Clinical correlates, outcomes and healthcare costs associated with early mechanical ventilation after kidney transplantation

The American Journal of Surgery (2013) 206, 686-692 Association of Women Surgeons: Clinical Science Clinical correlates, outcomes and healthcare costs associated with early mechanical ventilation after kidney transplantation Hui Yuan, M.D. a, Janet E. Tuttle-Newhall, M.D. a,b, Mark Dy-Liacco, M.D. b, Mark A. Schnitzler, Ph.D. b,d, Nino Dzebisashvili, Ph.D. c, Huiling Xiao, M.S. d, David Axelrod, M.D., M.B.A. c, Brian Holt, M.D. b, Krista L. Lentine, M.D., Ph.D. b,d, * a Department of Anesthesia, Saint Louis University School of Medicine, St Louis, MO, USA; b Department of Surgery, Saint Louis University School of Medicine, St Louis, MO, USA; c Department of Surgery, Dartmouth Hitchcock Medical Center, Hanover, NH, USA; d Center for Outcomes Research, Saint Louis University School of Medicine, St. Louis, MO, USA KEYWORDS: Graft failure; Health care costs; Kidney transplantation; Mechanical ventilation; Medicare Abstract BACKGROUND: Information is lacking on the frequency, clinical implications, and costs of respiratory failure requiring mechanical ventilation after kidney transplantation. METHODS: US Renal Data System records for Medicare-insured kidney transplant recipients (1995 to 2007; n 5 88,392) were examined to identify post-transplantation mechanical ventilation from billing claims within 30 days after transplantation. RESULTS: Post-transplantation mechanical ventilation was required among 2.1% of the cohort. Independent correlates of early mechanical ventilation included recipient age, low body mass index, coronary artery disease, and cerebrovascular disease. Post-transplantation mechanical ventilation was twice as likely with delayed graft function (adjusted odds ratio, 2.13; P,.001) and 35% lower among recipients of living versus deceased donor allografts. Patients needing early mechanical ventilation experienced 5-fold higher 1-year mortality, as well as significantly higher Medicare costs during the transplant hospitalization and first post-transplantation year. CONCLUSIONS: Recognition of patients at risk for post-transplantation respiratory failure may help direct protocols for reducing the incidence and consequences of this complication. Ó 2013 Elsevier Inc. All rights reserved. The data examined herein have been supplied by the US Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US government. The authors declare no conflicts of interest. * Corresponding author. Tel.: 11-314-977-9477; fax: 11-314-977-1101. E-mail address: lentinek@slu.edu Manuscript received May 30, 2013; revised manuscript July 27, 2013 Over the past several decades, the outcomes of kidney transplantation have improved dramatically through advances in surgical techniques, preservation of the donor organ, and immunosuppression therapy. Eligibility and organ selection criteria have also broadened to include transplantation of elderly recipients with greater baseline comorbidity burdens and use of organs from expanded criteria deceased donors. However, recipient comorbidity and organ quality affect the frequency of complications, outcomes, and the costs of kidney transplantation. Early 0002-9610/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2013.07.008

H. Yuan et al. Early mechanical ventilation after kidney transplantation 687 pulmonary complications are understudied potential risks of the transplantation procedure. Transplantation surgery may result in perioperative acute lung injury through mechanisms including ischemia, cardiogenic edema, pulmonary infection, sepsis, and the use of blood products, while patients with impaired cardiopulmonary reserve before transplantation may require extended mechanical ventilation support after surgery. However, the risks and consequences of post-transplantation respiratory failure requiring mechanical ventilation early after kidney transplantation are not well defined. To advance understanding of this potential complication of transplant surgery, we examined national registry data for a large cohort of Medicare-insured kidney transplant recipients in the United States. Specifically, we sought to quantify the incidence of mechanical ventilation requirements early after transplantation and to define the clinical correlates, associated posttransplantation survival, and cost implications to Medicare. Methods Study data were drawn from records of the US Renal Data System (USRDS), which integrate Organ Procurement and Transplantation Network and United Network for Organ Sharing (UNOS) records with Medicare billing claims. The primary study sample comprised recipients of first singleorgan kidney transplants in the United States in 1995 to 2007 with Medicare as the primary payer at the time of transplantation. This study was approved by the Saint Louis University institutional review board and the USRDS. Information on recipient clinical and demographic traits, donor characteristics, and transplant factors were drawn from the Organ Procurement and Transplantation Network transplant candidate registration and transplant recipient registration forms incorporated in the USRDS. Standard criteria deceased donors (SCD) and expanded criteria deceased donors (ECD) were categorized according to UNOS definitions. Clinical and economic outcome definitions Early post-transplantation mechanical ventilation. Early post-transplantation mechanical ventilation was defined by the identification of Medicare claims (Part A or B) with corresponding International Classification of Diseases, Ninth Revision, Clinical Modification procedure codes (96.71 or 96.72) within the first 30 days after transplantation. Graft and patient survival. Mortality was defined as death from any cause. Graft failure was defined as the earliest reported date of return to maintenance dialysis or preemptive retransplantation. Observation time was censored at the date of each individual s last expected UNOS follow-up report or the end of the study (December 2007). Cost outcome. The primary economic measure was actual payments for all health care services made by Medicare. Payments were evaluated during the transplantation hospitalization and from this hospitalization to the end of the 1st year after transplantation. Monetary figures were adjusted to the prices in the year 2007 medical care component of the Consumer Price Index (the year of most recent available cost data). Statistical analysis Data management and analysis were performed using SAS version 9.2 for Windows (SAS Institute Inc, Cary, NC). Continuous data for the whole cohort and for living donor (LD), SCD, and ECD recipients are summarized as means 6 SD, and categorical data are summarized as proportions. The incidence of the requirement for early post-transplantation mechanical ventilation was computed overall and among subgroups defined by given baseline traits. Adjusted associations of baseline factors with the likelihood of early post-transplantation mechanical ventilation (adjusted odds ratio [aor]) were quantified with multivariable logistic regression. Variation in 1-year patient and death-censored graft survival according to the requirement for early mechanical ventilation was compared using the Kaplan-Meier method, and the log-rank test was used to assess the statistical significance of differences in unadjusted survival. Multivariate Cox regression was used to examine associations of the requirement for early post-transplantation mechanical ventilation with the relative risks for death and graft failure including adjustment for baseline recipient, donor, and transplant factors listed in Table 1. Analyses were performed for the whole cohort and also separately by organ donor type. Finally, the marginal impact of the requirement for mechanical ventilation on health care costs to Medicare during the transplantation hospitalization and 1st year after transplantation were computed using a linear regression equation as E(Y) 5 b 1 X 1 1 b 2 X 2 1. b k X k, where E(Y) is Medicare payments within a period of interest, X k is the value of a given predictor variable, and b k is the marginal costs associated with a 1-unit change in a given variable after adjustment for other observed factors in the model. Thus, for binary variables such as mechanical ventilation, the b k parameter quantifies the marginal costs incurred by patients who required mechanical ventilation compared with those who did not, adjusted for recipient, donor, and transplant factors. Results Demographic and clinical characteristics The total sample of Medicare-insured kidney transplant recipients from 1995 to 2007 (n 5 88,392) included 51,284 SCD, 20,235 LD, and 16,873 ECD graft recipients. Distributions of clinical traits in the full sample are shown in Table 1. Recipients of LD allografts tended to be younger, were more commonly white, and less frequently had diabetes, sensitization, or delayed graft function (DGF) compared

688 The American Journal of Surgery, Vol 206, No 5, November 2013 Table 1 Sample trait distributions and adjusted clinical correlates of the likelihood of early mechanical ventilation requirement after kidney transplantation in the full cohort and by donor type All donors All donors LDs SCDs ECDs Baseline factor (n 5 88,392) x aor (95% CI) aor (95% CI) aor (95% CI) aor (95% CI) Demographic and clinical factors Recipient age (y),18 2.7 1.92 (1.38 2.67) 1.67 (.75 3.76) 1.86 (1.26 2.75)* 4.12 (1.40 12.16)* 18 30 11.8 Reference Reference Reference Reference 31 44 27.0 1.34 (1.09 1.65)* 2.17 (1.29 3.65)* 1.13 (.89 1.45) 1.72 (.95 3.12) 45 59 35.4 1.65 (1.35 2.02) 2.13 (1.26 3.61)* 1.45 (1.14 1.84)* 2.14 (1.20 3.82)* R60 23.0 2.25 (1.83 2.78) 3.00 (1.75 5.11) 2.02 (1.57 2.60) 2.75 (1.53 4.93) Male gender 60.6.97 (.88 1.07) 1.03 (.79 1.35).94 (.83 1.06) 1.05 (.86 1.28) Recipient race White 62.5 Reference Reference Reference Reference Black 29.6 1.10 (.98 1.23).76 (.34 1.68) 1.06 (.92 1.21) 1.15 (.93 1.43) Other 8.0.85 (.69 1.03).68 (.33 1.38).72 (.55.95)* 1.03 (.72 1.47) Body mass index (kg/m 2 ),18.5 4.0 1.49 (1.18 1.88) 1.92 (1.10 3.34)* 1.31 (.97 1.76) 1.88 (1.15 3.08)* 18.5 25 38.6 Reference Reference Reference Reference 25 30 34.2.93 (.83 1.04).97 (.71 1.33).95 (.83 1.09).85 (.68 1.07).30 23.2 1.00 (.89 1.13) 1.05 (.75 1.48).97 (.83 1.14) 1.04 (.81 1.33) Cause of ESRD Diabetes 23.3 0.97 (0.82 1.15) 1.10 (.69 1.74).93 (.75 1.16).97 (.70 1.35) Glomerulonephritis 18.4.81 (.69.94)*.66 (.43 1.02).87 (.72 1.06).73 (.52 1.02) Hypertension 21.8 1.04 (.90 1.19) 1.08 (.74 1.59) 1.09 (.91 1.30).91 (.70 1.20) Polycystic kidney disease 6.2.80 (.64 1.00).97 (.52 1.83).86 (.65 1.13).60 (.38 0.96)* Other 30.3 Reference Reference Reference Reference Recipient comorbidities Diabetes 26.0 1.09 (.94 1.26) 1.12 (.74 1.69) 1.09 (.90 1.32) 1.03 (.78 1.37) Coronary disease/angina 9.6 1.23 (1.06 1.42)*.98 (.66 1.48) 1.34 (1.11 1.61)* 1.18 (.89 1.56) Arrhythmia 0.2 1.00 (.44 2.28) 5.38 (1.24 23.34)*.74 (.23 2.36).53 (.07 3.95) Peripheral vascular disease 4.6 1.08 (.88 1.32) 1.17 (.70 1.96) 1.13 (.87 1.47).93 (.61 1.43) COPD 1.1 1.35 (.94 1.95).81 (.25 2.57) 1.69 (1.09 2.62)*.99 (.43 2.27) Cerebrovascular disease 2.8 1.29 (1.02 1.64)* 1.17 (.61 2.26) 1.14 (.82 1.57) 1.77 (1.15 2.73)* Previous transplantation 18.5.88 (.64 1.22) 1.16 (.65 2.06).69 (.42 1.11) 1.07 (.55 2.11) Preemptive transplantation 2.8.85 (.77.93).82 (.64 1.04).93 (.83 1.05).69 (.57.83) Peak panel reactive antibody level,10 68.6 Reference Reference Reference Reference 10 79 20.5 1.06 (.94 1.19).91 (.64 1.30).98 (.84 1.13) 1.34 (1.07 1.69)* R80 8.3.99 (.83 1.19) 1.04 (.55 1.98).87 (.70 1.09) 1.39 (.95 2.03) Missing 2.6.95 (.70 1.28).73 (.34 1.56).78 (.51 1.20) 1.72 (1.02 2.89)* HLA mismatches Zero A, B, and DR 11.7.69 (.58.82).65 (.38 1.14).61 (.50.75) 1.01 (.71 1.44) Zero DR 17.3.85 (.75.97)* 1.09 (.80 1.49).74 (.62.87) 1.05 (.81 1.35) Other 71.0 Reference Reference Reference Reference Delayed graft function 22.4 2.13 (1.93 2.35) 3.96 (2.89 5.42) 2.12 (1.87 2.39) 1.83 (1.52 2.22) Donor variables Donor race White 79.0 Reference Reference Reference Reference Black 13.8.97 (.84 1.12) 1.87 (.83 4.18) 1.00 (.83 1.20).70 (.52.96)* Other Race 7.2.80 (.65.99)* 1.57 (.82 2.98).78 (.60 1.03).56 (.35.90)* Donor type SCD 58.0 Reference NA NA NA LD 22.9.65 (.56.75) NA NA NA ECD 19.1 1.00 (.89 1.12) NA NA NA aor 5 adjusted odds ratio; CI 5 confidence interval; COPD 5 chronic obstructive pulmonary disease; ECD 5 expanded criteria deceased donor; ESRD 5 end-stage renal disease; HLA 5 human leukocyte antigen; LD 5 living donor; NA 5 not applicable; SCD 5 standard criteria deceased donor. *P,.001 to,.05. P,.0001 to,.001. P,.0001. x Percentages indicate fractions of recipients of a given transplantation type with the indicated clinical trait (column percentages).

H. Yuan et al. Early mechanical ventilation after kidney transplantation 689 with deceased donor recipients. ECD recipients tended to be the oldest group and to have the highest frequency of diabetes. Frequency and clinical correlates of early posttransplantation mechanical ventilation Overall, early mechanical ventilation was required in 2.1% of the study sample, including higher unadjusted proportions of ECD (2.7%) and SCD (2.3%) recipients compared with LD recipients (1.2%). Multivariate regression with adjustment for recipient, donor, and transplant factors in the full study sample demonstrated that the likelihood of early mechanical ventilation was approximately twice that among recipients,18 years of age (aor, 1.92; 95% confidence interval [CI], 1.38 to 2.67) compared with those aged 18 to 30 years and increased in a graded manner with recipient age.30 years to more than 2-fold risk among those aged R60 years (aor, 2.25; 95% CI, 1.83 to 2.78; Table 1). The adjusted likelihood of early mechanical ventilation was increased by 49% in underweight compared with normalweight recipients (aor, 1.49; 95% CI, 1.18 to 1.88), by 23% in recipients with baseline coronary artery disease (aor, 1.23; 95% CI, 1.06 to 1.42), and by 29% in those with cerebral vascular disease (aor, 1.29; 95% CI, 1.02 to 1.64). Recipients who developed DGF had approximately twice the adjusted likelihood of early mechanical ventilation compared with recipients without DGF (aor, 2.13; 95% CI, 1.93 to 2.35). The likelihood of early mechanical ventilation was decreased by 19% in patients with end-stage renal disease due to glomerulonephritis compared with other causes (aor,.81; 95% CI,.69 to.94), and was 15% lower in those transplanted preemptively (aor,.85; 95% CI,.77 to.93). Early mechanical ventilation was also lower in those without A, B, and DR mismatches as well as in those with 0 DR mismatches compared with higher degrees of mismatches. After adjustment for captured factors, recipients of LD kidneys had a 35% reduction in the likelihood of early mechanical ventilation compared with SCD recipients (aor,.65; 95% CI,.56 to.75), but the adjusted risk for early mechanical ventilation was not different among recipients of ECD compared with SCD kidneys. Relative patterns were generally similar when considered among recipients of each allograft type, except that history of pretransplantation arrhythmia was a strong predictor of early mechanical ventilation requirement among LD recipients, and chronic obstructive pulmonary disease (COPD) history was a significant predictor among SCD recipients. DGF was a stronger correlate of early mechanical ventilation requirement regardless of allograft type. Patient and graft survival according to early post-transplantation mechanical ventilation status First-year patient survival was markedly lower among patients who required early mechanical ventilation (70.3% vs 95.2%, P,.0001), with much of the decline in survival occurring in the first 2 months (Fig. 1A). Similar adverse associations of the need for mechanical ventilation with reduced survival were observed among recipients stratified by donor type (Fig. 1B D). Death-censored graft survival at 1 year was Figure 1 Patient survival according to the requirement for early mechanical ventilation after kidney transplantation.

690 The American Journal of Surgery, Vol 206, No 5, November 2013 84.9% among those who required early mechanical ventilation compared with 93.1% among those who did not. Adjusted associations of early posttransplantation mechanical ventilation with death and graft loss In multivariate Cox regression analysis including adjustment for baseline factors, patients who required early mechanical ventilation had 6 times the relative risk for death (adjusted hazard ratio [ahr], 5.99; 95% CI, 5.47 to 6.55) in the 1st year after transplantation compared with those who did not require early mechanical ventilation. Other significant predictors of 1st-year mortality included older recipient age, baseline coronary disease (ahr, 1.28; 95 CI%, 1.18 to 1.39) and peripheral vascular disease (ahr, 1.14; 95% CI, 1.01 to 1.28), elevated level of peak panel reactive antibody (ahr, 1.35; 95% CI, 1.21 to 1.50), DGF (ahr, 1.90; 95% CI, 1.78 to 2.02), and ECD allograft type (ahr, 1.29; 95% CI, 1.21 to 1.38). When considered by donor type, the relative mortality implications of early mechanical ventilation were similar, with approximately 5 to 6 times the relative mortality among LD (ahr, 5.66; 95% CI, 4.23 to 7.58), SCD (ahr, 6.47; 95% CI, 5.75 to 7.27), and ECD (ahr, 5.33; 95% CI, 4.53 to 6.27) recipients. The requirement for early post-transplantation mechanical ventilation was also associated with significantly increased risk for death-censored graft loss in the 1st year after transplantation, although the effect sizes were smaller than for patient mortality. Specifically, the relative risk for death-censored graft loss was increased by 76% in the full study cohort (ahr, 1.76; 95% CI, 1.54 to 2.00), with similar effect sizes among SCD (ahr, 1.69; 95% CI, 1.42 to 2.01), ECD (ahr, 1.36; 95% CI, 1.36 to 2.41), and LD (ahr, 2.29; 95% CI, 1.58 to 3.30) recipients. Associations of early post-transplantation mechanical ventilation with adjusted Medicare costs In multivariate linear regression including covariate adjustment, the requirement for early mechanical ventilation was associated with approximately $8,412 higher marginal costs to Medicare during the transplantation hospitalization and $16,206 higher costs during the remaining 1st year after transplantation. Other correlates of increased Medicare costs during these periods included recipient age.60 years, low body mass index (BMI), cerebrovascular disease, prior transplantation status, desensitization, and DGF, while adjusted Medicare payments decreased in more recent compared with previous years. Comments Limited information is available on the frequency and outcomes of mechanical ventilation use early after kidney transplantation. We examined mechanical ventilation requirements among a large sample of Medicare-insured kidney transplant recipients and observed several key findings. First, the need for mechanical ventilation early after kidney transplantation was relatively uncommon overall, occurring in 2.1% of the full cohort, including 1.2% of LD recipients and 2.7% of ECD recipients. Second, independent clinical correlates of early mechanical ventilation requirements included recipient age, low BMI, coronary artery disease, and cerebrovascular disease. After multivariate adjustment, the need for mechanical ventilation was twice as likely among recipients with DGF and 35% lower among recipients of LD versus deceased donor allografts. Third, the requirement for early mechanical ventilation was a potent predictor of adverse outcomes, including 5 times the relative risk for 1 year mortality and 76% higher risk for death-censored graft loss. Finally, patients who required early mechanical ventilation incurred significantly higher Medicare costs during the transplantation hospitalization and the 1st year after transplantation. Mechanical ventilation is often used to treat patients who develop postoperative respiratory failure. The VA Surgery Quality Improvement Program and National Surgical Quality Improvement Program data sets reported postoperative respiratory failure requiring mechanical ventilation support among 3.1% to 3.4% of patients within 30 days after a variety of forms of general surgery in the past decade. 1,2 The present study demonstrated an overall incidence of mechanical ventilation requirements within 30 days of 2.1% in a national kidney transplantation cohort, although the incidence exceeded 3% in some subgroups, such as transplant recipients with COPD, and was nearly 4% in those who experienced DGF. Although kidney transplantation offers better long-term survival compared with chronic dialysis for most patient subgroups, age and baseline comorbidity affect the likelihood of post-transplantation complications. 3 We found that recipient age, comorbidities such as coronary artery disease, cerebrovascular disease, and COPD, and low BMI significantly influenced the likelihood of needing mechanical ventilation within 30 days after transplantation surgery. Specifically, we observed a U-shaped association between age and the risk for early mechanical ventilation after kidney transplantation, such that compared with recipients aged 18 to 30 years in the full cohort, those aged,18 years had approximately twice the risk, and there was also a graded risk increase with age.30 years. Both pediatric and geriatric patients have unique risks profiles for complications after kidney transplantation. Often, when transplanting adult-sized kidneys into smaller children, requirements for ventilation relate more to issues of abdominal domain and the need to expand blood volumes to perfuse the kidney than to underlying pulmonary abnormalities. 4 In a large single-center study, one third of patients rehospitalized after kidney transplantation were aged R50 years, and elderly patients requiring readmission more commonly required intensive care (8.6% vs 2.4%) and had higher in-hospital mortality (10.2% vs 2.6%) compared with young patients requiring readmission. 5

H. Yuan et al. Early mechanical ventilation after kidney transplantation 691 Geriatric transplant recipients warrant particular attention in the preoperative, perioperative, and postoperative periods for full assessment of baseline comorbidity, correction of reversible processes when possible, and attention to risks of delayed organ function, especially when organs from expanded criteria donors are used. Independent of age, the presence of baseline comorbidity is associated with requirements for mechanical ventilation after kidney transplantation. In the present study, 17% of recipients had diagnoses of atherosclerotic vascular disease before transplantation, and coronary and cerebrovascular diseases predicted increased risk for early mechanical ventilation requirements. Cardiovascular diseases contribute to approximately 30% of post-transplantation mortality, with particularly high risk in the peritransplantation period. 6 Although cause and effect cannot be determined from this observational study, at a minimum, increased vigilance to respiratory complications in patients with vascular disease is warranted. Furthermore, the impact of aggressive cardiac protection and maintenance of adequate organ perfusion according to current clinical practice guidelines 7 deserve study as a potential means for reducing the risk for respiratory complications. Controversy exists regarding the use of BMI and other metrics of body composition in the selection of recipients for kidney transplantation. 8 In the present study, we did not find significant associations of overweight and obese BMI status with requirements for post-transplantation mechanical ventilation after covariate adjustment. However, patients with underweight BMIs had a 49% increase in the risk for needing post-transplantation ventilator support. Although our models were adjusted for recipient, donor, and transplant factors recorded in the registry, this finding might reflect uncaptured comorbidities and frailty in underweight patients. In a large national study, Meier-Kriesche et al 9 found that low BMI (,18 kg/m 2 ) was associated with adverse clinical outcomes, including reduced patient and graft survival. Underweight states due to comorbidity and/or malnutrition could predispose patients to infection and organ dysfunction during stressful situations such as kidney transplantation, resulting in higher chances of postoperative respiratory failure. DGF, defined as the need for dialysis within the 1st week after transplantation, was strongly associated with the need for mechanical ventilation after kidney transplantation in the present study. Poor renal allograft function is often associated with hypervolemia, which in turn may precipitate acute respiratory failure in the postoperative setting. Decisions on when to use dialysis in a patient with poor allograft function can be complicated by competing concerns for maintaining allograft perfusion versus managing pulmonary edema. Prospective study of the optimal timing of use of dialysis in patients with poor initial allograft function, considering both renal and pulmonary outcomes, is warranted. In the VA Surgical Quality Improvement Program and National Surgical Quality Improvement Program cohorts, patients requiring ventilator support after general surgery had high mortality rates of 26.7% to 30% within 30 days 1,2 ;by comparison, 30-day mortality in kidney transplant recipients requiring mechanical ventilation was 16.5%. The present study quantifies the prognostic implications of early mechanical ventilation requirements for adverse outcomes, including 6 times the risk for death, 76% higher death-censored allograft failure, and increased costs of care compared with absence of the need for mechanical ventilation. The identification of markers of reduced survival in kidney transplant recipients may guide the development of management protocols to improve outcomes, such as aggressive medical management aimed at optimizing medical conditions that could compromise pulmonary function after the surgical procedure, as well as intraoperative and postoperative management and monitoring protocols. Furthermore, the identification of high-risk patients by transplantation centers is critically important in the current regulatory environment, in which centers are graded for their recipient and graft survival compared with expected outcomes according to the Scientific Registry of Transplant Recipients. 10 Consequences of a citation for recipient death and/or graft loss rates that exceed expected from the United Network for Organ Sharing Membership and Professional Standards Committee, the Centers for Medicare and Medicaid Services, or private insurance networks are serious and can include costly systems improvement agreements, exclusions from insurance plan participation, and even program closure. Transplantation centers must also manage costs of care within relatively fixed reimbursement limits, and costly complications may lead to resource utilization and expenses that exceed reimbursement. 11 Better understanding of patients at risk for posttransplantation respiratory failure, which in turn is marker for mortality and expenses, may help centers optimize preparation and post-transplantation monitoring and management for particularly high-risk subgroups. Limitations of this study include selection of the sample as Medicare beneficiaries. Although our findings may not generalize to beneficiaries of other insurance systems, clinical and economic studies among Medicare-insured transplant recipients are relevant as Medicare is the largest single payer for transplantation services in the United States. The retrospective design is also a limitation. It is possible that future changes in clinical practice may modify the outcomes implications of early post-transplantation mechanical ventilation requirements. Information was limited to variables captured in the USRDS and we did not have access to other clinical factors relevant to this topic, such as the cause of respiratory failure or detailed descriptions of the surgical procedure (eg, operative times, ischemic times, blood loss), or intraoperative volume management. These factors warrant examination in future research. With respect to the structure of our economic models, alternatives to ordinary least squares models, such as regressions estimating the determinants of the natural logarithm of Medicare payments, may be more efficient but also may produce biased estimates and are difficult to interpret. Because we had access to cost data for very large samples, we used the unbiased estimator. Our past work has demonstrated nearly identical results with ordinary

692 The American Journal of Surgery, Vol 206, No 5, November 2013 least squares cost regression and regressions on the natural logarithm of Medicare payments, and ordinary least squares regression has become our standard in analyses of the economic impact of complications in transplantation. 12,13 Conclusions This analysis of a large national cohort demonstrates that overall, 2% of kidney transplant recipients develop respiratory failure requiring mechanical ventilation early after transplantation. Clinical correlates of increased risk include pediatric and elderly recipient age, low BMI, atherosclerotic vascular disease, COPD, and donor type. Furthermore, DGF is strongly associated with need for post-transplantation mechanical ventilation. Although uncommon, mechanical ventilation requirements after kidney transplantation predict markedly reduced patient survival, lower allograft survival, and increased costs. Better recognition of patients at risk for post-transplantation respiratory failure may guide protocol development to attempt to reduce the incidence and consequences of this complication. References 1. Gupta H, Gupta PK, Fang X, et al. Development and validation of a risk calculator predicting postoperative respiratory failure. Chest 2011;140:1207 15. 2. Johnson RG, Arozullah AM, Neumayer L, et al. Multivariable predictors of postoperative respiratory failure after general and vascular surgery: results from the patient safety in surgery study. J Am Coll Surg 2007;204:1188 98. 3. Port FK, Wolfe RA, Mauger EA, et al. Comparison of survival probabilities for dialysis patients vs cadaveric renal transplant recipients. JAMA 1993;270:1339 43. 4. Moudgil A, Martz K, Stablein DM, et al. Good outcome of kidney transplants in recipients of young donors: a NAPRTCS data analysis. Pediatr Transplant 2011;15:167 71. 5. Nemati E, Saadat AR, Hashemi M, et al. Causes of rehospitalization after renal transplantation; does age of recipient matter? Transplant Proc 2007;39:970 3. 6. U.S. Renal Data System. Seven: transplantation. In: USRDS 2008 annual data report: atlas of ESRD. Available at: http://www.usrds.org/ 2008/view/esrd_07.asp. Accessed March 31, 2013. 7. Lentine KL, Costa SP, Weir MR, et al. Cardiac disease evaluation and management among kidney and liver transplantation candidates: a scientific statement from the American Heart Association and the American College of Cardiology Foundation. J Am Coll Cardiol 2012;60: 434 80. 8. Lentine KL, Santos RD, Axelrod D, et al. Obesity and kidney transplant candidates: how big is too big for transplantation? Am J Nephrol 2012;36:575 86. 9. Meier-Kriesche HU, Arndorfer JA, Kaplan B. The impact of body mass index on renal transplant outcomes: a significant independent risk factor for graft failure and patient death. Transplantation 2002; 73:70 4. 10. Scientific Registry of Transplant Recipients. Risk adjustment models. Available at: http://www.srtr.org/csr/current/modtabs.aspx. Accessed March 31, 2013. 11. Englesbe MJ, Ads Y, Cohn JA, et al. The effects of donor and recipient practices on transplant center finances. Am J Transplant 2008;8: 586 92. 12. Schnitzler MA, Johnston K, Axelrod D, et al. Associations of renal function at 1-year after kidney transplantation with subsequent return to dialysis, mortality, and healthcare costs. Transplantation 2011;91: 1347 56. 13. Gheorghian A, Schnitzler MA, Axelrod DA, et al. The implications of acute rejection and reduced allograft function on health care expenditures in contemporary US kidney transplantation. Transplantation 2012;94:241 9.