SEARCH

SEARCH BY CITATION

Keywords:

  • African American;
  • immune monitoring;
  • interferon;
  • kidney transplantation;
  • T cell alloimmunity

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Final crossmatch testing is routinely used to assess the risk of antibody-mediated graft injury/rejection post-transplant. Analogously, we postulated that quantitative measurements of anti-donor effector/memory T cells pre-transplant would independently assess post-transplant risk. To address this hypothesis, we determined the frequencies of pre-transplant, donor-specific interferon-γ (IFN-γ) enzyme-linked immunosorbent spots (ELISPOTs) and correlated the results with post-transplant outcomes in 37 African American recipients of deceased donor kidney transplants treated with tacrolimus- and sirolimus-based immunosuppression.

A positive ELISPOT test (>25 spots/300 000 cells) was detected in 14 (38%) of 37 patients. The incidence of biopsy-proven acute rejection was 50% (7/14) in ELISPOT-positive versus 17% (4/23) in ELISPOT-negative patients (p = 0.036). Calculated glomerular filtration rate (MDRD) at 12 months was 37 ± 16 mL/min in ELISPOT-positive versus 55 ± 20 mL/min in ELISPOT-negative patients (p = 0.01). ELISPOT status remained a correlate of allograft function at 12 months by linear regression analysis (p = 0.001), independent of rejection and other contributing variables. Pre-transplant donor-directed IFN-γ ELISPOT assessment of anti-donor cellular immunity may function as a ‘cellular crossmatch’ and independently correlates with renal allograft function in African Americans receiving tacrolimus- and sirolimus-based immunosuppression.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Both immune and nonimmune mechanisms contribute to the pathogenesis of chronic allograft nephropathy (CAN), which remains a leading cause of late allograft failure (1). Analyses linking episodes of acute rejection (AR) and long-term allograft dysfunction lend support to an immune etiology for CAN. Indeed, early AR remains an important predictor of poor long-term allograft function even in the era of modern immunosuppression (2,3). The ability to stratify allograft recipients into high or low immunologic risk could allow for more intelligent and informed titration of immunosuppressive therapy (4). HLA matching and the assessment of pre-transplant alloantibodies by crossmatching techniques are accepted measures of risk stratification, but these tests do not directly appraise anti-donor cellular immunity.

Noninvasive tests measuring cellular immunity pre-transplant could help predict immunologic risk and subsequent allograft dysfunction. Because T cells play a central role in graft rejection and because effector/memory T cells respond more vigorously than naïve T cells, we postulated that measures of the pre-transplant frequency of anti-donor effector/memory T cells would provide an additional independent assessment of post-transplant risk. To address this hypothesis, we developed a highly sensitive enzyme-linked immunosorbent spot (ELISPOT) assay that quantifies the number of interferon-γ (IFN-γ) producing effector/memory T cells in human peripheral blood. We previously demonstrated a crude correlation between pre-transplant frequencies of anti-donor IFN-γ producing PBLs and subsequent AR in 19 kidney transplant recipients (5). In addition, we have shown that serial post-transplant measurements of anti-donor IFN-γ producing cells correlate with late allograft function and with CAN in an ethnically heterogeneous population of kidney transplant recipients (6,7). Nickel and colleagues recently confirmed these results in a German population of kidney transplant recipients treated with a variety of immunosuppressive regimens (8).

In the study reported herein, we analyzed the relationship of pre-transplant donor-reactive IFN-γ ELISPOTs with both early AR and allograft function at 1-year post-transplant in a group of African American deceased donor kidney transplant recipients treated uniformly with a tacrolimus- and sirolimus-based regimen. We found that pre-transplant donor-stimulated ELISPOTs were associated with both early AR and poor renal allograft function at 1 year.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Patient selection

Patients were enrolled under the approved guidelines of the Institutional Review Board for Human Studies at the University Hospitals of Cleveland. The study population consisted of 37 primary renal transplant recipients who received a transplant between January 2000 and December 2003. All patients were African American and were recipients of deceased donor kidney transplants. Patients provided informed consent and were selected for pre-transplant immune monitoring based on the availability of donor stimulator cells consisting of donor splenocytes. Immune monitoring studies were performed prior to the initiation of immunosuppression.

Maintenance immunosuppression consisted of tacrolimus, sirolimus and corticosteroids. Induction therapy with rabbit antithymocyte globulin was given in instances of positive B-cell crossmatch by flow cytometry (n = 6). Positive T-cell crossmatch by flow cytometry or cytotoxic assays precluded transplantation. Adjustment of immunosuppression was not influenced by the results of pre-transplant ELISPOTs, which were not prospectively available to the transplant clinicians. Renal biopsies were performed in all patients with suspected rejection episodes. Biopsies were graded using the Banff 97 classification system (9) by one of two pathologists without knowledge of ELISPOT results. Antibody therapy (either rabbit antithymocyte globulin or OKT3) was used in patients with biopsy-proven rejection who first failed a trial of high-dose steroid therapy. Delayed graft function (DGF) was defined as the need for dialysis during the first week post-transplant. The simplified MDRD formula was used to calculate glomerular filtration rate (GFR) at 12 months (10).

HLA typing and alloantibody determinations

Antigens encoded by HLA class I loci (A and B) were identified by the basic microlymphocytotoxicity assay using local antisera. Class II alleles were determined by sequence-specific priming and PCR. Pre-transplant panel reactive antibody (PRA) was determined by flow cytometry using Flow PRA® beads (One Lambda, Canoga Park, CA) according to manufacturer's recommendations. Patients were defined as sensitized if their final pre-transplant flow PRA was >20%.

Enzyme-linked immunosorbent spot (ELISPOT) assays

IFN-γ ELISPOT assays were performed as previously described in detail (5,11). A total of 300 000 responder PBLs in 100 μL of T-cell medium (90% RPMI 1640 medium) (Sigma Chemical Co., St. Louis, MO), plus 10% human serum with L-glutamine plus penicillin/streptomycin (BioWhittaker, Walkersville, MD), were immediately placed in 96-well ELISPOT plates (Millipore, Bedford, MA) pre-coated with capture anti-IFN-γ antibody (Endogen, Woburn, MA). PBLs were stimulated with medium alone (negative control), T-cell depleted stimulator cells, and a positive control, phytohemagglutinin (PHA at 1 μg/mL of medium (Murex Diagnostics, Dartford, UK)). Plates were then incubated overnight at 37°C. Following three washes with PBS and PBS-Tween, a biotinylated, anti-IFN-γ antibody (Endogen) was added to detect bound cytokine, and the plates were incubated overnight at 4°C. After an additional wash, streptavidin horseradish peroxidase conjugate (Dako, Denmark) was added for 1 h at room temperature. After a final wash, the plates were developed with aminoethylcarbazole (10 mg/mL in N,N-dimethylformamide; Pierce Chemicals, Rockford, IL), prepared in 0.1 M sodium acetate buffer (pH 5.0) mixed with H2O2.

The resulting spots were counted with a Series 1 Immunospot computer-assisted ELISPOT image analyzer (Cellular Technology, Cleveland, OH). Results were depicted as the mean number of IFN-γ spots/300 000 recipient PBLs based on duplicate or triplicate measurements in a given assay. Previous work has demonstrated that <10 spots/300 000 cells represent background reactivity. A positive test was defined as a >25 spots/300 000 PBLs. The well-to-well and assay-to-assay variability is 20–30% (11). Control wells assessing cytokine production by stimulators alone were included in all assays (<20 spots/300 000) and detected spots in these control wells were subtracted from the total number of spots in wells in which responders and stimulators were mixed.

Statistical analyses

All analyses were performed using SPSS version 11.5 (SPSS, Chicago, IL). Values are shown as mean ± SD, median (range) or percentage. Baseline demographic data between patient groups were analyzed using Student's t-test for continuous variables and Fisher's exact test for dichotomous variables. Mann-Whitney U test was used to compare the total number of IFN-γ spots between groups, and also to compare serum creatinine between groups at 12 months. Logistic regression analysis was used to determine factors associated with AR in the first year post-transplant. Pearson correlation was used for ELISPOT (+) status and GFR, and Spearman correlation for absolute number of IFN-γ ELISPOTs and GFR. Finally, univariate and multivariate linear regression analyses were performed to determine correlates of MDRD GFR at 12 months. Two-sided p-values of less than 0.05 were considered to indicate statistical significance.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

A total of 37 consecutive African American patients with pre-transplant donor-specific ELISPOT testing were included in the analysis. Fourteen (38%) of 37 were ELISPOT (+). Baseline demographic data comparing ELISPOT (+) and (−) patients is shown in Table 1. There was no difference in age, gender or donor age between groups. The number of HLA mismatches, percent of patients sensitized, time on dialysis and the use of induction therapy were also similar between groups. The median number of IFN-γ spots/300 000 cells was 1 (range: 0–23) in ELISPOT (−) and 48 (range: 27–236) in ELISPOT (+) patients (p < 0.001).

Table 1.  Baseline characteristics
 ELISPOT (−) (n = 23)ELISPOT (+) (n = 14) p-value
Age47 ± 1347 ± 12ns
Male, %6557ns
Cold time, h20 ± 522 ± 5ns
Donor age45 ± 1442 ± 15ns
HLA mismatch4.8 ± 1.04.2 ± 1.5ns
Sensitized, %2614ns
Dialysis, months58 ± 2962 ± 31ns
Induction Rx, %1714ns
DGF, %3529ns
Med. IFN-γ spots (range) (per 300 000 cells)1 (0–23)48 (27–236)<0.001

Seven (50%) of 14 ELISPOT (+) patients experienced AR versus 4 (17%) of 23 in the ELISPOT (−) group (p = 0.036). AR tended to occur early in both groups, with 73% (8/11) of episodes occurring in the first 3-month post-transplant. Neither ELISPOT (+) test nor the absolute number of IFN-γ spots correlated with HLA mismatch, PRA or DGF. Logistic regression analysis was performed controlling for known risk factors for AR (Table 2). The association between ELISPOT (+) status and AR showed borderline significance in this model (odds ratio for ELISPOT (+): 6.3, 95% CI: 1.0–41.1, p = 0.055).

Table 2.  Logistic regression analysis for acute rejection
VariableOdds ratio95% CI p-value
HLA mismatch0.50.2–1.4ns
Sensitized1.10.1–13.1ns
Induction Rx0.20.1–4.8ns
DGF15.11.6–145.00.019
ELISPOT (+)6.31.0–41.10.055

All four cases of AR in the ELISPOT (−) patients were graded as Banff IA. Three of these patients responded to pulse steroids alone, while one required antibody therapy. All four had functioning allografts at 2-year post-transplant. In the ELISPOT (+) group, the Banff scores were IA (four); IB (one); IIA (one) and IIB (one). Five of seven patients required antibody therapy following pulse steroids. Two of the seven patients lost their allografts and returned to dialysis within 2 years.

We next examined the relationship between pre-transplant IFN-γ ELISPOTs and renal allograft function, using MDRD GFR at 12-month post-transplant. Mean serum creatinine at 12 months was 2.5 ± 1.0 mg/dL for ELISPOT (+) patients versus 1.8 ± 0.8 mg/dL for ELISPOT (−) patients (p = 0.013). The difference in 12-month GFR between ELISPOT groups is shown in Figure 1. Mean GFR was 37 ± 16 mL/min in ELISPOT (+) versus 55 ± 20 mL/min in ELISPOT (−) (p = 0.01). There was a negative Pearson correlation with ELISPOT (+) status and 12-month GFR (r=−0.43, p = 0.01), and a similar negative Spearman correlation with absolute number of IFN-γ spots and GFR (r=−0.54, p = 0.001).

image

Figure 1. MDRD GFR (mL/min) at 12-month post-transplant in IFN-γ ELISPOT (−) patients (n = 23), and IFN-γ ELISPOT (+) patients (n = 14). Shown are the median values with inter-quartile range (box borders), and the extreme values (whiskers).

Download figure to PowerPoint

Univariate logistic regression was used to determine variables associated with 12-month GFR (Table 3). Donor age, DGF, AR and ELISPOT (+) status were all associated with lower GFR. Recipient age, donor gender, HLA mismatch, pre-transplant sensitization, time on dialysis and the use of induction therapy did not correlate with GFR in this analysis, although the small sample size precluded conclusions regarding these associations. Multivariate linear regression analysis, using MDRD GFR as the dependent variable, is shown in Table 4. Along with older donor age and DGF, ELISPOT (+) status remained associated with lower GFR at 12 months (slope =−19.9, 95% CI: −31.1 to −8.6, p = 0.001).

Table 3.  Univariate analysis for variables associated with GFR at 12 months
Variable Slope95% CI of slope p-value
Older donor age (years)−0.5−1.0 to −0.10.025
DGF−19.0−33.0 to −5.40.008
Acute rejection−15.3−29.9 to −0.70.041
ELISPOT (+)−17.7−30.9 to −4.50.010
Table 4.  Multivariate linear regression for GFR at 12 months
VariableSlope Slope95% CI of slope p-Value
Older donor age (years)−0.4−0.8 to −0.10.024
DGF−16.7−28.8 to −4.60.009
Acute rejection−4.7−17.1 to 7.6ns
ELISPOT (+)−19.9−31.1 to −8.60.001

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Our findings suggest that the anti-donor ELISPOT assay performed pre-transplant may function as a ‘cellular crossmatch’ that is associated with post-transplant allograft function in a homogeneous population of African American deceased donor kidney transplant recipients. Notably, this association was independent of variables known to impact on the outcome of kidney allografts, such as AR and DGF. Because renal function at 1 year is predictive of long-term renal graft survival (12), the results of pre-transplant ELISPOT testing may also correlate with graft survival.

We chose to study a population of African American kidney transplant recipients, with hopes of gaining more insight into risk factors for poor outcomes in this group. Kidney transplant outcomes in African Americans are inferior to those of other ethnic groups (13–15). Immunologic risk factors such as relatively poor HLA matching influence outcomes in African Americans (13,14). However, differences in graft survival are not entirely explained by traditional immunologic risk factors. In fact, outcomes remain inferior to those of other ethnic groups after controlling for HLA matching (14,15). This has led to speculation that socioeconomic and behavioral factors might be to blame for poor outcomes (16). However, non-HLA genetic factors influencing the strength of immune responses also may play a role.

The ELISPOT test of T-cell alloreactivity may be a measure of immune responsiveness between a recipient and donor that is independent of HLA matching. Previous studies have suggested that African American patients generally exhibit high levels of T-cell reactivity. For example, compared to non-African Americans, African Americans exhibit greater nonspecific T-cell activity measured by mixed lymphocyte culture (MLC) (17). A study in a pediatric transplant population found greater expression of B7 costimulatory molecules CD80 and CD86 from antigen-presenting cells in African Americans, with subsequent increased cell proliferation in MLC (18). Recently, preliminary data from Poggio et al. found that African American ethnicity, along with younger age, independently predicted a higher number of IFN-γ ELISPOTs in a cohort of dialysis patients awaiting transplant (19).

In our study and in previous reports of ELISPOT immune monitoring, IFN-γ activity did not correlate with HLA matching (5,20). In addition, IFN-γ ELISPOTs appear to be independent of PRA as a measure of humoral immunity. As an interesting example, one ELISPOT (+) patient in our cohort received a zero HLA-mismatched deceased donor kidney and had a pre-transplant flow PRA of 0%. Nevertheless, he had AR at 10-month and ultimate graft failure at 42-month post-transplant. This was despite good compliance in this patient with medications and appointments. His mean trough levels of tacrolimus and sirolimus in the 3 months prior to rejection were 6.5 ± 2.6 and 13.5 ± 4.1 ng/mL, respectively, and levels 3 days prior to biopsy for rejection were 7.3 and 18.2 ng/mL, respectively. Thus, ELISPOT testing may offer immunologic insight not available through current HLA matching or alloantibody testing. Analogous to the final crossmatch that is used to determine the presence of donor-specific alloantibodies, the pre-transplant donor-stimulated IFN-γ ELISPOT assay may be viewed as a measure of donor-directed T-cell alloreactivity just prior to transplantation. We did not use ELISPOT testing against third parties in this analysis. Therefore, it is not clear if higher levels of IFN-γ were donor-specific or reflected overall nonspecific increases in T-cell reactivity in ELISPOT (+) patients.

The majority of deceased donor recipients included in this study had good early outcomes on a regimen of tacrolimus, sirolimus and corticosteroids. Approximately two-thirds of the patients were ELISPOT (−), and outcomes in these patients were generally favorable despite minimal use of induction antibody therapy. AR occurred in 4 of 23 ELISPOT (−) patients but responded to steroids alone in 3 of these patients, and allograft function remained stable in the first year in these patients as well as in those without AR. Thus, within a higher-risk African American population, ELISPOT testing may help to define not only patients at high risk of AR, but also those with low immunologic risk who may not require as much immunosuppression.

Whether routine induction antibody therapy would alter outcomes related to the ELISPOT assay in this population is unclear. Only one ELISPOT (+) patient received induction therapy. Notably, she had a high number of IFN-γ ELISPOTs (236 spots) pre-transplant. She had no rejection and maintained stable allograft function in the first year, with a serum creatinine of 1.1 mg/dL at 12 months. Repeat ELISPOT testing in this patient at 6-month post-transplant using donor cells revealed only five spots. Thus, induction therapy with antithymocyte globulin may mitigate higher T-cell reactivity measured pre-transplant, although future studies using ELISPOT with induction therapy are obviously necessary to make any conclusions.

AR was more frequent in ELISPOT (+) patients by univariate analysis. In addition, AR in ELISPOT (+) patients tended to be more severe, and required more aggressive therapy. However, after controlling for known risk factors for AR, the association with ELISPOT status became borderline for statistical significance. With the small number of patients in this study, the multivariate analysis using AR as the dependent variable was flawed by a lack of statistical power. Thus, the data associating ELISPOT and AR is provocative but inconclusive, and further testing is required to prove a definitive link.

Renal allograft function at 12 months correlated with a positive ELISPOT test in this analysis and was independent of other factors, including donor age, DGF and AR. It is possible that ELISPOT (+) patients had higher rates of subclinical immunologic damage to the allograft. Lack of protocol biopsies is one limitation of this study as subclinical rejection on such biopsies has been shown to contribute to CAN (21). Future studies that incorporate both ELISPOT testing and protocol biopsies may determine whether immune-mediated renal damage correlates with IFN-γ expression in the absence of overt rejection.

In conclusion, we found that pre-transplant, donor-directed stimulation of recipient T cells, as measured by IFN-γ ELISPOTs, was independently associated with allograft function at 1 year. Future studies are necessary to determine whether ELISPOT testing has an independent association with clinical and subclinical rejection, and whether adjustment of immunosuppression based on ELISPOT testing will allow for improved outcomes after kidney transplantation.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

This work was supported in part by NIH Contract N01-A1-05410 and by the Leonard Rosenberg Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
  • 1
    Hariharan S, Johnson CP, Bresnahan BA, Taranto SE, McIntosh MJ, Stablein D. Improved graft survival after renal transplantation in the United States, 1988–1996. N Engl J Med 2000; 342: 605612.DOI: 10.1056/NEJM200003023420901
  • 2
    Cosio FG, Pelletier RP, Falkenhain ME et al. Impact of acute rejection and early allograft function on renal allograft survival. Transplantation 1997; 63: 16111615.
  • 3
    Meier-Kriesche HU, Ojo AO, Hanson JA et al. Increased impact of acute rejection on chronic allograft failure in recent era. Transplantation 2000; 70: 10981100.DOI: 10.1097/00007890-200010150-00018
  • 4
    Heeger PS, Hricik D. Immune monitoring in kidney transplant recipients revisited. J Am Soc Nephrol 2002; 13: 288290.DOI: 10.1159/000057858
  • 5
    Heeger PS, Greenspan NS, Kuhlenschmidt S et al. Pretransplant frequency of donor-specific, IFN-γ-producing lymphocytes is a manifestation of immunologic memory and correlates with the risk of posttransplant rejection episodes. J Immunol 1999; 163: 22672275.
  • 6
    Hricik DE, Rodriguez V, Riley J et al. Enzyme linked immunosorbent spot (ELISPOT) assay for interferon-gamma independently predicts renal function in kidney transplant recipients. Am J Transplant 2003; 3: 878884.DOI: 10.1034/j.1600-6143.2003.00132.x
  • 7
    Poggio ED, Clemente M, Riley J et al. Alloreactivity in renal transplant recipients with and without chronic allograft nephropathy. J Am Soc Nephrol 2004; 15: 19521960.DOI: 10.1097/01.ASN.0000129980.83334.79
  • 8
    Nickel P, Presber F, Bold G et al. Enzyme-linked immunosorbent spot assay for donor-reactive interferon-gamma-producing cells identifies T cell presensitization and correlates with graft function at 6 and 12 months in renal-transplant recipients. Transplantation 2004; 78: 16401646.DOI: 10.1097/01.TP.0000144057.31799.6A
  • 9
    Racusen LC, Solez K, Colvin RB et al. The Banff 97 working classification of renal allograft pathology. Kidney Int 1999; 55: 713723.DOI: 10.1046/j.1523-1755.1999.00299.x
  • 10
    Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of diet in Renal Disease Study Group. Ann Intern Med 1999; 130: 461470.
  • 11
    Gebauer BS, Hricik DE, Atallah A et al. Evolution of the enzyme-linked immunosorbent spot assay for post-transplant alloreactivity as a potentially useful immune monitoring tool. Am J Transplant 2002; 2: 857866.DOI: 10.1034/j.1600-6143.2002.20908.x
  • 12
    Hariharan S, McBride MA, Cherikh WS, Tolleris CB, Bresnahan BA, Johnson CP. Post-transplant renal function in the first year predicts long-term kidney transplant survival. Kidney Int 2002; 62: 311318.DOI: 10.1046/j.1523-1755.2002.00424.x
  • 13
    Cecka JM. The UNOS Scientific Renal Transplant Registry—2000. Clin Transpl 2000; 118.
  • 14
    Koyama H, Cecka JM, Terasaki PI. Kidney transplant in black recipients; HLA matching and other factors affecting long-term graft survival. Transplantation 1994; 57: 10641068.
  • 15
    Isaacs RB, Nock SL, Spencer CE et al. Racial disparities in renal transplant outcomes. Am J Kidney Dis 1999; 34: 706712.
  • 16
    Curtis JJ. Kidney transplantation: racial or socioeconomic disparities? Am J Kidney Dis 1999; 34: 756758.
  • 17
    Kerman RH, Kimball PM, Van Buren CT, Lewis RM, Kahan BD. Possible contribution of pretransplant immune responder status to renal allograft survival differences of black versus white recipients. Transplantation 1991; 51: 338342.
  • 18
    Hutchings A, Purcell WM, Benfield MR. Increased costimulatory responses in African-American kidney allograft recipients. Transplantation 2001; 71: 692695.DOI: 10.1097/00007890-200103150-00021
  • 19
    Poggio ED, Clemente M, Roddy M, Hricik DE, Heeger PS. Factors influencing humoral and cellular alloimmunity in hemodialysis patients awaiting renal transplantation [abstract]. J Am Soc Nephrol 2004; 15(Suppl.): 743A.DOI: 10.1097/01.ASN.0000113315.81448.CA
  • 20
    Naether BJ, Nickel P, Presber F et al. Highly significant correlation of donor-specific T cell frequencies with 6 months renal function by improved ELISPOT technique [abstract]. Am J Transplant 2004: 4(Suppl. 8); 162.
  • 21
    Nankivell BJ, Borrows RJ, Fung CL, O'Connell PJ, Allen RD, Chapman JR. Natural history, risk factors, and impact of subclinical rejection in kidney transplantation. Transplantation 2004; 78: 242249.DOI: 10.1097/01.TP.0000128167.60172.CC