 years). Eleven percent of donors were <18 years old, 36% were 18–29 years old, 43% were 30–54 years old, 9% were 55–64 years old and 1.1% (n = 99) were ≥65 years old. Compared to younger donors, donors age 55–64 years had similar gas exchange and ischemic times, but were more commonly female and diabetic, had more frequently been smokers and were more likely to have died as a result of cerebrovascular disease rather than trauma or anoxic injury (Tables 1 and S1). Recipients of lungs from donors age 55 to 64 years were more likely to be female, to have smoked and to have had obstructive lung disease.

Table 1. Donor, recipient and procedure characteristics at the time of transplantation
 No. of with dataDonor age, years
 <1818–2930–5455–64≥65
  • Data presented as mean (SD), median (IQR) and frequency (percentage).

  • PaO2 on FiO2 of 1.0, partial pressure of oxygen in arterial blood with fractional concentration of inspired oxygen of 1.0; BMI, body mass index; CNS, central nervous system; IQR, interquartile range; LAS, Lung Allocation Score; PA, pulmonary artery; Wedge, pulmonary capillary wedge pressure; M, male; F, female; CMV (D+/R−), cytomegalovirus donor seropositive, recipient seronegative.

  • 1

    At the time of transplantation.

Number of transplanted88609373218383776999
Donor characteristics
PaO2 on FiO2 of 1.08768429 (247–505)433 (228–500)415 (245–487)418 (259–492)424 (332–491)
Height, cm8860170 (12)174 (9)171 (10)169 (10)168 (10)
BMI, kg/m2886023 (5)25 (5)26 (5)27 (5)26 (5)
Female gender8860275 (29%)827 (26%)1902 (50%)452 (59%)59 (60%)
≥20 Pack-years smoking87905 (0.5%)201 (6.3%)774 (20%)144 (19%)5 (5.3%)
Heavy alcohol use87353 (0.3%)305 (9.6%)732 (19%)102 (13%)10 (10%)
Pulmonary infection8860333 (36%)1291 (40%)1498 (39%)266 (35%)27 (27%)
Diabetes883412 (1.3%)39 (1.2%)344 (9%)130 (17%)18 (18%)
Cause of death8645     
Anoxic brain injury1051121 (13%)415 (13%)469 (13%)45 (6%)1 (1%)
Cerebrovascular321479 (9%)359 (11%)2096 (56%)589 (77%)91 (92%)
Head trauma4305699 (77%)2337 (75%)1135 (30%)127 (17%)7 (7%)
CNS tumor756 (1%)24 (1%)41 (1%)4 (1%)0 (0%)
Recipient characteristics
Age, years886058 (45–63)57 (47–63)58 (48–63)60 (53–64)62 (56–65)
Female gender8860444 (47%)1151 (36%)1645 (43%)376 (49%)45 (45%)
LAS score886039 (35–47)39 (34–47)39 (34–49)38 (34–47)39 (35–46)
Height, cm8859168 (11)171 (10)170 (10)168 (10)169 (10)
BMI, kg/m2886025 (5)25 (5)25 (5)25 (4)25 (4)
Underweight841115 (12%)323 (10%)344 (9%)51 (7%)8 (8%)
Normal weight3475374 (40%)1258 (39%)1503 (39%)300 (39%)40 (40%)
Overweight3154310 (33%)1135 (35%)1360 (35%)311 (40%)38 (38%)
Obese1390138 (15%)502 (16%)630 (16%)107 (14%)13 (13%)
Recipient diagnosis8860     
Obstructive lung disease2953272 (29%)1071 (33%)1279 (33%)297 (39%)34 (34%)
Vascular disease38246 (5%)131 (4%)175 (5%)30 (4%)0 (0%)
Cystic fibrosis1110136 (15%)434 (13%)481 (13%)54 (7%)5 (5%)
Interstitial lung disease4415483 (52%)1582 (49%)1902 (50%)388 (50%)60 (61%)
Ever smoker8404512 (58%)1910 (62%)2301 (63%)500 (69%)66 (73%)
Prior cardiac surgery826331 (4%)97 (3%)124 (3%)30 (4%)5 (6%)
Mean PA pressure,1 mmHg778027 (11)27 (10)28 (11)27 (11)26 (10)
PA wedge pressure,1 mmHg777111 (6)11 (5)11 (6)11 (6)10 (5)
Oxygen requirement,1 L/min88173 (2–5)4 (2–6)4 (2–6)4 (2–5)4 (2–6)
Cardiac output,1 L/min73845.4 (1.4)5.4 (1.5)5.3 (1.4)5.2 (1.4)5.2 (1.3)
Insurance8860     
Private5048534 (57%)1866 (58%)2146 (56%)439 (57%)63 (64%)
Medicare2838304 (32%)993 (31%)1263 (33%)246 (32%)32 (32%)
Medicaid62268 (7%)211 (7%)294 (8%)47 (6%)2 (2%)
Other35231 (3%)148 (5%)134 (3%)37 (5%)2 (2%)
LAS increase >5 in 30 days before transplantation8740145 (16%)473 (15%)623 (16%)129 (17%)19 (19%)
Diabetes8601130 (14%)500 (16%)652 (18%)101 (14%)24 (24%)
Mechanical ventilation1886057 (6.1%)166 (5.2%)255 (6.7%)43 (5.6%)4 (4.0%)
Procedure characteristics
Ischemic time, h83515.1 (1.7)5.1 (1.7)5.1 (1.7)5.1 (1.6)5.7 (1.8)
Bilateral transplant8860622 (66%)2057 (64%)2504 (65%)482 (63%)65 (66%)
Donor–recipient gender8860     
F–F2209184 (20%)525 (16%)1161 (30%)302 (39%)37 (37%)
F–M130691 (10%)302 (9%)741 (19%)150 (20%)22 (22%)
M–F1452260 (28%)626 (19%)484 (13%)74 (10%)8 (8%)
M–M3893402 (43%)1765 (55%)1451 (38%)243 (32%)32 (32%)
CMV (D+/R−)8860226 (24%)788 (24%)877 (23%)185 (24%)30 (30%)
Center characteristics
Annualized center volume886038 (21–55)37 (23–54)39 (24–57)43 (27–64)57 (35–117)

Compared to younger donors, donors age ≥65 years had similar gas exchange, were more commonly female and diabetic, had a lower prevalence of bronchopulmonary infection, were less likely to have smoked, and almost all died from cerebrovascular disease. Donors age ≥65 years also had longer ischemic times (mean 5.7 h vs. 5.1 h for all younger age groups). Recipients of lungs from donors age ≥65 years were somewhat older and more commonly had a history of cigarette smoking and ILD (Tables 1 and S1).

Figure 2 shows the multivariable-adjusted fractional polynomial model fitting the continuous relationship between donor age and the log relative hazard of graft failure in the first year after lung transplantation (p for linearity = 0.02). Compared to donors age 18–29 years, there was a subtle increase in the rate of 1-year graft failure for donors age <18 years. The rate of 1-year graft failure was the lowest among donors age 30–50 years, and began to rise for donors older than 50 years, most markedly for donors older than 65 years.

image

Figure 2. Multivariable-adjusted fractional polynomial model fitting the continuous association between donor age (years) and the log relative hazard of graft failure in the first year after lung transplantation for the cohort with complete data (n = 7823). Estimates are adjusted for covariates listed in Model 3 in Table 2. Gray area—95% confidence interval bands. p for linearity = 0.02.

Download figure to PowerPoint

Figure 3 shows adjusted graft survival curves by donor age category (p < 0.001), and Figure S2 shows unadjusted survival by donor age category. Compared to donors age 30–54 years, recipients of lungs from donors age 55 to 64 years had an unadjusted nonsignificant 14% increase in the rate of graft failure within 1 year (HR 1.14, 95% CI 0.95–1.37; Table 2). Adjustment for recipient factors did not substantially change the effect estimate (Model 1, Table 2), but after adjustment for donor factors, the effect estimate decreased to an 8% increased rate (95% CI 0.89–1.31; Models 2 and 3, Table 2), suggesting that some of the risk associated with older donor age was explained by other donor factors.

image

Figure 3. Multivariable-adjusted survival curves for the entire cohort (n = 8860) for donor age <18, 18–29, 30–54, 55–64 and ≥65 years. Adjusted survival estimates are adjusted for covariates in Model 3. Adjusted p < 0.001. Bottom: Numbers indicate the number of surviving lung transplant recipients at each time point. The unadjusted survival curves for the entire cohort are shown in Figure S1.

Download figure to PowerPoint

Table 2. Associations between donor age and the rate of 1-year graft failure
 OverallDonor age, years
<1818–2930–5455–64≥65
  1. HR, hazard ratio; IQR, interquartile range.

  2. Unadjusted: transplant center as stratification variables.

  3. Model 1: recipient characteristics (age, gender, height, BMI, Lung Allocation Score [LAS], prior cardiac surgery, with previous lung transplantation prior to May 4, 2005, LAS diagnosis category, mechanical ventilation at time of transplantation and transplant center as stratification variables). BMI, body mass index.

  4. Model 2: Model 1 + donor characteristics (gender, height, BMI, PaO2 on FiO2 of 1, pulmonary infection, heavy alcohol use, smoked ≥20 pack-years, cause of death, donor–recipient gender mismatch, a multiplicative interaction variable for donor–recipient height, with donor diabetes as an additional stratification variable). PaO2 on FiO2 of 1.0, partial pressure of oxygen in arterial blood with fractional concentration of inspired oxygen of 1.0.

  5. Model 3: Model 2 + procedure characteristics (single vs. double lung transplant, and allograft ischemic time).

Number of transplanted88609373218383776999
Median follow-up time, years (IQR)2.7 (1.1–4.0)2.8 (1.1–4.1)2.8 (1.1–4.0)2.7 (1.1–4.0)2.7 (1.0–4.0)2.1 (0.6–3.0)
No. of donors with 1-year graft failure (%)1511 (17%)171 (18%)524 (16%)642 (17%)142 (18%)32 (32%)
1-Year graft failure rate, per person-year0.19 (0.18–0.20)0.21 (0.18–0.24)0.18 (0.17–0.20)0.19 (0.17–0.20)0.21 (0.18–0.25)0.41 (0.29–0.57)
HR for 1-year graft failure (95% CI)
Unadjusted 1.10 (0.92–1.30)0.96 (0.86–1.08)1 (Ref)1.14 (0.95–1.37)2.19 (1.53–3.20)
Model 1 1.11 (0.93–1.33)1.01 (0.90–1.14)1 (Ref)1.13 (0.94–1.36)2.16 (1.51–3.08)
Model 2 1.23 (1.01–1.51)1.12 (0.97–1.29)1 (Ref)1.08 (0.89–1.31)2.18 (1.50–3.19)
Model 3 1.23 (1.01–1.51)1.12 (0.97–1.28)1 (Ref)1.08 (0.89–1.31)2.15 (1.47–3.15)

HRs for 1-year graft failure among donors 55–64 were similar across all subgroups in stratified analyses except for ischemic time (HR 1.57, 95% CI 1.01–2.44 in the shortest ischemic time quartile [ischemic time <3.9 h], n = 2069, p for interaction = 0.04); LAS (HR 1.57, 95% CI 1.13–2.19 in the highest LAS quartile [LAS > 47.7], n = 2215, p for interaction = 0.02); and mechanically ventilated recipients (HR 1.52, 95% CI 0.78–2.95, n = 525, p for interaction = 0.04). There was no discernible increase in early graft failure among recipients of donors age 55–64 years with long ischemic times (upper quartile >6 h, HR 1.18, 95% CI 0.79–1.76) (Figure 4 and Tables S2–S4).

image

Figure 4. Stratified analyses across recipient age quartiles, ischemic time quartiles, LAS quartiles, LAS diagnosis, donor lung oxygenation (PaO2 on FiO2 of 1), recipient mechanical ventilation at the time of transplantation and donor cause of death (intracranial tumor not shown because there were too few observations to calculate meaningful effect estimates in stratified analysis). Hazard ratios (HRs) and 95% CI shown are from Model 3 (Tables S2 and S3) for graft failure within 1 year across donor age groups <18, 18–29 and 55–64 years (reference 30–54 years). LAS, Lung Allocation Score; PaO2 on FiO2 of 1.0, partial pressure of oxygen in arterial blood with fractional concentration of inspired oxygen of 1.0.

Download figure to PowerPoint

Recipients of lungs from donors age ≥65 had a more than twofold increased rate of 1-year graft failure (HR 2.15, 95% CI 1.47–3.15; Model 3, Table 2). Stratified analyses are shown in Tables S2 and S3, but should be interpreted cautiously given the small number of recipients of lungs from donors age ≥65 (n = 99).

Compared to donors age 30–54 years, recipients of lungs from donors age <18 years had a 23% increased rate of 1-year graft failure in a multivariable-adjusted analysis (HR 1.23, 95% CI 1.01–1.50; Model 3, Table 2). HRs for 1-year graft failure among recipients of pediatric lungs were similar across all subgroups in stratified analyses (Figure 4 and Tables S2–S4) except for those with poor donor gas exchange (HR 1.90, 95% CI 1.22–2.96 for PaO2/FiO2 < 242, n = 2192, p for interaction = 0.89). In addition, there was a subtle trend toward higher rates of 1-year graft failure across increasing quartiles of recipient age (HR 0.69, 95% CI 0.44–1.08 for recipients <48 years, n = 2148, and HR 1.37, 95% CI 0.97–1.94 for >63 years, p for interaction = 0.10; Figure 4 and Tables S2–S4).

Sensitivity analyses that varied the categorization of donor age showed similar findings to our primary analyses (Table S5). Compared to donors age 30–39 years, recipients of lungs from donors age 50 to 59 years had a nonsignificant 10% increase in the rate of 1-year graft failure (HR 1.10, 95% CI 0.90–1.35; Table S5), and those who received lungs from donors ≥60 years had a 59% increase in the rate of 1-year graft failure (HR 1.59, 95% CI 1.21–2.09). Additional adjustment for mean pulmonary artery pressure, pulmonary artery wedge pressure, cardiac output, supplemental oxygen flow rate, recipient diabetes, insurance coverage, cytomegalovirus mismatch, LAS change >5 units within 30 days of transplantation and annualized center volume did not appreciably changes our findings (Table S6). HRs for 5-year graft failure among 1-year graft survivors were similar across all donor lung age group, with the exception of a nonsignificant HR of 1.30 (95% CI 0.82–2.05) for recipients of donors ≥65 years old (Table S7).

The PAFs for donors age 55–64 and ≥65 years were 0.7% and 1.1%, respectively, suggesting that older donor age (and related unmeasured and poorly measured confounders) contributed to up to 1.8% of all graft failures in the first year after transplantation (Table 3). Transplanting lungs from donors age <18 years contributed up to 2.1% of graft failures in the first year after transplantation.

Table 3. Multivariable-adjusted population attributable fractions by categories of donor age
Donor age (years)Number of transplantedPAF at 1 year (%)
  1. PAF, population attributable fraction.

  2. The PAF for all recipients at 1 year was estimated using hazard ratios from a multivariable Cox proportional hazard model (Model 3 from Table 2). The PAF is the proportional reduction in the number of deaths if the exposure of interest was eliminated from the population. For example, if all lung transplant recipients who received lungs from donors age 55 to 64 years instead received lungs from donors age 30 to 54 years (reference group), up to 0.7% of graft failures might be prevented in the first year after transplantation.

<189372.1
18−548681.8
55–647690.7
≥65991.1

For the LTOG analysis, we excluded 29 recipients with missing data and 14 recipients of donor lungs age ≥65 years leaving 1212 who formed our cohort (see Supplement Methods and Figure S1). Baseline characteristics of the cohort are shown in Table S8. The median (IQR) donor age was 31 (21–46) years and ranged from 9 to 63 years, and 202 (17%) had Grade 3 PGD on postoperative Day 2 or 3. In a fully adjusted analysis, there was a marginally significant 27% increased risk of PGD among recipients of lungs from donors age 55 to 64 years compared to those age 30–54 years (relative risk [RR] 1.27, 95% CI 0.99–1.63; Table 4). When we categorized donor age by quintiles, there was no meaningful difference in the risk of PGD across donor age categories (Table S9). When we examined donor age as a continuous variable using a multivariable-adjusted generalized additive fitted model (Figure 5), there was no evidence of either a linear or nonlinear association (p for association = 0.66; p for linearity = 0.12). While there was some discernible variation in the magnitude of PGD risk among recipients of older lungs in stratified analyses, tests for interaction were nonsignificant (p > 0.50 for all) and the magnitude of the findings was small (Table S10).

Table 4. Multivariable-adjusted risk ratios for the association between donor age and Grade 3 primary graft dysfunction on postoperative Day 2 or 3 after lung transplantation
 NDonor age, years
<1818–2930–5455–64
  1. RR, risk ratio; PGD, primary graft dysfunction.

  2. Unadjusted model: transplant center is included as a random intercept.

  3. Adjusted: recipient characteristics (age, gender, height, BMI, lung transplantation prior to May 4, 2005, diagnosis, mean PA pressure); donor characteristics (gender, height, BMI, best PaO2, smoked ≥20 pack-years smoking, cause of death and donor–recipient gender mismatch); procedure characteristics (single vs. double lung transplant, cardiopulmonary bypass, FiO2 at reperfusion, packed red blood cell transfusion volume and allograft ischemic time) and transplant center as a random intercept. BMI, body mass index.

  4. There were too few recipients of lungs from donors age ≥65 years (n = 14) to calculate meaningful effect estimates, and therefore, we excluded these subjects from these analyses.

Number of transplanted121212742457091
PGD, n (%) 21 (17)71 (17)90 (16)20 (22)
RR for PGD (95% CI)
Unadjusted12121.02 (0.75–1.37)1.08 (0.82–1.42)1 (Ref)1.50 (1.06–2.12)
Adjusted12121.28 (0.90–1.82)1.19 (0.93–1.53)1 (Ref)1.27 (0.99–1.63)
image

Figure 5. Continuous relationship of donor age (years) and risk of Grade 3 PGD between 48 and 72 h after lung transplantation. Estimates are adjusted for covariates listed in Model 3 in the footnote to Table S9. p-value for linearity = 0.12. PGD, primary graft dysfunction.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

In a US registry-based nationwide sample of adult lung transplant recipients, we were unable to confirm our hypothesis that older donor age was uniformly associated with a higher rate of early graft failure after lung transplantation. We instead found a complex relationship between donor age and early outcomes after adult lung transplantation. There appeared to be a linear relationship between older donor age and a higher early rate of graft failure for recipients of lungs from donors over the age of 50 years. The magnitude of this association, however, was small and unlikely to be clinically meaningful for donors under the age of 65 years. The use of lungs from donors over the age of 65, however, was associated with a greater than twofold increased risk of early graft failure. We also found a 23% increased risk of early graft failure among adult recipients of lungs from pediatric donors. In an analysis of LTOG participants, we were unable to detect a meaningful association between donor age and the risk of PGD, although small increases in risk for recipients of very young or very old donors could not be excluded.

The use of older donor lungs for transplantation raises concerns about graft durability and long-term function. The aging lung has long been known to have decreased elastic recoil and increased alveolar size (so-called “senile emphysema”) [25, 26], leading to a progressive decline in the forced expiratory volume in 1 s (FEV1) and FEV1/forced vital capacity ratio [27]. Aging is also associated with a decline in the diffusion capacity of the lung for carbon monoxide due to loss of alveolar surface area and the capillary bed [28]. The higher risk of early graft failure observed in our study among recipients of lungs from donors ≥65 years old could potentially be attributable to a variety of overlapping mechanisms, including increased susceptibility to oxidative stress, impaired autophagy and cellular senescence and age-related changes in the lung extracellular matrix that lead to impaired wound healing and fibrosis [29-32].

Independent of the influence of cellular and physiological aging, older donor lungs may also suffer from a lifetime of environmental exposures, such as tobacco smoke, air pollution and occupational exposures. It is plausible that we did not adequately control for the effects of donor smoking, a risk factor for PGD and mortality after transplantation that might explain the observed association between older donor age ≥65 years and early graft failure in our study [18, 33]. Likewise, lack of measurement of the effects of other inhaled exposures or inadequate control of donor comorbidities in our study could also explain some of this increased risk [34].

Our results stand in contrast to earlier studies in which the use of older donor lungs for transplantation were linked to early impairment in gas exchange, a higher risk of bronchiolitis obliterans syndrome, and early mortality [9, 11-13]. In our study, we were unable to detect an increase in either early graft failure or PGD risk among recipients of lungs from donors age 55 to 64 years except among those with higher LAS scores and those receiving mechanical ventilation. While chance is a plausible explanation for our findings in stratified analyses, it seems prudent to avoid offering older donor lungs to candidates who are severely ill. It is appealing to conjecture that older donor lungs might be more susceptible to clinical or subclinical injury triggered by a recipient's pro-inflammatory state. Future studies should test such a hypothesis.

In a pre-LAS era analysis, Meyer et al. [11] found that a long ischemic time when combined with older age was associated with increased mortality after lung transplantation. Our analysis suggests that this association may no longer exist for donors age <65 years in the current LAS era. Indeed, in analyses using the LTOG cohort, donor age >50 was not a risk factor for PGD even among those with ischemic times >5.5 h. It is not clear whether these findings are related to improvements in organ procurement and cold ischemic storage (such as the now widespread use of low-potassium preservation fluid), avoidance of very long ischemic times or other confounding factors that may be difficult to ascertain.

We found that adult recipients of lungs from pediatric age donors had higher rates of 1-year graft failure. The reasons for this finding are not yet clear. Structural differences between pediatric and adult lungs, including differences in antigen presentation or extracellular matrix composition, might influence donor–recipient interactions, but data here seem to be lacking. In animal models, there are conflicting data on whether younger lungs are more [35] or less [36] susceptible to ventilator-associated lung injury. It may be that the use of pediatric donor lungs for adult recipients leads to worse outcomes because of inappropriately high recipient size-based tidal volumes during postoperative mechanical ventilation and/or donor–recipient height mismatch, a recently described association that is being actively investigated [37-39]. The higher prevalence of head trauma among younger donors in our study may have been a marker for a greater burden of pulmonary contusion, predisposing to PGD and/or alloimmune recognition of lung antigens [40], but we found no evidence of confounding or interaction by donor cause of death in this age group, and a recent study suggests that donor cause of death is not associated with PGD [18].

Our study has several limitations. Unmeasured or poorly measured confounders might account for the associations between pediatric and oldest donor lung ages and higher 1-year graft failure rates, but we included extensive donor and recipient clinical data in the multivariable analyses, including a stratification variable for transplant center. The small number of recipients who received lungs from donors age ≥65 years (n = 99) in the LAS era limits our ability to make inferences about the potential explanations for the association with early graft failure after lung transplantation. Selection bias was unlikely since we included consecutive patients undergoing lung transplantation in the United States since the initiation of the LAS system using few exclusion criteria. Examination of different thresholds to define donor age categories suggests that our selection of clinically relevant thresholds did not unduly influence our findings.

In summary, we found that lung transplantation using donors between the ages of 18 and 64 years was not associated with early graft failure among adult recipients, whereas the use of pediatric donors and donors 65 years and older is associated with an increased risk of early graft failure. Our findings suggest that lungs from donors age 55 to 64 years may be suitable for nonmechanically ventilated candidates with lower LAS scores. Increasing the number of older donors may help close the gap between supply and demand for lung transplantation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

This work was supported by National Institutes of Health (Grants R01 HL10367603 and R25HL09626004) and a Health Resources and Services Administration (Grant 231-00-0115).

Lung Transplant Outcomes Group collaborators:

Jason Christie, MD, MS (PI), Steven M. Kawut, MD, MS, Alberto Pocchetino, MD, Y. Joseph Woo, MD, Ejigayehu Demissie, MSN, Robert M. Kotloff, MD, Vivek N. Ayha, MD, James Lee, MD, Denis Hadjiliadis, MD, MHS, Melanie Doran, BS, Richard Aplenc, MD, Clifford Deutschman, MD, MS and Benjamin Kohl, MD—University of Pennsylvania (coordinating site); David J. Lederer, MD, MS (PI), Selim M. Arcasoy, MD, Joshua R. Sonett, MD, Jessie Wilt, MD, Frank D'Ovidio, MD, Matthew Bacchetta, MD, Hilary Robbins, MD, Nilani Ravichandran, NP, Nadine Al-Naamani, MD, Nisha Philip, MBBS, Debbie Rybak, BA, Shefali Sanyal, BS, Michael Koeckert, BA and Robert Sorabella, BA—Columbia University; Lorraine Ware, MD (PI), Pali Shah, MD and Stephanie Logan, RN—Vanderbilt University; Ann Weinacker, MD (PI), Ramona Doyle, MD, Susan Spencer Jacobs, MSN and Val Scott, MSN—Stanford University; Keith Wille, MD (PI), David McGiffin, MD and Necole Harris, BS—University of Alabama, Birmingham; Jonathan Orens, MD (PI), Ashish Shah, MD and John McDyer, MD—Johns Hopkins University; Vibha Lama, MD, MS (PI), Fernando Martinez, MD, MS and Emily Galopin, BS—University of Michigan; Scott M. Palmer, MD, MHS (PI), David Zaas, MD, MBA, R. Duane Davis, MD and Ashley Finlen-Copeland, MSW—Duke University; Sangeeta Bhorade, MD (PI)—University of Chicago; Maria Crespo, MD (PI)—University of Pittsburgh.

Disclosure

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. References
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ajt12428-sm-0001-SuppData.docx430K

Figure S1: Flow chart of Lung Transplant Outcomes Group (LTOG) study participants.

Figure S2: Unadjusted survival curves for the entire cohort (n = 8860) for donor age <18, 18–29, 30–54, 55–64 and >64 years. Unadjusted p < 0.001. Bottom: Numbers indicate the number of surviving lung transplant recipients at each time point.

Table S1: Donor, recipient and procedure characteristics at the time of transplantation with p-values.

Table S2: A priori stratified analyses with Model 3.

Table S3: Post hoc stratified analyses with Model 3.

Table S4: Tests for interaction with donor age categorized as <18, 18–54 and ≥55 years using Model 3.

Table S5: Associations between (A) decades and (B) deciles of donor age and the rate of 1-year graft failure.

Table S6: Post hoc analyses for associations between donor age and the rate of 1-year graft failure.

Table S7: Associations between donor age and the rate of 5-year graft failure conditional on 1-year survival.

Table S8: Recipient, donor and procedure characteristics for the LTOG cohort.

Table S9: Risk ratios for the association between donor age and Grade 3 primary graft dysfunction on postoperative Day 2 or 3 after lung transplantation by (A) quintile of donor age and (B) donor age as a continuous variable.

Table S10: Odds ratios for the association between donor age (categorized as <50 years vs. ≥0 years) and Grade 3 primary graft dysfunction on postoperative Day 2 or 3 after lung transplantation.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.