Advertisement

The ripple effect of a complication in lung transplantation: Evidence for increased long-term survival risk

Open ArchivePublished:December 12, 2015DOI:https://doi.org/10.1016/j.jtcvs.2015.11.058

      Abstract

      Objective

      Lung transplantation is a life-saving procedure for patients who have end-stage lung disease. The frequency and severity of complications have not been fully characterized. We hypothesized that early in-hospital, postoperative complications decrease long-term survival.

      Methods

      We retrospectively identified in-hospital complications in lung transplant recipients, from the period January 2007 to October 2013. Complications were graded using the extended Accordion Severity Grading System (ASGS). Complications were categorized by event and organ system. Survival analysis was performed (P < .05) using a time-dependent model.

      Results

      Among 748 eligible patients, 3381 independent in-hospital, postoperative complications occurred in 92.78% of patients. Median follow-up was 5.4 years. Complications associated with significant decrease in 5-year survival were: renal (hazard ratio [HR] 2.58, 95% confidence interval [CI] 1.40-4.48); hepatic (HR 4.08, 95% CI 2.86-5.82); cardiac (HR 1.95, 95% CI 1.56-2.45). The maximum ASGS of ≥5 (18.5% vs 73.8%), and the weighted ASGS sum >10 (2.5% vs 73.8%), were found to be significant predictors of long-term survival. Multivariate analysis identified a weighted ASGS sum of >10, and renal, cardiac, and vascular complications as predictors of decreased long-term survival.

      Conclusions

      Rigorous delineation of complications after lung transplantation showed that grade 5 ASGS in-hospital postoperative complications, and a weighted ASGS sum >10, were independent predictors of decreased long-term survival well beyond the initial perioperative period. These results may identify important targets for best practice guidelines and quality-of-care measures after lung transplantation.

      Key Words

      Abbreviations and Acronyms:

      ASGS (Accordion Severity Grading System), CCI (Charlson Comorbidity Index), CI (confidence interval), HR (hazard ratio), ICD-9 (International Classification of Diseases, 9th edition)
      Figure thumbnail fx1
      Time-dependent survival curves for patients who have any postoperative complications, versus none.
      Grade ≥5 ASGS in-hospital postoperative complications, and a weighted ASGS sum of >10 were independent predictors of long-term survival.
      The overall 90-day postoperative course influences long-term survival. Both severity of the intervention needed to overcome a complication, and having many less-severe complications, have negative effects on long-term survival. These results support the need to establish best-practice guidelines outlining the best way to avoid complications, particularly during the initial period.
      See Editorial Commentary page 1181.
      Advancements in operative techniques and postoperative management have greatly evolved and served to improve survival outcomes after lung transplantation surgery during the past 3 decades.
      • Yusen R.D.
      Survival and quality of life of patients undergoing lung transplant.
      • Yusen R.D.
      • Shearon T.H.
      • Qian Y.
      • Kotloff R.
      • Barr M.L.
      • Sweet S.
      • et al.
      Lung transplantation in the United States, 1999-2008.
      Despite these improvements, however, 5-year survival of patients undergoing lung transplantation remains in the range of 55%. Some argue
      • Valapour M.
      • Skeans M.A.
      • Heubner B.M.
      • Smith J.M.
      • Hertz M.I.
      • Edwards L.B.
      • et al.
      OPTN/SRTR 2013 Annual Data Report: Lung.
      that survival rates have plateaued in 2005, owing to a change in candidate selection from the implementation of the lung allocation score. Nevertheless, our goal should be to identify methods in which we can improve medical care given to these patients to push the field forward and improve outcomes.
      An evolving literature suggests that postoperative course is correlated with long-term survival.
      • Bach P.B.
      • Cramer L.D.
      • Schrag D.
      • Downey R.J.
      • Gelfand S.E.
      • Begg C.B.
      The influence of hospital volume on survival after resection for lung cancer.
      • Rueth NM
      • Parsons H.M.
      • Habermann E.B.
      • Groth S.S.
      • Virnig B.A.
      • Tuttle T.M.
      • et al.
      The long-term impact of surgical complications after resection of stage I nonsmall cell lung cancer: a population-based survival analysis.
      High-volume centers performing lung resection surgery have decreased perioperative morbidity rates, compared with lower-volume centers, indicating potentially more-effective management of postoperative complications that affect long-term survival. Using the American College of Surgeons National Surgical Quality Improvement Program Database, Ghaferi and colleagues
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Complications, failure to rescue, and mortality with major inpatient surgery in Medicare patients.
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Variation in hospital mortality associated with inpatient surgery.
      attributed a 2.5-fold–higher mortality rate in the “worst” hospitals to their much higher failure-to-rescue rates. They proposed standardization of the management of postoperative complications as an effective strategy in decreasing variability in mortality after major surgery, across institutions.
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Complications, failure to rescue, and mortality with major inpatient surgery in Medicare patients.
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Variation in hospital mortality associated with inpatient surgery.
      In addition, center volume may have a strong association with improved survival, even beyond the initial perioperative period of 30 days in many fields, including thoracic surgery and transplantation.
      • Livingston E.H.
      • Cao J.
      Procedure volume as a predictor of surgical outcomes.
      • Rajagopal K.
      • Lima B.
      • Petersen R.P.
      • Mesis R.G.
      • Daneshmand M.A.
      • Lemaire A.
      • et al.
      Infectious complications in extended criteria heart transplantation.
      • Kilic A.
      • George T.J.
      • Beaty C.A.
      • Merlo C.A.
      • Conte J.V.
      • Shah A.S.
      The effect of center volume on the incidence of postoperative complications and their impact on survival after lung transplantation.
      • Campos Hernandez J.P.
      • Gomez Gomez E.
      • Carrasco Valiente J.
      • Marques Lopez F.J.
      • Ruiz Garcia J.
      • Angelada Curado F.J.
      • et al.
      Influence of surgical complications on kidney graft survival in recipients of simultaneous pancreas kidney transplantation.
      • Thabut G.
      • Christie J.D.
      • Kremers W.K.
      • Fournier M.
      • Halpern S.D.
      Survival differences following lung transplantation among US transplant centers.
      Recently, our group demonstrated a negative association between in-hospital complications and 5-year overall and cancer-specific survival in patients undergoing lung resection surgery for stage I non–small cell lung cancer using the Surveillance, Epidemiology, and End Results (SEER) database.
      • Rueth NM
      • Parsons H.M.
      • Habermann E.B.
      • Groth S.S.
      • Virnig B.A.
      • Tuttle T.M.
      • et al.
      The long-term impact of surgical complications after resection of stage I nonsmall cell lung cancer: a population-based survival analysis.
      Others
      • Kilic A.
      • George T.J.
      • Beaty C.A.
      • Merlo C.A.
      • Conte J.V.
      • Shah A.S.
      The effect of center volume on the incidence of postoperative complications and their impact on survival after lung transplantation.
      have demonstrated equivalent incidences of individual postoperative complications in high-volume centers after lung transplantation, but they have identified these same high-volume centers as best equipped for minimizing adverse effects of complications on short- and long-term survival. Similar associations were identified between postoperative complications and survival in cardiac and renal transplant recipients.
      • Rajagopal K.
      • Lima B.
      • Petersen R.P.
      • Mesis R.G.
      • Daneshmand M.A.
      • Lemaire A.
      • et al.
      Infectious complications in extended criteria heart transplantation.
      • Campos Hernandez J.P.
      • Gomez Gomez E.
      • Carrasco Valiente J.
      • Marques Lopez F.J.
      • Ruiz Garcia J.
      • Angelada Curado F.J.
      • et al.
      Influence of surgical complications on kidney graft survival in recipients of simultaneous pancreas kidney transplantation.
      These important findings highlight the need for further in-depth analysis into an intriguing aspect of surgical management of complications after high-risk procedures.
      Complication types such as infections have been extensively studied in the lung transplantation literature.
      • Schuurmans M.M.
      • Benden C.
      • Inci I.
      Practical approach to early postoperative management of lung transplant recipients.
      • Chhajed P.N.
      • Tamm M.
      • Malouf M.A.
      • Glanvaille A.R.
      Lung transplantation: management and complications.
      However, few reports have rigorously examined the overall postoperative complications profile of lung transplant recipients. The purpose of the current study is to identify and create the postoperative complication profile for this population. Further, we sought to investigate the relationship between long-term survival and the way in which it is influenced by each adverse event during the postoperative course. Using the Accordion Severity Grading System (ASGS) to assess severity, we hypothesized that in-hospital, postoperative complications have an independent, negative impact on long-term survival.

      Methods

       Data Sources

      The University of Pittsburgh Medical Center Transplant Patient Management System database was utilized to identify patients suitable for this study. Prior to extracting data, institutional review board approval was obtained to ensure patient confidentiality. Between January 2007 and October 2013, a comprehensive database on all postoperative complications was prospectively recorded on all 748 patients transplanted at that time. Follow-up included the day of transplantation until death or censoring on August 31, 2015. No patient was lost to follow-up for postoperative complications or mortality.

       Data Variables

      All variables were obtained from the University of Pittsburgh Medical Center Cardiothoracic Transplant database. These included, among others, demographics such as comorbidities and diagnosis, characteristics of the surgical procedure, cause of death, and annotated in-hospital postoperative complications occurring within 90 days of surgery. Comorbidities were quantified using the Charlson Comorbidity Index (CCI).
      • Austin SR
      • Wong YN
      • Uzzo RG
      • Beck JR
      • Egleston BL
      Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser Score work.
      In-hospital complications up to 90 days were included in the analysis to achieve a comprehensive list of complications while excluding medical events that may have been unrelated to the surgery or postoperative course. Postoperative complications that occurred outside of the initial 90 days, including outside-hospital events resulting in readmission, were excluded from this study, owing to the relative ambiguity of their origin. Adverse events were identified via chart review by a dedicated team of database professionals and adjudicated by a lung transplantation surgical-quality committee prior to updating of patient profiles in the database. The date of occurrence was verified as that between the day of surgery and the date of discharge, prior to labeling them as in-hospital postoperative complications.
      Following a methodologic approach that paralleled our previous work, 94 postoperative complications were identified based on International Classification of Diseases, 9th edition (ICD-9) diagnostic definitions, as well as findings in the literature, and classified into categories
      • Rueth NM
      • Parsons H.M.
      • Habermann E.B.
      • Groth S.S.
      • Virnig B.A.
      • Tuttle T.M.
      • et al.
      The long-term impact of surgical complications after resection of stage I nonsmall cell lung cancer: a population-based survival analysis.
      (Table 1). No additions to the ICD-9 diagnostic codes were necessary. Our primary outcome measure was 1-, 3-, and 5-year overall survival, in keeping with other similar published reports.
      • Kilic A.
      • George T.J.
      • Beaty C.A.
      • Merlo C.A.
      • Conte J.V.
      • Shah A.S.
      The effect of center volume on the incidence of postoperative complications and their impact on survival after lung transplantation.
      • Campos Hernandez J.P.
      • Gomez Gomez E.
      • Carrasco Valiente J.
      • Marques Lopez F.J.
      • Ruiz Garcia J.
      • Angelada Curado F.J.
      • et al.
      Influence of surgical complications on kidney graft survival in recipients of simultaneous pancreas kidney transplantation.

      International Society for Heart and Lung Transplantation. ISHLT Transplant Registry Quarterly Reports for Lung in North America. Available at: https://www.ishlt.org/registries/quarterlyDataReportResults.asp?organ=LU&rptType=all&continent=4.

      Table 1List of complications with 1-, 3-, and 5-year survival
      Complication typeNo. of type (%)n (% of 748 patients)Survival (%)Hazard ratio (95% CI)
      1-y3-y5-y
      None54 (7.22)94.486.773.8
      Any3381 (100)694 (92.78)82.4

      P < .01
      66.9

      P = .001
      53.3

      P = .00017
      2.50 (1.40, 4.48)
      Renal271 (8.02)271 (36.23)66.7

      P < .0001
      49.8

      P < .0001
      35.4

      P < .0001
      2.58 (2.05, 3.23)
      Hepatic48 (1.42)48 (6.42)42.5

      P < .0001
      22.6

      P < .0001
      18.1

      P < .0001
      4.08 (2.86, 5.82)
      Cardiac329 (9.73)268 (35.83)72.4

      P < .0001
      54.4

      P < .0001
      39.5

      P < .0001
      1.95 (1.56, 2.45)
      Vascular114 (3.37)109 (14.57)64.2

      P = .001
      49.1

      P = .001
      29.4

      P < .0001
      2.00 (1.52, 2.65)
      Neurologic115 (3.40)86 (11.50)55.0

      P < .0001
      41.4

      P < .0001
      32.6

      P < .0001
      2.07 (1.52, 2.81)
      Musculoskeletal50 (1.48)48 (6.42)71.7

      P < .05
      42.5

      P < .0001
      27.4

      P < .0001
      2.47 (1.70, 3.59)
      Pleural space517 (15.29)345 (46.12)76.6

      P < .0001
      6.4

      P < .0001
      48.7

      P < .0001
      1.60 (1.27, 2.00)
      Infectious892 (26.38)520 (69.52)82.0

      P < .01
      64.9

      P < .0001
      51.4

      P < .001
      1.65 (1.26, 2.16)
      Pulmonary713 (21.09)536 (71.66)8.8

      P < .001
      65.5

      P = .001
      51.3

      P < .001
      1.57 (1.20, 2.05)
      Gastrointestinal59 (1.75)49 (6.55)6.2

      P = .001
      52.3

      P = .01
      28.7

      P = .001
      1.98 (1.35, 2.91)
      Psychiatric episode134 (3.96)125 (16.71)79.6

      P = .057
      58.7

      P < .01
      44.1

      P < .01
      1.56 (1.18, 2.05)
      Wound-healing complication27 (0.80)27 (3.61)54.3

      P < .001
      4.7

      P = .001
      32.6

      P < .01
      2.22 (1.36, 3.63)
      Hematologic7 (0.21)7 (0.94)24.0

      P < .01
      16.0

      P < .01
      16.0

      P < .01
      4.24 (1.75, 10.26)
      Endocrine36 (1.06)36 (4.81)88.9

      P = .608
      71.6

      P = .835
      6.9

      P = .628
      0.87 (0.49, 1.55)
      Ear, nose, throat69 (2.04)67 (8.96)85.1

      P = .931
      72.9

      P = .661
      58.1

      P = .841
      0.96 (0.64, 1.45)
      P values are comparing differences in survival between patients with who experienced a complication type versus those who did not (eg, renal vs nonrenal). CI, Confidence interval.

       Accordion Severity Grading System

      The extended ASGS version was chosen to classify adverse events. This standardized system quantifies the severity of complications, based on the magnitude of intervention (treatment for complications) necessary to overcome the event.
      • Strasberg S.M.
      • Linehan D.C.
      • Hawkins W.G.
      The accordion severity grading system of surgical complications.
      Details illustrating the course of complications, as well as the extent of treatment the patients underwent, were obtained directly from chart review and included in the database. Grades from the ASGS were weighted based on the burden of total morbidity, as quantified by Porembka and colleagues.
      • Porembka M.R.
      • Hall B.l.
      • Hirbe M.
      • Strasberg S.M.
      Quantitative weighting of postoperative complications based on the accordion severity grading system: demonstration of potential impact using the American College of Surgeons National Surgical Quality Improvement Program.
      Analysis of the ASGS grades included patients' highest-graded in-hospital postoperative complications, as well as the sum of their weighted ASGS scores, as previously described.

       Statistical Analysis

      Cox proportional hazard models were used to test the univariate associations between time-independent covariates, such as demographics, operative characteristics, comorbidities, and post-transplant survival. The Cox model, with time-dependent covariates, mainly postoperative complication categories, was used to test for univariate and multivariate associations with survival. Survival curves were estimated via counting processes in Cox regression models with time-dependent covariates. Data were analyzed using the R environment for statistical analysis and graphics,
      • Ihaka R.
      • Gentleman R.
      A language for data analysis and graphics.
      particularly the survival package.
      • Therneau T.M.
      • Grambsch P.M.
      Modeling survival data: extending the Cox model. (Statistics for Biology and Health).

      Results

       Patient Characteristics

      A total of 748 patients underwent single or double lung transplantation for end-stage lung disease, from January 2007 to October 2013, at our institution. A total of 91% of our patients were Caucasian. Median follow-up time was 5.4 years (95% confidence interval [CI] 5.2-5.8 years). Demographics of the cohort and 5-year survival estimates are given in Table 2. A total of 82.2% of patients underwent a double lung transplantation, and approximately 21% of our patients were categorized as obese (body mass index >30). The mean lung allocation score was 48.2 (standard deviation: 19). Mean intensive care unit stay for our cohort was 10.5 days (standard deviation: 15.3), with a median of 5 days (range = 1-150 days). Intraoperative cardiopulmonary bypass was utilized in one third of our patients. Fifty nine (7.9%) patients required extracorporeal membrane oxygenation. Most indications for transplantation were for idiopathic pulmonary fibrosis (47.7%) and chronic obstructive pulmonary disease (36.2%).
      Table 2Patient demographics
      VariableValue5-y survivalP value
      Age at transplant (y)
      57.2 ± 13.755.6 (51.8, 59.7)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      <.0001
       >60408 (54.5)46.1 (41.1, 51.7)<.0001
       ≤60340 (45.5)66.0 (60.7, 71.7)
       >65262 (35.0)44.2 (38.1, 51.2)<.0001
       ≤65486 (65.0)61.2 (56.7, 66.1)
      Gender
       Female316 (42.2)59.6 (54.1, 65.7).060
       Male432 (57.8)51.5 (46.5, 57.0)
      Race
       Caucasian681 (91.0)54.8 (50.8, 59.0).729
       Non-Caucasian67 (9.0)59.0 (47.6, 73.1)
      Survival, unadjusted
       1-y83.7 (81.1, 86.4)
       3-y68.7 (65.4, 72.1)
       5-y55.0 (51.2, 59.0)
       Median follow-up (y)5.4 (5.2, 5.8)
      Body mass index
      25.4 ± 5.155.0 (51.3, 59.1)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      .358
       >30155 (20.7)51.5 (43.6, 60.9).488
       ≤30593 (79.3)55.9 (51.7, 60.5)
       >3524 (3.2)48.7 (32.0, 74.3).320
       ≤35724 (96.8)55.2 (51.3, 59.3)
      Charlson Comorbidity Index
       0246 (32.9)56.6 (50.0, 64.0).819
       1-2404 (54.0)53.8 (48.9, 59.3)
       3-476 (10.2)52.7 (41.5, 67.0)
       ≥522 (2.9)63.6 (46.4, 87.3)
      Preoperative PFTs
       FVC % predicted40.9 ± 22.555.1 (51.1, 59.3)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      <.05
       FVC % predicted51.6 ± 19.255.2 (51.2, 59.4)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      .334
       DLCO % predicted33.8 ± 15.754.1 (50.0, 58.6)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      .988
      LAS48.2 ± 19.055.0 (51.2, 59.0)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      .556
      Type of transplant
       Single133 (17.8)41.6 (33.7, 51.3).00216
       Double615 (82.2)58.4 (54.3, 62.8)
      Days in ICU
      Days in ICU are modeled as a time-dependent predictor of mortality. HR = 1.029, 95% CI 1.024, 1.034.
      10.5 ± 15.350.4 (45.6, 55.7)
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      <.0001
      Median 5 (range: 1-150)
      Intraoperative life support
       CPB244 (32.6)51.4 (45.1, 58.5).309
       ECMO59 (7.9)66.2 (49.1, 89.5)
       Both12 (1.6)75.0 (54.1, 1.00)
       Neither433 (57.9)56.0 (51.3, 61.3)
      Transplant indication
       Obstructive271 (36.2)56.9 (51.0, 63.5).401
       Fibrotic357 (47.7)51.6 (46.0, 57.8)
       Suppurative93 (12.4)63.3 (53.3, 75.3)
       PH17 (2.3)52.9 (33.8, 82.9)
       Other10 (1.3)43.8 (20.0, 95.7)
      Values are mean ± standard deviation, n (%), or % (95% confidence interval), unless otherwise specified. P values are from likelihood ratio tests in Cox models. PFT, Pulmonary fibrosis; FVC, forced vital capacity; DLCO, diffusion capacity of the lungs for carbon dioxide; LAS, Lung Allocation Score; ICU, intensive care unit; CPB, cardiopulmonary bypass; ECMO, extracorporeal membrane oxygenation; PH, pulmonary hypertension.
      5-year survival estimates for quantitative predictors from univariate Cox models with predictors fixed at mean values.
      Days in ICU are modeled as a time-dependent predictor of mortality. HR = 1.029, 95% CI 1.024, 1.034.

       In-Hospital Postoperative Complication Distribution and Survival Analysis

      The overall postoperative complication distribution of our cohort is summarized in Table 1. Fifty-four (7.22%) patients had an uneventful postoperative course. A total of 3381 postoperative complications occurred in the remaining 694 (92.78%) patients during the first 90 days postoperatively. The most commonly occurring complication categories were pulmonary (71.66%), infectious (69.52%), pleural space–related (46.12%), renal (36.23%), and cardiac (35.83%). Figure 1, A depicts the difference in survival observed between those without any complications and those who developed at least one complication within 90 days of surgery. Survival at 5 years (73.8% vs 53.3%, P = .00017; hazard ratio [HR] 2.50) was greater for individuals who experienced a benign postoperative course; this difference was statistically significant.
      Figure thumbnail gr1
      Figure 1A, Survival estimates among patients who have any postoperative complications, versus none; and (B and C) with ASGS grade. ASGS, Accordion Severity Grading System; max, maximum.
      Figure 2 depicts the survival curves for the postoperative complications associated with decreases in 5-year survival that have the highest levels of statistical significance. Renal complications were associated with the most statistically significant decrease in 5-year survival, at 35.4%, compared with 64.4% in patients without renal complications (HR 2.58, 95% CI 2.05-3.23; P < .0001). Patients who developed hepatic (18.1% vs 57.3%, HR 4.08, 95% CI 2.86-5.82; P < .0001) and cardiac (39.5% vs 62.3%, HR 1.95, 95% CI 1.56-2.45, P < .0001) complications had significantly worse 5-year survival rates after those with renal complications. The 4 complication categories associated with the most statistically significant decreases in 5-year survival after renal, hepatic, and cardiac were: vascular (29.4% vs 58.5%, HR 2.00, 95% CI 1.52-2.65; P < .0001); neurologic (32.6% vs 57.1%, HR 2.07, 95% CI 1.52-2.81); musculoskeletal (27.4.0% vs 56.8%, HR 2.47, 95% CI 1.70-3.59; P < .0001); and pleural-space events (48.7% vs 60.3%, HR 1.60, 95% CI 1.27-2.00; P < .0001) (Table 1).
      Figure thumbnail gr2
      Figure 2Survival estimates among patients with (A) renal and hepatic, (B) vascular and cardiac, and (C) neurologic and musculoskeletal postoperative complication types. Musculo., Musculoskeletal.

       Accordion Severity Grading System Survival Analysis

      The complication ASGS grading distribution is noted in Table 3. The most commonly graded ASGS complication in our cohort was a grade 2 (n = 1392 [41.17%]), followed by grade 3 (n = 586 [17.33%]). Subsequent survival analysis was done with the ASGS system. Patients were first assigned an ASGS score equivalent to their highest ASGS-graded postoperative complication. The most-common maximum ASGS grade assigned was 4 (n = 260 [34.76%]), followed by 3 (n = 176 [23.53%]). Figure 1, B depicts the survival curves comparing patients who had no complications, versus those with ASGS grades 1 to 4 versus ASGS grades 5 to 6. A statistically significant difference was found in 5-year survival between patients with no postoperative complications and patients who were assigned ASGS scores 1 to 4 (73.8% vs 61.0%, P = .015). A more drastic statistically significant difference in 5-year survival was noted between patients with no complications and patients whose ASGS score was 5 to 6 (73.8% vs 18.5%, P < .00001).
      Table 3Distribution of ASGS grades, maximum scores, and weighted sums
      GroupingN (%)
      ASGS grades
       1654 (19.34)
       21392 (41.17)
       3586 (17.33)
       4470 (13.90)
       5-6279 (8.25)
      Maximum ASGS score
       0 (no complication)54 (7.22)
       111 (1.47)
       295 (12.70)
       3176 (23.53)
       4260 (34.76)
       5-6152 (20.32)
      Weighted ASGS score sum
       054 (7.22)
       0 < X ≤ 10589 (78.74)
       >10105 (14.04)
      Values are n (%); for grades, % is of complications; for scores and sum scores, % is per cohort. ASGS, Accordion Severity Grading System.
      Next, each individual ASGS grade was weighted, based on the system of Porembka and colleagues,
      • Porembka M.R.
      • Hall B.l.
      • Hirbe M.
      • Strasberg S.M.
      Quantitative weighting of postoperative complications based on the accordion severity grading system: demonstration of potential impact using the American College of Surgeons National Surgical Quality Improvement Program.
      and the sum was calculated. Weighted sums were grouped into categories in Table 3. Most weighted sums were 0 < X ≤ 10 (78.74%). Figure 1, C depicts the differences in survival based on weighted ASGS sum. A statistical difference was found in long-term survival between the weighted ASGS sum of 0 (no complications) and weighted sums 0 < X ≤ 10 (73.8% vs 58.8%, P = .007). Compared with patients with no complications, those with an ASGS sum >10 (n = 105) exhibited significantly worse 5-year survival (73.8% vs 2.5%, P < .00001). A statistically significant difference in 5-year survival was also noted between those whose weighted ASGS sum was 0 < X ≤ 10 versus >10 (58.8% vs 2.5%, P < .0001).

       Charlson Comorbidity Index Analysis

      The CCI distribution among our lung transplant–recipient cohort is provided in Table 2. Most of our patients (54%) had a CCI of 1 to 2, with the next most frequent CCI being 0 (32.9%). The CCI was not found to be a statistically significant predictor of survival (P = .819). Analysis of variance testing confirmed no statistical evidence of any difference in distribution between CCI groups in overall weighted ASGS sum (P = .969) or maximum ASGS (P = .864).

       Multivariate Analysis

      The multivariate analysis performed included age at transplantation, year of surgery, type of surgery, and all complication categories. The following predictors were selected in the multivariate analysis of long-term survival: age at transplantation >65 years (adjusted HR 1.01; P = .0012); renal events (adjusted HR 1.70; P < .0001); cardiac events (adjusted HR 1.29; P = .037); vascular events (adjusted HR 1.33; P = .037); and weighted ASGS sum (adjusted HR 1.08; P = .046). No other complication categories were selected in the multivariate model of long-term survival. The addition of CCI into the multivariate model had no significant effect (P = .67).

      Discussion

      This study is the first to systematically evaluate and profile complications that occur after lung transplantation, and to relate these data to outcome. According to Van Trigt and colleagues,
      • Van Trigt P.
      • Davis R.D.
      • Shaeffer G.S.
      • Gaynor J.W.
      • Landolfo K.P.
      • Higginbotham M.B.
      Survival benefits of heart and lung transplantation.
      in an analysis using the United Network for Organ Sharing database, the 1- and 5-year survival for patients who had undergone lung transplantation decreased from 78% to 50%. North American 5-year survival in the past decade, per the International Society for Heart and Lung Transplantation, is estimated

      International Society for Heart and Lung Transplantation. ISHLT Transplant Registry Quarterly Reports for Lung in North America. Available at: https://www.ishlt.org/registries/quarterlyDataReportResults.asp?organ=LU&rptType=all&continent=4.

      at roughly 57%, compared with 45% in patients who underwent lung transplantation in the period from 1990 to 1997. Although improved, these rates are far lower than those for other organ transplantation, such as the heart, which had a 5-year survival of 70% in the same study. As of 2011, the US Renal Disease System annual data report

      United States Renal Data System. United States Renal Data System Annual Report: Transplant. Available at: http://www.usrds.org/adr.aspx.

      states a 5-year survival of 83%. All of these figures indicate the need to identify areas in which improvements may be made in the care of patients undergoing lung transplantation.

       Pretransplantation Factors

      Pretransplantation factors, such as comorbidities, have been shown to have an association with survival after cardiothoracic and transplantation procedures.
      • Battafarano R.J.
      • Piccirillo J.F.
      • Meyers B.F.
      • Hsu H.S.
      • Guthrie T.J.
      • Cooper J.D.
      • et al.
      Impact of comorbidity on survival after surgical resection in patients with stage I non-small cell lung cancer.
      • Hernandez J.C.
      • Lok A.S.
      • Marrero J.A.
      Modified Charlson Comorbidity Index for predicting survival after liver transplantation.
      Recently, the presence of diabetes mellitus has been shown to increase death in cystic fibrosis patients who are on the wait list, before lung transplantation, but not to influence survival after lung transplantation.
      • Hayes D.
      • Patel A.V.
      • Black S.M.
      • McCoy K.S.
      • Kirkby S.
      • Tobias J.D.
      • et al.
      Influence of diabetes on survival in patients with cystic fibrosis before and after lung transplantation.
      Although our study was not designed to investigate pretransplantation status and the incidence of postoperative events, we were able to calculate the CCI for each patient, to quantify comorbidity profiles.
      Based on our analysis, CCI was not found to be a predictor of long-term survival. Moreover, no statistical differences were found in distribution of weighted ASGS sums and maximum ASGS scores within the various CCI groups. This homogeneity in the ASGS distribution per CCI group indicates that CCI may not be suitable for predicting the postoperative course of lung transplant recipients. The lack of an association between CCI and survival after lung transplantation was further confirmed in our study with our multivariate model.

       Incidence of In-Hospital Postoperative Complications

      The incidence of in-hospital postoperative complications in patients who have undergone lung transplantation in our study was >90%. This figure may be staggering, compared with other literature accounts of surgical complications. However, many of these studies use a voluntary reporting system. Moreover, many investigators focus mainly on the major adverse events categories, such as infectious profiles, especially as they pertain to the immunocompromised state of the recipient, acute and chronic rejection, airway complications, renal complications, diabetes, and diaphragmatic paralysis. In 2002, Chhajed and colleagues
      • Chhajed P.N.
      • Tamm M.
      • Malouf M.A.
      • Glanvaille A.R.
      Lung transplantation: management and complications.
      published a review that explored several other categories not often studied. Our main goal was to provide a comprehensive profile for any possible complications. Therefore, we selected patients from a time period when postoperative complications had been meticulously recorded, adjudicated, and documented in our database.
      Among the patients who developed complications in this study, 5-year survival remained consistent with published survival rates.
      • Strasberg S.M.
      • Linehan D.C.
      • Hawkins W.G.
      The accordion severity grading system of surgical complications.
      However, compared with patients who had an uneventful postoperative course, the difference in 5-year survival rate was significantly longer. Our results indicate that 90-day postoperative complications have a significant association with long-term survival and can serve as a determinant for patient outcomes.
      These results are further highlighted by the impact of having specific types of postoperative complications. In our cohort, the 7 complication categories that had the most significant adverse impact on long-term survival were renal, hepatic, cardiac, vascular, neurologic, musculoskeletal, and pleural space. These complications make up >47% of all our complications. Several other categories, including infectious, pulmonary, gastrointestinal, and psychiatric, also showed statistical significance and need to be further studied. This finding is an important step toward defining associations between donor and recipient characteristics that may significantly affect outcomes.

       Survival Analysis With the ASGS

      The ASGS was used in our analysis to assess the severity of each complication, based on the extent of the intervention necessary to overcome an adverse event. During data analysis, we observed the importance of using this systematic approach to further quantify and categorize each complication, owing to the potential complex course each may undergo. We recognized that although every complication type may occur within the context of the standard of care, each individual event may require a different level of intervention.
      Moreover, each institution, as well as every medical care provider, has preferences for treatment choice. This variability in postoperative complication identification and treatment algorithm across institutions was previously identified to be a major contributing factor to failure-to-rescue rates, which were found to be significantly higher in institutions with higher mortality after major surgery.
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Complications, failure to rescue, and mortality with major inpatient surgery in Medicare patients.
      • Ghaferi A.A.
      • Birkmeyer J.D.
      • Dimick J.B.
      Variation in hospital mortality associated with inpatient surgery.
      In the setting of a single institution, we hope to strengthen the generalizability of our results with the analysis of each complication with a well documented standardized system to minimize the effects of variability in other centers. Implementing the ASGS system may give physicians in other institutions the ability to apply our results to their practice.
      Our analysis of the ASGS system was done using the maximum score, as well as the weighted sum of all postoperative events. In our cohort, a disparity was seen in the patients with a maximum ASGS score of ≥5 and a weighted ASGS sum >10. These thresholds represent essentially different populations. This uneven distribution can be attributed to the numerous combinations possible that lead to the respective ASGS status.
      Using the maximum score allows us to identify the specific goal of decreasing all complications that result in “multisystem organ failure or a complication that required the following: an invasive procedure under general anesthesia or requiring intensive care unit admittance,” as that is the singular way to meet that threshold.
      • Strasberg S.M.
      • Linehan D.C.
      • Hawkins W.G.
      The accordion severity grading system of surgical complications.
      However, the weighted ASGS sum threshold of >10 can be met through an infinite combination of many lower-scoring events, and a handful of higher-scoring incidences.
      • Porembka M.R.
      • Hall B.l.
      • Hirbe M.
      • Strasberg S.M.
      Quantitative weighting of postoperative complications based on the accordion severity grading system: demonstration of potential impact using the American College of Surgeons National Surgical Quality Improvement Program.
      This important distinction between the 2 ASGS methods may be the reason that only the weighted ASGS sum strategy was selected by our multivariate analysis on long-term survival. Because all complications contributed to the weighted sum, this method may provide more insight into the entire postoperative course. Ultimately, a prospective study investigating both methods is needed to compare the benefits of both systems.

       Multivariate Model

      Some of our multivariate analysis findings support results published in the literature.

      United States Renal Data System. United States Renal Data System Annual Report: Transplant. Available at: http://www.usrds.org/adr.aspx.

      Mainly, better survival is associated with double lung transplantation. However, decreased survival was associated with being aged >65 years. This finding differs from results of several studies that did not find increased age to have a significant impact on short- or long-term outcomes.
      • Schaffer J.M.
      • Singh S.K.
      • Reitz B.A.
      • Zamanian R.T.
      • Mallidi H.R.
      Single- vs double-lung transplantation in patients with chronic obstructive pulmonary disease and idiopathic pulmonary fibrosis since the implementation of lung allocation based on medical need.
      • Genao L.
      • Whitson H.E.
      • Zaas D.
      • Sanders L.L.
      • Schmader K.E.
      Functional status after lung transplantation in older adults in the post-allocation score era.
      Moreover, renal, cardiac, and vascular complications were found to be significant predictors of long-term survival, along with the weighted ASGS sum. Analysis of these complication types reveals a higher frequency of higher ASGS grades. All renal complications were labeled as “acute renal dysfunction” events, and 33% (88 of 271) resulted in an ASGS score of ≥4, representing at least the need for dialysis as treatment. A total of 68 of 271 (25%) of all renal events were graded 5 or 6, signifying multiorgan system failure. When these 3 complication categories were isolated, 185 of 748 (25%) of our patients experienced ≥1 event in ≥2 of these complication categories. Because of concurrent organ failure, attributing the association with decreased 5-year survival after lung transplantation to any specific one of these complication categories is difficult. Further studies are needed to definitively study this association.
      The ASGS score, either as a weighted sum or a maximum score, can be used to predict long-term survival of post–lung transplantation patients. Although these scores can be assigned only after a complication has been resolved, thereby limiting the real-time applicability of the ASGS system, it can be effectively used as a retrospective tool, to identify possible complications that may affect survival, as seen with renal complications in our cohort. Moreover, specific patients that meet these thresholds may need to be subjected to more-meticulous monitoring after transplantation.
      These results highlight the need to implement procedure-specific postoperative monitoring and intervention protocols, with the ultimate goal of eliminating these types of postoperative adverse events. Clinical pathways have been found to be effective in several other medical fields.
      • Tomaszek S.C.
      • Fibla J.J.
      • Dierkhising R.A.
      • Scott J.P.
      • Schen K.R.
      • Wigle D.A.
      • et al.
      Outcome of lung transplantation in elderly recipients.
      • Vigneswaran W.T.
      • Bhorade S.
      • Wolfe M.
      • Pelletiere K.
      • Garrity E.R.
      Clinical pathway after lung transplantation shortens hospital length of stay without affecting outcome.
      • Pitt H.A.
      • Murray K.P.
      • Bowman H.M.
      • Coleman J.
      • Gordon T.A.
      • Yeo C.J.
      • et al.
      Clinical pathway implementation improves outcomes for complex biliary surgery.
      • Collier P.E.
      Do clinical pathways for major vascular surgery improve outcomes and reduce cost?.
      Future studies are needed to measure the success of these pathways in lung transplantation surgery, but this study shows promise in this area.
      Our study has several important attributes. With 748 eligible patients, this series offers one of the largest evaluations of patients undergoing lung transplantation that has a specific focus on their postoperative course. Moreover, the reporting of any postoperative complication was not restricted to self-reporting by the medical team. Complications were maintained by a team of database professionals and were additionally adjudicated by a panel of cardiothoracic transplant surgeons before finalizing their addition to the database.

       Limitations

      We acknowledge several limitations in this study. First, it was conducted retrospectively and within a single institution. The retrospective nature invariably subjects the current study to selection, as well as information, bias. These biases were actively addressed through implementation of various data checkpoints, with multiple investigators. Nonetheless, such bias limits the generalizability of our findings.

      Conclusions

      With appropriate patient selection and contemporary surgical techniques, vigilant postoperative management and avoidance of adverse events may potentially offer patients better long-term outcomes. This study demonstrates the independent negative association between in-hospital, postoperative complications and long-term survival in patients who have undergone lung transplantation. The overall 90-day postoperative course has an influence on long-term survival.
      Specifically, both the severity of the intervention needed to overcome a complication and the occurrence of many less-severe complications have negative impacts regarding long-term survival. These results, however, support the need to establish best-practice guidelines outlining the best way to avoid complications in these patients, particularly during this initial crucial period. The next step is to identify interventions that effectively reduce the incidence, as well as severity, of in-hospital, postoperative complications, to potentially improve long-term survival in these recipients.

       Conflict of Interest Statement

      Authors have nothing to disclose with regard to commercial support.

      References

        • Yusen R.D.
        Survival and quality of life of patients undergoing lung transplant.
        Clin Chest Med. 2011; 32: 253-264
        • Yusen R.D.
        • Shearon T.H.
        • Qian Y.
        • Kotloff R.
        • Barr M.L.
        • Sweet S.
        • et al.
        Lung transplantation in the United States, 1999-2008.
        Am J Transplant. 2010; 10: 1047-1068
        • Valapour M.
        • Skeans M.A.
        • Heubner B.M.
        • Smith J.M.
        • Hertz M.I.
        • Edwards L.B.
        • et al.
        OPTN/SRTR 2013 Annual Data Report: Lung.
        Am J Transplant. 2015; 15: 1-28
        • Bach P.B.
        • Cramer L.D.
        • Schrag D.
        • Downey R.J.
        • Gelfand S.E.
        • Begg C.B.
        The influence of hospital volume on survival after resection for lung cancer.
        N Engl J Med. 2001; 345: 181-188
        • Rueth NM
        • Parsons H.M.
        • Habermann E.B.
        • Groth S.S.
        • Virnig B.A.
        • Tuttle T.M.
        • et al.
        The long-term impact of surgical complications after resection of stage I nonsmall cell lung cancer: a population-based survival analysis.
        Ann Surg. 2011; 254: 368-374
        • Ghaferi A.A.
        • Birkmeyer J.D.
        • Dimick J.B.
        Complications, failure to rescue, and mortality with major inpatient surgery in Medicare patients.
        Ann Surg. 2009; 250: 1029-1034
        • Ghaferi A.A.
        • Birkmeyer J.D.
        • Dimick J.B.
        Variation in hospital mortality associated with inpatient surgery.
        N Engl J Med. 2009; 361: 1368-1375
        • Livingston E.H.
        • Cao J.
        Procedure volume as a predictor of surgical outcomes.
        JAMA. 2010; 304: 95-97
        • Rajagopal K.
        • Lima B.
        • Petersen R.P.
        • Mesis R.G.
        • Daneshmand M.A.
        • Lemaire A.
        • et al.
        Infectious complications in extended criteria heart transplantation.
        J Heart Lung Transplant. 2008; 27: 1217-1221
        • Kilic A.
        • George T.J.
        • Beaty C.A.
        • Merlo C.A.
        • Conte J.V.
        • Shah A.S.
        The effect of center volume on the incidence of postoperative complications and their impact on survival after lung transplantation.
        J Thorac Cardiovasc Surg. 2012; 144 (discussion 1508-9): 1502-1508
        • Campos Hernandez J.P.
        • Gomez Gomez E.
        • Carrasco Valiente J.
        • Marques Lopez F.J.
        • Ruiz Garcia J.
        • Angelada Curado F.J.
        • et al.
        Influence of surgical complications on kidney graft survival in recipients of simultaneous pancreas kidney transplantation.
        Transplant Proc. 2015; 47: 112-116
        • Thabut G.
        • Christie J.D.
        • Kremers W.K.
        • Fournier M.
        • Halpern S.D.
        Survival differences following lung transplantation among US transplant centers.
        JAMA. 2010; 304: 53-60
        • Schuurmans M.M.
        • Benden C.
        • Inci I.
        Practical approach to early postoperative management of lung transplant recipients.
        Swiss Med Wkly. 2013; 143: w13773
        • Chhajed P.N.
        • Tamm M.
        • Malouf M.A.
        • Glanvaille A.R.
        Lung transplantation: management and complications.
        Indian J Chest Dis Allied Sci. 2002; 44: 31-43
        • Austin SR
        • Wong YN
        • Uzzo RG
        • Beck JR
        • Egleston BL
        Why summary comorbidity measures such as the Charlson Comorbidity Index and Elixhauser Score work.
        Med Care. 2015; 53: e65-e72
      1. International Society for Heart and Lung Transplantation. ISHLT Transplant Registry Quarterly Reports for Lung in North America. Available at: https://www.ishlt.org/registries/quarterlyDataReportResults.asp?organ=LU&rptType=all&continent=4.

        • Strasberg S.M.
        • Linehan D.C.
        • Hawkins W.G.
        The accordion severity grading system of surgical complications.
        Ann Surg. 2009; 250: 177-186
        • Porembka M.R.
        • Hall B.l.
        • Hirbe M.
        • Strasberg S.M.
        Quantitative weighting of postoperative complications based on the accordion severity grading system: demonstration of potential impact using the American College of Surgeons National Surgical Quality Improvement Program.
        J Am Coll Surg. 2010; 210: 286-298
        • Ihaka R.
        • Gentleman R.
        A language for data analysis and graphics.
        J Comput Graph Stat. 1996; 5: 299-314
        • Therneau T.M.
        • Grambsch P.M.
        Modeling survival data: extending the Cox model. (Statistics for Biology and Health).
        Springer, New York, NY2000
        • Van Trigt P.
        • Davis R.D.
        • Shaeffer G.S.
        • Gaynor J.W.
        • Landolfo K.P.
        • Higginbotham M.B.
        Survival benefits of heart and lung transplantation.
        Ann Surg. 1996; 223: 576-584
      2. United States Renal Data System. United States Renal Data System Annual Report: Transplant. Available at: http://www.usrds.org/adr.aspx.

        • Battafarano R.J.
        • Piccirillo J.F.
        • Meyers B.F.
        • Hsu H.S.
        • Guthrie T.J.
        • Cooper J.D.
        • et al.
        Impact of comorbidity on survival after surgical resection in patients with stage I non-small cell lung cancer.
        J Thorac Cardiovasc Surg. 2002; 123: 280-287
        • Hernandez J.C.
        • Lok A.S.
        • Marrero J.A.
        Modified Charlson Comorbidity Index for predicting survival after liver transplantation.
        Liver Transplant. 2007; 13: 1515-1520
        • Hayes D.
        • Patel A.V.
        • Black S.M.
        • McCoy K.S.
        • Kirkby S.
        • Tobias J.D.
        • et al.
        Influence of diabetes on survival in patients with cystic fibrosis before and after lung transplantation.
        J Thorac Cardiovasc Surg. 2015; 150: 707-713.e2
        • Schaffer J.M.
        • Singh S.K.
        • Reitz B.A.
        • Zamanian R.T.
        • Mallidi H.R.
        Single- vs double-lung transplantation in patients with chronic obstructive pulmonary disease and idiopathic pulmonary fibrosis since the implementation of lung allocation based on medical need.
        JAMA. 2015; 313: 936-948
        • Genao L.
        • Whitson H.E.
        • Zaas D.
        • Sanders L.L.
        • Schmader K.E.
        Functional status after lung transplantation in older adults in the post-allocation score era.
        Am J Transplant. 2013; 13: 157-166
        • Tomaszek S.C.
        • Fibla J.J.
        • Dierkhising R.A.
        • Scott J.P.
        • Schen K.R.
        • Wigle D.A.
        • et al.
        Outcome of lung transplantation in elderly recipients.
        Eur J Cardiothorac Surg. 2011; 39: 726-731
        • Vigneswaran W.T.
        • Bhorade S.
        • Wolfe M.
        • Pelletiere K.
        • Garrity E.R.
        Clinical pathway after lung transplantation shortens hospital length of stay without affecting outcome.
        Int Surg. 2007; 92: 93-98
        • Pitt H.A.
        • Murray K.P.
        • Bowman H.M.
        • Coleman J.
        • Gordon T.A.
        • Yeo C.J.
        • et al.
        Clinical pathway implementation improves outcomes for complex biliary surgery.
        Surgery. 1999; 126 (discussion 756-8): 751-756
        • Collier P.E.
        Do clinical pathways for major vascular surgery improve outcomes and reduce cost?.
        J Vasc Surg. 1997; 26: 179-185