SEARCH

SEARCH BY CITATION

Keywords:

  • Medicare costs;
  • pneumonia;
  • renal transplantation;
  • sepsis;
  • USRDS

Abstract

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

We compared the graft survival and accumulative costs associated with sepsis and pneumonia pre- and post-transplantation. We analyzed 44 916 first kidney transplants from 1995 to 2001 USRDS where Medicare was the primary payer. We drew five cohorts for each disease from the baseline population: patients who had a disease onset in the first or second years pre-transplantation (cohorts 1 and 2) or post-transplantation (cohorts 3 and 4) and patients who were disease-free (cohort 5). For each cohort, we calculated graft survival and average accumulated Medicare payments (AAMPs) for the two pre- and post-transplantation years. Graft survival: new-onset sepsis and pneumonia both significantly (p <0.01) lowered graft survival during the year of onset. AAMPs: the AAMPs incurred by sepsis- (pneumonia-) free patients during the first and second years post-transplantation were $50 000 and 13 000 ($51 100 and 13 500), respectively. Patients with a sepsis (pneumonia) onset post-transplantation cost on average $48 400 ($38 400) extra (p<0.01). Episodes of sepsis and pneumonia have a strong and independent impact on graft survival and costs.


Background

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

Despite generally improved outcomes in kidney transplantation over the years, infectious complications remain a substantial cause of morbidity and mortality in adults (1). Post-transplant infections have recently become a more important cause of hospitalizations than acute rejections in pediatric patients (2). Infections and febrile episodes reduce patient and graft survival at the first and third years after transplantation (3,4).

Sepsis and pneumonia are the two most commonly diagnosed infections in dialysis patients and kidney transplant recipients. Based on the United States data for the period 1994–1997, the incidence of hospitalizations for septicemia among renal transplant recipients was approximately 42 times that of the general population after adjustments for age and gender (5). It has also been demonstrated that the risk of sepsis was much higher among transplant recipients than among patients on dialysis (6). Among transplant recipients, hospitalizations for septicemia have been associated with substantially decreased patient survival (5).

Pneumonia is a less frequent but no less severe morbidity. In the transplanted population, pneumonia has been estimated to occur in 2.86 cases per 100 person-years and has been considered a major cause of patient death (7,8). Patients are 3.62 times more likely to contract pneumonia within the first year post-transplant than in the subsequent years (9). The pulmonary infectious mortality rate is also high among dialysis patients (10).

The goal of the current study was to compare the incidence and the financial costs of sepsis and pneumonia in the 2 years after renal transplantation with the incidence and the costs incurred during the 2 years prior to transplantation. In addition, the impacts of sepsis and pneumonia on graft survival were assessed.

Data and Methods

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

We examined data provided by the United States Renal Data System (USRDS) (11). The USRDS is a joint effort of the National Institute of Diabetes and Digestive and Kidney Diseases, and the Centers for Medicare and Medicaid Services (CMS). The collaboration is designed to collect, analyze, and distribute data describing ESRD in the United States including prevalence, treatment modality, survival and cost of care. We used the USRDS transplant and outcome data on all kidney transplantations recorded by the United Network of Organ Sharing (UNOS) and long-term follow-up information from both CMS and UNOS.

Our baseline sample included all 81 712 first kidney transplants reported between January 1, 1995 and December 31, 2001. Of those, Medicare was the primary payer for transplantation (as determined by accumulated institutional claims of at least $5000 and accumulated physician/supplier claims of at least $50 in the first year post-transplant) for 44 929 patients. (We also tested the robustness of our results by employing a stricter definition of the “primary payer.” In particular, we redefined Medicare as the primary payer at transplant if the following three criteria were met: (i) accumulated institutional claims of at least $15 000 in the first year post-transplant; (ii) accumulated physician/supplier claims of at least $2400 in the first year post-transplant and (iii) Medicare indicated as the primary payer at transplant by the TXUNOS and/or the PAYHIST (payer history) file. Our results were robust to this change.) The analysis was performed on 44 916 recipients for whom we could identify the ZIP code area of residence and without obviously miscoded graft failure date.

The number of patients followed on any day diminished in either direction from the date of transplantation (Figure 1). We considered a patient to have been followed from his/her first physician/supplier or institutional claim. Each patient was censored on the day of last claim (3018 cases), last follow-up (312 cases), death (4001 cases), or December 31, 2001 (18 115 cases), whichever occurred first.

image

Figure 1. Number of patients followed.

Download figure to PowerPoint

The incidences of sepsis and pneumonia diagnoses were calculated daily among followed individuals, but were reported as the number of individuals with sepsis/pneumonia claims per 100 patient-years, a measure common in the USRDS. Sepsis/pneumonia claims were drawn both from the USRDS institutional and physician/supplier files and were based on ICD-9-CM diagnosis codes 038 for sepsis and 480–487 for pneumonia. For the purpose of calculating Medicare costs, “onset” of sepsis or pneumonia was defined as the date of the first inpatient and/or the second outpatient claim with a relevant sepsis or pneumonia ICD-9-CM code. Since it may be argued that a patient with true sepsis would not be treated on an outpatient basis, we also conducted our analyses redefining sepsis onset as the date of the first inpatient claim. As expected, this modification leads to somewhat higher cost estimates. Notable differences in results are discussed below.

For our cost calculations, five cohorts for each disease were drawn from the baseline transplant population: (i) patients who had a sepsis (2824) or pneumonia (2142) onset in the period between 730 and 365 days before transplant, (ii) patients who did not have any sepsis or pneumonia claim during that period but had a sepsis (2998) or pneumonia (2240) onset in the first year before transplant, (iii) patients who did not have any sepsis or pneumonia claim in the 2 years before transplant but had a new onset sepsis (4086) or pneumonia (4263) within the first year post transplant, (iv) patients who neither had any sepsis or pneumonia claim in the 2 years before nor the first year post-transplant, but who did have a new onset sepsis (1283) or pneumonia (1304) within the second year post-transplant and (v) patients who did not have any sepsis or pneumonia claim in either the 2 years before or the 2 years after transplant (sepsis: 29 328 patients; pneumonia: 27 739 patients).

Two substantially different methods were used to assess Medicare's costs of treating patients with sepsis or pneumonia from each cohort. The first, which served as a rough proxy for the costs of treating each disease, calculated Medicare's institutional and physician/supplier payments only on days with claims that included a sepsis or pneumonia diagnostic code. These payments were accumulated over the 2 years before and the 2 years after transplantation. Note that in our graphical representation, the accumulated payments were scaled by setting the value at the day of transplantation equal to 0. This transformation portrayed the cost estimates as negative values for the days before transplant and positive values for the days after transplant.

The second method accumulated all costs of treating an average patient in each of our five cohorts. For each cohort, we calculated the Average (i.e. per patient) Accumulated Medicare (institutional and physician/supplier) Payments (AAMPs) for the 2 years before and after transplantation. AAMPs for each day “t” were calculated according to the following formula: AAMP(t) = AAMP(t − 1) +[total Medicare payments(t)/number of individuals followed(t)]. Bootstrap techniques (12) were used to identify significant differences between the AAMPs of each pair of cohorts at four benchmarks: 365 and 730 days before transplantation and 365 and 730 days after transplantation. Note that in our graphical representation of AAMPs, costs before transplant were represented as negative values and calculated backward from the day of transplantation.

Multiple regression analyses of Medicare payments at 1 year post-transplant (for individuals followed at least that year) were used to further confirm the size and significance of the AAMP differences associated with sepsis and pneumonia. In a stepwise procedure, significant determinants of post-transplant Medicare payments were selected from the following list of recipient, donor and transplant characteristics: recipient's age, gender, race and ethnicity; donor type (living vs. cadaveric); donor's age, gender and race; hepatitis C infection at transplant; diabetes as a cause of end stage renal disease; histories of diabetes, cardiovascular disease and peripheral vascular disease; donor and recipient cytomegalovirus (CMV) status; the number of human leukocyte antigen (HLA) mismatches; warm ischemia time; cold ischemia time; panel-reactive antibody percent; immunosuppression regimen (modified cyclosporine, azathioprine and/or mycophenolate mofetil) and the year of transplantation.

Finally, we calculated the graft survival rates over the 2 years post-transplantation separately for our five patient cohorts for each of the two diseases. In particular, we drew Kaplan-Meier graft survival curves for patients who had a sepsis or pneumonia onset in the second or the first year before transplant, patients who had a sepsis or pneumonia onset in the first or the second year after transplant and patients who did not have any sepsis or pneumonia claim either in the 2 years before or in the 2 years after transplant. Pairwise log-rank statistics were used to test the equality of graft survival across our patient cohorts at both the first and the second years post-transplant. In addition, death-censored graft survival rates were calculated for comparison purposes.

Results

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

Incidence

The baseline incidence of sepsis during the 2 years before transplant was 52 episodes per 100 patient-years (Figure 2). During the first post-transplant month, it rose to 133 episodes per 100 patient years and then gradually declined to a post-transplant steady state of 46 episodes per 100 patient years (p<0.001) by the end of the first year.

image

Figure 2. Sepsis incidence 2 years before and after transplant.

Download figure to PowerPoint

The baseline incidence of pneumonia during the 2 years before transplant was 37 episodes per 100 patient years (Figure 3). During the first post-transplant month, it rose to 216 episodes per 100 patient years and then gradually declined to a post-transplant steady state of 60 episodes per 100 patient years (p<0.001) by the end of the first year. As our reading of the medical literature suggests, the high incidence of post-operative pneumonia may be associated with the use of a mechanical ventilator (13,14). According to a recent CDC report, the rate of ventilator-associated pneumonia per 1000 ventilator days ranges between 2.2 in pediatric ICUs and 14.7 in trauma ICUs and patients on mechanical ventilation have 6–21 times the risk of developing hospital-associated pneumonia compared with patients who are not receiving mechanical ventilation (14).

image

Figure 3. Pneumonia incidence 2 years before and after transplant.

Download figure to PowerPoint

Characteristics of the five cohorts

The descriptive statistics associated with our five cohorts differed with respect to various recipient, donor and procedure characteristics (Tables 1 and 2). For example, the proportion of patients that were at least 60 years old with sepsis in the first year post-transplant (26.11%) was greater than the proportion of equally old patients with sepsis in the first year pre-transplant (16.51%, p<0.001). The percentage of living donor transplantations with a sepsis onset in the year immediately preceding transplantation was higher than the percentage of living donor transplantations without any sepsis diagnosis (30.52% vs. 24.92%, p<0.001). Transplantations to patients with pneumonia onset in the first or the second year after transplantation were associated with longer mean cold ischemia time (17.42 and 17.55 h for pneumonia in the first and second years post-transplant, respectively, vs. 16.04 h for no pneumonia, p<0.001). No clinical hypotheses were derived from any of these differences.

Table 1.  Descriptive statistics: sepsis cohorts
VariableSample proportion or mean
Sepsis second year before tx (n = 2824)Sepsis first year before tx (n = 2998)Sepsis first year after tx (n = 4086)Sepsis second year after tx (n = 1283) No sepsis (n = 29 328) Significance of overall differences
  1. When constructing bivariate explanatory variables, missing values were coded as the absence of the status or condition.

  2. * and ** denote statistical significance at the 90% and 99% confidence level, respectively.

  3. Table values are percentages unless otherwise labelled.

Recipient age ≤ 172.413.902.772.182.70**
Recipient age ≥ 45 and < 6031.6231.2933.4632.2733.20 
Recipient age ≥ 6017.2816.5126.1120.8119.63**
Recipient male59.9959.2157.5456.2760.91**
Recipient Black34.9529.3227.2131.8827.02**
Recipient Hispanic13.6012.6411.8211.5413.15*
Living donor17.7830.5218.6020.1124.92**
Donor age (years)35.4035.0437.9837.0235.40**
Donor male53.1252.1751.5955.2653.36*
Donor Black14.9814.1813.3615.5112.25**
Hepatitis C at transplant7.296.746.027.484.83**
Diabetes as a cause of ESRD28.6131.1534.9233.6725.35**
History of diabetes31.9434.5937.1334.9227.38**
History of cardiovascular disease2.802.673.112.262.14**
History of peripheral vascular disease5.494.845.656.393.73**
Donor CMV+/Recipient CMV+42.6939.9041.9242.0540.50 
Donor CMV−/Recipient CMV+24.1423.3321.4720.9324.17**
Donor CMV+/Recipient CMV−19.2620.5022.8320.2019.00**
Number of HLA mismatches3.493.233.383.423.24**
Warm ischemia time (min)22.8022.1622.1822.0822.50 
Cold ischemia time (h)17.4515.5117.0516.6516.08**
Panel-reactive antibody percent ≥ 504.783.973.553.982.93**
Modified cyclosporine54.5755.8760.8265.0059.40**
Azathioprine15.1617.2421.1222.5319.83**
Mycophenolate mofetil60.3060.1754.5555.7359.01**
Table 2.  Descriptive statistics: pneumonia cohorts
VariableSample proportion or mean
Pneumonia second year before tx (n = 2142)Pneumonia first year before tx (n = 2240)Pneumonia first year after tx (n = 4263)Pneumonia second year after tx (n = 1304) No pneumonia (n =27 739) Significance of overall differences
  1. When constructing bivariate explanatory variables, missing values were coded as the absence of the status or condition.

  2. * and ** denote statistical significance at the 95% and 99% confidence level, respectively.

  3. Table values are percentages unless otherwise labelled.

Recipient age ≤ 171.962.412.933.532.82*
Recipient age ≥ 45 and < 6031.2630.0433.6134.8933.26**
Recipient age ≥ 6022.3420.8925.4523.9318.48**
Recipient male57.9057.8661.6759.8260.45**
Recipient Black30.5628.7527.8229.3727.71*
Recipient Hispanic12.7612.7711.6611.1213.40**
Living donor18.9326.3818.0420.7825.13**
Donor age (years)36.0735.9337.4837.1735.22**
Donor male53.0852.2853.0654.3753.39 
Donor Black13.9314.5513.2114.0312.43**
Hepatitis C at transplant6.365.986.196.674.88**
Diabetes as a cause of ESRD30.4232.7231.0829.9825.86**
History of diabetes31.9235.8033.1232.5228.11**
History of cardiovascular disease2.242.902.842.912.18*
History of peripheral vascular disease4.915.675.024.913.96**
Donor CMV+/Recipient CMV+43.6642.7241.3842.8239.94**
Donor CMV−/Recipient CMV+23.7024.1523.8324.3923.50 
Donor CMV+/Recipient CMV−18.2218.2821.9917.7119.55**
Number of HLA mismatches3.333.253.423.303.24**
Warm ischemia time (min)22.2721.8122.4222.6222.48 
Cold ischemia time (h)17.7516.0217.4217.5516.04**
Panel-reactive antibody percent ≥ 504.864.603.243.763.00**
Modified cyclosporine56.8256.4360.4363.3458.84**
Azathioprine16.5418.9720.3621.0919.22**
Mycophenolate mofetil60.8957.9555.9054.7559.40**

Disease-related Medicare payments

Medicare payments for claims that included sepsis diagnoses were smaller when the new onset occurred before transplant. Interestingly, additional sepsis claims continued in subsequent years regardless of when the new onset sepsis occurred (Figure 4). The 2824 patients with sepsis admissions during the period between 730 and 365 days before transplantation averaged $8736 of sepsis-related Medicare payments during that period and $2953, 3628, and 1735 in each subsequent year. The 2998 patients with new onset sepsis during the 365 days before transplant accumulated $8619 of sepsis-related Medicare payments during that period and $5581 and 1499 in each of the subsequent years. The 4086 patients with new onset sepsis during the first post-transplant year accumulated $19 029 of sepsis-related Medicare payments during that year and $2509 in the following year. The 1283 patients with new onset sepsis during the second post-transplant year accumulated $14 964 of sepsis-related Medicare payments during that year. As expected, the sepsis-related Medicare payments calculated when disease onset had been redefined as the date of the first inpatient claim were larger than our baseline estimates. For example, the payments reached $13 409, 12 518, 28 076 and 21 591 during the year of sepsis onset for patients with onset in the second year pre-transplant, the first year pre-transplant, the first year post-transplant, and the second year post-transplant, respectively.

image

Figure 4. Sepsis: claim-related Medicare payments accumulated over the 2 years before and after transplant.

Download figure to PowerPoint

Medicare payments for claims that included a pneumonia diagnosis were also smaller when the new onset occurred before transplant and also continued at lower levels in subsequent years (Figure 5). The 2142 patients with pneumonia admissions during the period between 730 and 365 days before transplantation averaged $6997 of pneumonia-related Medicare payments during that period and $1476, 3664 and 2192 in each subsequent year. The 2240 patients with new onset pneumonia during the 365 days before transplant accumulated $6491 of pneumonia-related Medicare payments during that period and $6782 and 2065 in each of the subsequent years. The 4263 patients with new onset pneumonia during the first post-transplant year accumulated $15 719 of pneumonia-related Medicare payments during that year and $1768 in the following year. The 1304 patients with new onset pneumonia during the second post-transplant year accumulated $11 973 of pneumonia-related Medicare payments during that year.

image

Figure 5. Pneumonia: claimrelated Medicare payments accumulated over the 2 years before and after transplant.

Download figure to PowerPoint

Sepsis AAMPs

The AAMPs represent the accumulation of all claims for patients in each cohort. Not surprisingly, the AAMPs for patients with no sepsis claim during the entire 4-year period were significantly lower than the payments for any of the other four cohorts. On average, Medicare payments for these entirely sepsis-free patients amounted to $27 400 annually during the 2 years before transplantation. Medicare paid approximately $50 000 and 13 000 for these patients during the first and the second year post-transplant, respectively (Figure 6 and Table 3).

image

Figure 6. Sepsis: average accumulated Medicare payments.

Download figure to PowerPoint

Table 3.  Sepsis: average accumulated Medicare payments
 Day −730Day −365Day 365Day 730
  1. The number of patients followed in each cohort on each day is given in parentheses.

Sepsis second year before tx−$102 688−$47 447$67 318$90 615
(2345)(2824)(2269)(1768)
Sepsis first year before tx−$85 631−$54 288$67 888$88 835
(1694)(2465)(2435)(1968)
Sepsis first year after tx−$60 363−$32 417$101 612$130 193
(2409)(3317)(3163)(2555)
Sepsis second year after tx−$60 518−$32 292$58 466$116 633
(716)(996)(1283)(1035)
No sepsis−$54 854−$29 387$49 975$62 953
(15 939)(22 676)(24 177)(19 779)

The additional Medicare payments, as measured by AAMPs, for individuals with sepsis were much greater than anticipated by the payments for sepsis claims reported in the previous section. Patients with a sepsis onset during the first or the second year before transplantation cost an extra $27 400 during the year of the sepsis onset. Patients with a sepsis onset in the first or the second year post-transplant cost on average $48 400 extra during the year of the sepsis onset.

Moreover, patients with a sepsis onset in any year remained more expensive in all subsequent years. When compared to patients entirely free of sepsis, patients with a sepsis onset 2 years before transplant cost Medicare an extra $18 060 during the year before transplant. Patients with a sepsis onset 1 or 2 years before transplantation cost Medicare an additional $17 600 and 9 100 in the first and the second year post-transplant, respectively, when compared to patients without any sepsis. Similarly, patients with a sepsis onset during the first post-transplant year cost Medicare an extra $15 600 during the second post-transplant year.

Finally, patients who eventually had a sepsis claim differentiated themselves from patients who never developed sepsis by having significantly higher Medicare payments before the sepsis onset. For example, patients with a sepsis onset in the year immediately preceding transplantation were $5900 more expensive in the period between 730 and 365 days before transplant than patients without a sepsis onset at any time. Patients with a sepsis onset following transplantation were $2800 more expensive in each year prior to transplantation than patients without sepsis.

Pneumonia AAMPs

A similar set of results was obtained for pneumonia. The AAMPs for patients without any pneumonia during the entire 4-year period were significantly lower than the payments for any of the other four cohorts. Pre-transplant, patients without pneumonia cost Medicare approximately $28 400 annually. Medicare paid approximately $51 100 and 13 500 for these patients during the first and the second year post-transplant, respectively (Figure 7 and Table 4).

image

Figure 7. Pneumonia: average accumulated Medicare payments.

Download figure to PowerPoint

Table 4.  Pneumonia: average accumulated Medicare payments
 Day −730Day −365Day 365Day 730
  1. The number of patients followed in each cohort on each day is given in parentheses.

Pneumonia second year before tx−$95 566−$45 013$68 631$91 774
(1829)(2141)(1712)(1357)
Pneumonia first year before tx−$83 202−$51 610$69 097$91 818
(1355)(1962)(1819)(1463)
Pneumonia first year after tx−$60 549−$32 204$91 529$117 699
(2561)(3517)(3421)(2723)
Pneumonia second year after tx−$61 088−$32 368$57 225$107 059
(737)(1034)(1302)(1061)
No pneumonia−$56 709−$30 291$51 059$64 586
(14 904)(21 269)(22 505)(18 247)

The additional Medicare payments, as measured by AAMPs, for individuals with pneumonia were also much greater than anticipated by the payments for pneumonia-related claims reported previously. In particular, patients with a pneumonia onset during the first or the second year before transplantation cost Medicare an extra $22 800 during the year of the pneumonia onset. Patients with pneumonia in the first or the second year post-transplant cost Medicare on average $38 400 extra during the year of the pneumonia onset.

Moreover, as in the case of sepsis, patients with a pneumonia onset remained more expensive than patients without pneumonia in all subsequent years. Patients with a pneumonia onset 2 years before transplant cost Medicare an extra $14 700 during the year before transplant. Patients with a pneumonia onset 1 or 2 years before transplantation cost Medicare an additional $17 800 and 9400 in the first and the second year post-transplant, respectively. Patients with a pneumonia onset during the first post-transplant year cost Medicare an extra $12 600 during the second post-transplant year.

Finally, patients who eventually acquired pneumonia differentiated themselves from patients who never acquired pneumonia by significantly higher Medicare payments before the pneumonia onset. Patients with a pneumonia onset in the year immediately preceding transplantation were $5200 more expensive in the period between 730 and 365 days before transplant than patients without pneumonia. Likewise, patients with a pneumonia onset following transplantation were $2100 more expensive than the cohort without pneumonia in each of the two years preceding transplantation.

Confirmatory multivariate regressions

Multivariate regression analyses controlling for significant donor, recipient and transplant characteristics confirmed the size of the costs attributable to sepsis or pneumonia during the first post-transplant year. Approximately 12% and 13% of our total sample had sepsis and pneumonia during the first post-transplant year, respectively. Thirty-two percent of the sepsis patients also had pneumonia and thirty percent of the pneumonia patients also had sepsis. Transplant recipients with sepsis (but not pneumonia) in the first post-transplant year incurred 1-year accumulated Medicare payments that were $29 787 (p<0.001) higher than those of patients without sepsis at any time, all other factors held equal (Table 5). Transplant recipients with pneumonia (but not sepsis) in the first post-transplant year incurred 1-year accumulated Medicare payments that were $18 107 (p<0.001) higher than those of patients without pneumonia. And, finally, patients with both sepsis and pneumonia in the first post-transplant year incurred an extra payment of $10 964 (p<0.001) in addition to the sum of the costs of the individual diseases.

Table 5.  Confirmatory multivariate regression: significant determinants of 1-year accumulated Medicare payments
VariableParameter estimate p-level
  1. The following variables were also included in the model but were found to be insignificant: recipient age ≤17 (compared to ≥18 and ≤44), recipient age ≥60 (compared to ≥18 and ≤44), recipient male, living donor, donor male, history of diabetes, donor CMV+/recipient CMV+ (compared to –/–), donor CMV–/recipient CMV+ (compared to –/–), transplant year 1996 (compared to 1995), transplant year 1997 (compared to 1995), and transplant year 2000 (compared to 1995).

Sepsis (transplant year)$29 787<0.001
3713 patients 
Pneumonia (transplant year)$18 107<0.001
4014 patients 
Sepsis and pneumonia interaction (transplant year)$10 964<0.001
1205 patients 
Recipient age ≥45 and <60 (compared to ≥18 and ≤44)−$939=0.045
Recipient Black$1617=0.003
Recipient Hispanic$1511=0.037
Donor age (years)$69<0.001
Donor black$1534=0.029
Hepatitis C at transplant$5509<0.001
Diabetes as a cause of ESRD$5513<0.001
History of cardiovascular disease$3858=0.009
History of peripheral vascular disease$4014<0.001
Donor CMV+/Recipient CMV– (compared to –/–)$1561=0.005
Number of HLA mismatches$875<0.001
Warm ischemia time (min)$27=0.008
Cold ischemia time (h)$140<0.001
Panel-reactive antibody percent ≥50$3707=0.003
Modified cyclosporine (compared to “no CsA”)−$4700<0.001
Azathioprine (compared to “no Aza”)−$2558<0.001
Mycophenolate mofetil (compared to “no MMF”)−$1155=0.047
Transplant year 1998 (compared to 1995)−$2643<0.001
Transplant year 1999 (compared to 1995)−$3331<0.001

Effects of sepsis on graft survival

The onset of sepsis had a significant impact on post-transplant graft survival (Figure 8). While no significant difference was observed between 1-year graft survival of the cohorts with new onset sepsis during the first year before transplant (85.6%), during the second year before transplant (84.0%) and during the second year after transplant (86.2%), graft survival for patients in each of these groups was significantly lower (p<0.01) than for sepsis-free patients (90.9%). The 64.7% 1-year graft survival experienced by patients with a new onset sepsis during the first year post-transplant was markedly and significantly lower (p<0.01) than that of all other cohorts.

image

Figure 8. Graft survival by sepsis onset time.

Download figure to PowerPoint

Graft survival among patients who developed sepsis within the 2 years post-transplant was negatively affected. First, new onset sepsis during the second post-transplant year was associated with reduced graft survival from 86.2% at the end of the first year to 62.0% at the end of the second year. Second, graft survival among the cohort that experienced new onset sepsis in the first post-transplant year fell from 64.7% at the end of the first post-transplant year to 57.8% at the end of the second post-transplant year. When compared over the 2-year period following transplantation, all pairwise differences in graft survival across the five patient cohorts were statistically significant at the 99% confidence level.

For comparison, the death-censored graft survival rates at the end of the second post-transplant year were 90.3% for sepsis-free patients, 81.8% for patients with sepsis onset in the second year pre-transplant, 85.1% for patients with sepsis onset in the first year pre-transplant and 69.3% for patients with sepsis onset in the first or the second year post-transplant.

Effects of pneumonia on graft survival

The onset of pneumonia before or after transplantation had a lower but still very substantial impact on post-transplant graft survival (Figure 9). While no significant difference was observed between 1-year graft survival of the cohorts with new onset pneumonia during the first year before transplant (84.1%) and during the second year before transplant (84.2%), each of these was significantly different (p<0.01) than the 89.9% graft survival for patients without any indication of pneumonia. The 71.0% 1-year graft survival experienced by patients with a new onset pneumonia during the first year post-transplant was significantly lower (p<0.01) than that of all other cohorts.

image

Figure 9. Graft survival by pneumonia onset time.

Download figure to PowerPoint

The graft survival implications of new onset pneumonia at 2 years post-transplant are similar to those of new onset sepsis. First, new onset pneumonia during the second post-transplant year reduced graft survival from 88.5% at the end of first year to 67.6% at the end of the second year. Second, graft survival among the cohort that experienced new onset pneumonia in the first post-transplant year fell from 71.0% at the end of the first post-transplant year to 63.7% at the end of the second post-transplant year. When compared over the 2-year post-transplant period, all pairwise differences in graft survival across the five patient cohorts were statistically significant at the 99% confidence level.

For comparison, the death-censored graft survival rates at the end of the second post-transplant year were 89.2% for pneumonia-free patients, 82.5% for patients with pneumonia onset in the second year pre-transplant, 83.9% for patients with pneumonia onset in the first year pre-transplant and about 74.0% for patients with pneumonia onset in the first or the second year post-transplant.

Conclusion and Discussion

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

While the high incidences of sepsis and pneumonia following kidney transplantation are well documented, the financial and graft survival implications of these infections are less well understood. This study compares pre- and post-transplant outcomes and costs of patients who were actually transplanted. It analyzes the impact of sepsis and pneumonia in cohorts of transplanted patients during the 2 years following transplantation and compares their outcomes and costs with those incurred as dialysis patients.

This study of the U.S. registry data confirmed substantial, but temporary, increases in the incidences of the two diseases immediately following kidney transplantation. It also found that while the incidence of sepsis in the second year post-transplant was slightly lower than that observed in the 2 years before transplantation, the incidence of pneumonia remained significantly above that observed before transplantation.

The onset of sepsis and pneumonia is each associated with a strong and immediate increase in Medicare payments. Patients with a sepsis onset during the first or the second year before transplantation cost, on average, an extra $27 400 during the year of the sepsis onset. Patients with sepsis in the first or the second year post-transplant cost on average $48 400 extra during the year of the sepsis onset. Patients with pneumonia onset during the first or the second year before transplantation cost an extra $22 800 during the year of the pneumonia onset. And, patients with pneumonia in the first or second year post-transplant cost on average $38 400 extra during the year of the pneumonia onset. These cost figures seem substantial compared to published estimates for other post-transplant diseases. For example, using data on patients transplanted between 1996 and 1997, we previously concluded that, by 2 years post-transplant, Medicare paid an extra $21 500 per newly diabetic patient (15). In a different study, hospitalization charges for CMV-related readmissions among renal transplant patients were found to be $22 598 on average (16).

Finally, both sepsis and pneumonia are associated with significantly decreased graft survival rates. For example, at the end of the second year post-transplant, the graft survival rate was 57.8% for patients with a sepsis onset in the first year post-transplant as compared to 87.5% for patients without any indication of sepsis either before or after transplantation. For pneumonia, the corresponding values were 63.7% and 86.3%, respectively.

There is a bidirectional relationship between infection and rejection. Infection may lead to cytokine release and upregulation of adhesion and inflammatory molecules that may predispose patients to rejection. Also, decreases in immunosuppression during sepsis and pneumonia may allow for rejection. While this study is unable to assess these competing principles, it does emphasize that sepsis and pneumonia occurring both before and after transplantation are associated with poor graft survival and markedly increased costs. Thus, a history of sepsis or pneumonia prior to transplant should be assessed when considering post-transplant immunosuppression and in the development of organ allocation algorithms. Moreover, strategies to reduce the incidences of sepsis and pneumonia infections are needed to improve the graft outcomes and the cost-effectiveness of renal transplantation.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References

The authors are indebted to the USRDS for their data and to Novartis for their financial support. RSW is supported in part by R01 DK63230-01. DCB is supported in part by NIH 1 K24-02886.

References

  1. Top of page
  2. Abstract
  3. Background
  4. Data and Methods
  5. Results
  6. Conclusion and Discussion
  7. Acknowledgments
  8. References
  • 1
    Ponticelli C, Tarantino A, Vegeto A. Renal transplantation, past, present and future. J Nephrol 1999; 12: S105-S110.
  • 2
    Dharnidharka VR, Stablein DM, Harmon WE. Post-transplant infections now exceed acute rejection as cause for hospitalization: a report of the NAPRTCS. Am J Transplant 2004; 4: 384389.
  • 3
    Witzke O, Schmidt C, Kohnle M, Lutkes P, Philipp T, Heemann U. Impact of febrile infections on the long-term function of kidney allografts. J Urol 2001; 66: 20482052.
  • 4
    Brayman KL, Stephanian E, Matas AJ et al. Analysis of infectious complications occurring after solid-organ transplantation. Arch Surg 1992; 127: 3847.
  • 5
    Abbott KC, Oliver JD 3rd, Hypolite I et al. Hospitalizations for bacterial septicemia after renal transplantation in the United States. Am J Nephrol 2001; 21: 120127.
  • 6
    Abbott KC, Napier MG, Agodoa LY. Hospitalizations for bacterial septicemia in patients with end stage renal disease due to diabetes on the renal transplant waiting list. J Nephrol 2002; 15: 248254.
  • 7
    Tveit DJ, Hypolite IO, Poropatich RK et al. Hospitalizations for bacterial pneumonia after renal; transplantation in the United States. J Nephrol 2002; 15: 255262.
  • 8
    Sileri P, Pursell KJ, Coady NT et al. A standardized protocol for the treatment of severe pneumonia in kidney transplant recipients. Clin Transplant 2002; 16: 450454.
  • 9
    Chang GC, Wu CL, Pan SH, et al. The diagnosis of pneumonia in renal transplant recipients using invasive and noninvasive procedures. Chest 2004; 125: 541547.
  • 10
    Sarnak MJ, Jaber BL. Pulmonary infectious mortality among patients with end-stage renal disease. Chest 2001; 120: 18831887.
  • 11
    United States Renal Data System (USRDS). USRDS Data CD-ROMS. Minneapolis , MN : January 2003.
  • 12
    Efron B. Better bootstrap confidence intervals. J Am Stat Assoc 1987; 82: 171.
  • 13
    CDC. Guideline for prevention of nosocomial pneumonia. 1997; MMWR 46: RR-1.
  • 14
    CDC. Guidelines for preventing health-care-associated pneumonia, 2003: Recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee. (http://www.cdc.gov/ncidod/hip/guide/CDCpneumo_guidlines.pdf; Accessed June 7, 2005).
  • 15
    Woodward RS, Schnitzler MA, Baty J et al. Incidence and cost of new onset diabetes mellitus among U.S. wait-listed and transplanted renal allograft recipients. Am J Transplant 2003; 3: 590598.
  • 16
    Henderson R, Carlin D, Kohlhase K et al. Multicenter US study of hospital resource utilization associated with cytomegalovirus-related readmissions of renal and heart transplant recipients. Transpl Infect Dis 2001; 3: 5759.